ARCHIMEDES

in the Middle Ages

VOLUME FOUR
Memoirs of the A SUPPLEMENT ON THE MEDIEVAL
AMERICAN PHILOSOPHICAL SOCIETY
held at Philadelphia
LATIN TRADITIONS OF CONIC SECTIONS
for Promoting Useful Knowledge ( 1150- 1566)
Volume 137 Part A

P a r t I: Texts and Analysis

MARSHALL CLAGETT

THE AMERICAN PHILOSOPHICAL SOCIETY

Independence Square
Philadelphia
1980
 PREFACE

When I was preparing the edition of William of Moerbeke’s translations

of the works of Archimedes which appeared as Volume Two of my work,
I decided to investigate separately the question of whether the Archime­
dean and Eutocian treatment of conic sections exhibited in the translations
of Moerbeke exerted any influence on medieval mathematics. While I
found little or no influence of these translations on the medieval treat­
ment of conic sections, I did discover a number of medieval texts on
conic sections (largely but not exclusively a part of medieval optical tra­
ditions) that revealed a steadily increasing interest in the geometry of conic
sections. The results of my investigation constitute this fourth volume.
Copyright 1980 by the American Philosophical Society I have here followed the procedures of publication adopted in Volumes
Two and Three. Once more I have published the diagrams and indexes
in a separately bound fascicle. My system of numbering the diagrams
is obvious. The Arabic numeral before the period designates the chapter
for which the diagram is relevant and the succeeding Arabic numeral
or numerals the ordinal position of the diagram within the chapter. The
indexes of Latin mathematical terms, manuscripts cited, and names and
works are similar to those found in Volume Three. I have also added an
index of Apollonian definitions and propositions so that the reader may
quickly appreciate those parts of the Conics of Apollonius that were
known and used in the Latin Middle Ages and Early Renaissance.
Again I must thank the many libraries which provided me with access
to and microfilms of manuscripts and early printed editions, and most
especially the British Library where I spent many profitable hours on sev­
eral occasions. Furthermore, thanks are due to Professors David Lindberg
of the University of Wisconsin and Sabetai Unguru of the University of
Oklahoma for making available to me some of their microfilms. As usual,
I have learned much from discussions with my colleagues at The Institute
for Advanced Study of various problems of philology and mathematics,
and particularly from those with Professor Harold Cherniss and Dr. Her­
man Goldstine.
Library of Congress Catalog Card Number 62-7218 Thanks must also be given to my research assistants Mr. Glenn Sterr
International Standard Book Number 0-87169-137-X and Dr. Peter Marshall for their efforts in preparing the indexes and
US ISSN 0065-9738 reading proof, and to my secretary Mrs. Ann Tobias for typing and re­
typing this volume, for preparing the final diagrams, and for reading proof.
Needless to say, I am grateful to the American Philosophical Society for
undertaking the publication of these volumes, and especially to its edi­
torial staff for guiding them through the press. Finally, I must acknowledge
with special thanks the help and support of the staff of The Institute for
Advanced Study and its director Dr. Harry Woolf. Not only has the In­
stitute provided an ideal academic home but it has generously given finan­
cial support to these volumes.
Marshall Clagett
 Contents

PAGE
Part I: T exts and Analysis

Chap. 1: The Latin Works o f Alhazen and the K nowledge

o f Conic Sections: De speculis comburentibus and
Perspectiva 3
Chap. 2: John o f Palerm o’s Translation o f a Short Arabic
Tract on the Hyperbola 33
Text A: The Treatise on Two Lines Always
Approaching Each Other but Never Meeting 44
Text B: Another Version of the De duabus lineis 55
Chap. 3: William o f M oerbeke and W itelo 63
Chap. 4: The Speculi almukeficompositio 99
Text: The Composition o f theParabolic Mirror 114
Chap. 5: Conic Sections in the Fourteenth and Fifteenth
Centuries: Jean Fusoris, Giovanni Fontana, and
Regiomontanus 159
Text A: Johannes Fusoris, The Booklet on the
Parabolic Section 185
Text B: N otes for A lhazen’s De speculis com­
burentibus as prepared by Giovanni Fontana 200
Text C: De speculis comburentibus (= Speculi al­
mukefi compositio) as prepared by Regiomontanus 203
Text D: Additions by Regiomontanus to the Speculi
almukefi compositio 222
Chap. 6: The M edieval Traditions o f Conic Sections in the
Early Sixteenth Century: Giorgio Valla, Johann
Werner, and Albrecht Diirer 235
Text: Johann Werner, The Booklet on Twenty-two
Conic Elements 269
Chap. 7: The Last Stages o f the M edieval Traditions o f
Conic Sections 311
Text A: Orontius Finaeus, De speculo ustorio 360
Text B: Extracts from F. Barozzi’s Admirandam illud
geometricum problema 384
vii
Part II: Bibliography, Diagrams, and Indexes

Bibliography 465
Diagrams 471

Index of Latin Mathematical Terms 519

Index o f Manuscripts Cited 553

Index o f Apollonian Definitions and Propositions 557

Index o f N am es and Works 561

PART I

TEXTS A N D A N A L Y S IS

viii
 CHAPTER 1

The Latin W orks o f A lhazen and

the K now ledge o f Conic Sections:
De speculis comburentibus
and Perspectiva

Before the twelfth century the knowledge of conic sections in the Latin
West was non-existent. It is true that occasionally the terms ellipsis,
hyperbola, and parabola had been used in earlier Latin texts but without
their mathematical meanings.1 The first traces of any knowledge of conic
sections in the West came as the result of the Latin translations of two
works of Alhazen (Ibn al-Haytham). The first was the translation by Gerard
of Cremona of Alhazen’s Liber de speculis comburentibus, a work on
the mathematical theory and construction of paraboloidal mirrors.2 To
this work, Gerard prefaced a short fragment translated from the Arabic
text of the introduction to Book I of Apollonius’ Conics and it is this
fragment that demands our initial attention. It begins in the best manuscript
(Paris, BN lat. 9335, 85v):3 “ The things that follow are in the beginning
of the Liber de pyramidibus of Apollonius; they are axioms which he

1 See my Archimedes in the Middle Ages, Vol. 2 (Philadelphia, 1976), p. 433.

2 Gerard’s translation has been edited by J. L. Heiberg and E. Wiedemann, “ Ibnal Haitams
Schrift uber parabolische Hohlspiegel,” Bibliotheca mathematica, 3. Folge, Vol. 10 (1909-
10), whole article, pp. 201-37; Latin text, pp. 218-31; German translation of the Arabic
text, pp. 205-18. An English translation of the Arabic text was published by H. J. J.
Winter and W. 'Arafat, “ Ibn al-Haitham on the Paraboloidal Focussing Mirror,” Journal
of the Royal Asiatic Society o f Bengal. Science, Vol. 15 (1949), pp. 25-40. I have used the
Arabic text in Majamu' al-rasa’il . . . ibn al-Haitham (Hyderabad, 1357 A.H.), 3rd tract.
For the latest list of the manuscripts of Gerard’s translation, see D. Lindberg, A Catalogue
of Medieval and Renaissance Optical Manuscripts (Toronto, 1975), pp. 20-21.
3Cf. J. L. Heiberg, ed., Apollonii Pergaei Quae graece exstant, Vol. 2 (Leipzig, 1893),
p. LXXV. Concerning the high quality of MS Paris, BN lat. 9335 for the texts of the
translations of Gerard of Cremona, see my Archimedes in the Middle A ges, Vol. 1 (Madison,
Wise., 1964) p. 227, η. 1. Experts now date the manuscript as early 13c. The quoted
statement in BN lat. 9335, 85v, runs: “ Ista que sequuntur sunt in principio libri Apol[l]onii
de pyramidibus; sunt anxiomata (/) que premittit in libro illo.”

3
4 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 5

premises in that book.” Another manuscript adds: ”They are also valid straight line is produced in both directions, and the point is fixed so that it may
for the Liber de speculis comburentibus. ” 4 Then follow the First Defini­ not move,Hand the straight line revolves on the circumference of the circle
tions, to which I have added bracketed numbers that are equivalent to until it returns to the place from which it began: then 1 call each of the two
those used by Heiberg in his Latin translation of the Greek text.5 surfaces described by the line in revolution (and each of which surfaces is
opposite to its companion and susceptible to infinite extension since the
[1] When a straight line is drawn between some point and the circumference extension of the line is without end) the “ surface of a cone.” 7 And I call the
of a circle, and the circle and the point are not in the same plane, and the fixed point the “ apex of each of the two surfaces of the two cones.” And I
call the straight line which passes through this point and through the center
4 See Heiberg, Apollonii . . . Quae graece exstant, Vol. 2, p. LXXV: '‘valent etiam
ad librum de speculis comburentibus.” of the circle the “ axis of the cone.”
5 Ibid., Vol. 1 (Leipzig, 1891), pp. 6 -8 for the Greek text; and for the Latin text, Vol. [2] And I call the figure contained by the circle and the conic surface
2, pp. LXXV-LXXVIII. The latter varies little from BN lat. 9335, 85v, which I now quote: between the apex and the circle a “ cone.” And I again call the point which
” [1] Cum continuatur inter punctum aliquod et lineam continentem circulum per lineam is the apex of the surface of the cone the “ apex of the cone.” And I call the
rectam, et circulus et punctum non sunt in superficie una, et extrahitur linea recta in ambas straight line drawn from the apex of the cone to the center of the circle the
partes, et figitur punctum ita ut non moveatur, et revolvitur linea recta super periferiam “ axis of the cone.” 8 And I call the circle the “ base of the cone.”
circuli donec redeat ad locum a quo incepit, tunc ego nomino unamquamque duarum super- [3] And I call a cone “ right” when its axis is erected at right angles to
ficierum quas designat linea revoluta per transitum suum, et unaqueque quarum est opposita its base. I call it “ oblique” (declivem) when its axis is not erected orthogonally
sue compari et susceptibilis additionis infinite cum extractio linee recte est sine fine, super­
ficiem piramidis. Et nomino punctum fixum caput cuiusque duarum superficierum duarum
on its base.
piramidum. Et nomino lineam rectam que transit per hoc punctum et per centrum circuli
[4] And when, from a point9 of any curved line (linee munani) which is
axem piramidis. [2] Et nomino figuram quam continet circulus et quod est inter punctum in one plane, some straight line is drawn in that plane that bisects all lines
capitis et inter circulum de superficie piramidis piramidem. Et nomino punctum quod est drawn within the curved line and having their extremities on it and parallel10
caput superficiei piramidis caput piramidis iterum. Et nomino lineam rectam que protrahitur to some posited line; then I call the [bisecting] line the “ diameter of the
ex capite piramidis ad centrum circuli axem piramidis. Et nomino circulum basim piramidis. curved line.” And I call the extremity of that line [i.e. the diameter] which
[3] Et nomino piramidem orthogoniam cum eius axis erigitur super ipsius basim secundum is at the curved line the “ vertex of the curved line.” And I call the parallel
rectos angulos. Et nomino ipsam declivem (mailan) quando non est eius axis erectus ortho- lines which I have mentioned11 “ lines of order [i.e. double ordinates] to
gonaliter super ipsius basim. [4] Et cum a puncto omnis linee munani (munhanin) que est that diameter.”
in superficie una plana protrahitur in eius superficie linea aliqua recta secans omnes lineas
que protrahuntur in linea munani et quarum extremitates ad eam, et est equidistans linee
alicui posite, in duo media et duo media, tunc ego nomino illam lineam rectam diametrum 6 The clause “ so that it may not move” is a redundancy of the Arabic text. Cf. the
illius linee munani. Et nomino extremitatem illius linee recte que est apud lineam munani variant readings from the Arabic text reported by Heiberg, Gr 2, p. LXXI (1 p. 6, 5) and
caput linee munani. Et nomino lineas equidistantes quas narravi lineas ordinis (tartlbi) Nix, op. cit., Arabic text, p. 4, 1. 4 (I have numbered the lines beginning from the title).
illi diametro. [5] Et similiter iterum cum sunt due linee munani in superficie una, tunc From this note onward the two volumes of Heiberg’s edition of Apollonius will be abbre­
ego nomino quod cadit inter duas lineas munani de linea recta que secat omnes lineas viated respectively Gr 1 and Gr 2.
rectas egredientes in unaquaque duarum linearum munani equidistantes linee alique in duo 7 In the Greek text it is the double conical surface that is named the “ conic surface”
media et duo media diametrum mugeniben (mujaniban). Et nomino duas extremitates dia­ while here each of the two surfaces is so named. Furthermore, here each surface is rather
metri mugenibi (mujanibi) que sunt super duas lineas munani duo capita duarum linearum improperly called the “ surface of a cone” rather than a “ conic surface. Similarly the apex
munanieni (munhaniain). Et nomino lineam rectam que cadit inter duas lineas munanieni and the axis are improperly described as being of a cone rather than of a conic surface.
et punctum super diametrum mugenib et secat omnes lineas rectas equidistantes diametro I say “ improperly” for all three cases since the surfaces are considered indefinitely produced
mugenib, cum protrahuntur inter duas lineas munanieni donec perveniant earum extremi­ and the cones cut off by base circles have not yet been defined. Indeed in the last two
tates ad duas lineas munanieni, in duo media et duo media diametrum erectam (qutran cases the Greek text simply does not have the equivalent of “ duarum superficierum duarum
qa ’iman). Et nomino has lineas equidistantes lineas ordinis ad illam diametrum erectam. piramidum” and “ piramidis” (cf. the Arabic variants in Heiberg, Gr 2, LXXI [I p. 6, 12]
[6] Et cum sunt due linee recte que sunt due diametri linee munani aut duarum linearum and Nix, op. cit., Arabic text, p. 4, lines 10 and 11).
munanieni, et unaqueque secat lineas equidistantes alteri in duo media et duo media, tunc 8 Again “ piramidis” here and in the succeeding sentence are additions of the Arabic
nomino eas duas diametros muzdaguageni (muzdawijain). [7] Et nomino lineam rectam, text. Cf. Heiberg, Gr 2, LXXI (I. p. 6, 18 and 19) and Nix, op. cit., Arabic text, p. 4,
cum est diameter linee munani aut duarum linearum munanieni et secat lineas equidistantes line 16.
que sunt linee ordinis ei secundum angulos rectos, axem linee munani aut duarum linearum 9 The word “ point” has been added in the Arabic text. Cf. Heiberg, Gr 2, p. LXXI
munanieni. [8] Et nomino duas diametros, cum sunt muzdaguageni et secat unaqueque (I p. 6, 24) and Nix, op. cit., Arabic text, p. 5, line 3.
earum lineas equidistantes alteri secundum rectos angulos, duos axes muzdaguageni linee 10 The Latin mistakenly has “ est equidistans,” thus implying that it is the diameter
munani aut duarum linearum munanieni.” The Arabie text of these definitions has been that is parallel to the posited line rather than the bisected lines. The preceding “ and having
published by L. M. L. Nix, Das fiinfte Buch der Conica des Apollonius von Perga in der their extremities on it” comes out of an Arabic addition. See Heiberg, Gr 2, p. LXXII
arabischen LIbersetzung des Thabit ibn Corrah (Leipzig, 1889), Arabic text, pp. 4 -6 . I have (I p. 6, 26) and Nix, op. cit., Arabic text, p. 5, lines 4 -5 .
also used the text in the beautiful Arabic codex of the Conics in MS Oxford, Bodleian 11 The expression “ which I have mentioned” (narravi) is taken from the Arabic text.
Library, Marsh 667, lr-6r. See Heiberg, Gr 2, p. LXXII (I p. 6, 29) and Nix, op. cit., Arabic text, p. 5, line 8. The
6 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 7

[5] And similarly again when there are two curved lines in one plane, then the Conics), let us pause to comment briefly on Gerard’s translation of
I call that part of the straight line which falls [i.e. is placed] between the two the First Definitions. The first point to notice is Gerard’s practice of
curved lines12and bisects all the lines that are drawn within each of the curved transliterating the Arabic terms which were strange to him or were strange
lines and are parallel to some line the “ transverse diameter" (diametrum to the Latin mathematical vocabulary that existed in the twelfth century.
mugeniben). And I call “ vertices of the curved lines" the two extremities The curved line that is formed on the surface of the cone by a plane
of the transverse diameter that are on the two curved lines. And I call “ erect that does not pass through the apex of the cone was rendered by Gerard
diameter" the straight line falling [i.e. placed] between the curved lines and as linea munani (and at times in the dual linee munanieni), munani being
[intersecting] a point on the transverse diameter13 and bisecting all the lines
a transliteration of the Arabic munhani. The original Greek term was
parallel to the transverse diameter when they are drawn between the two
καμπύλης. The other transliterations from the Arabic used by Gerard
curved lines so that their extremities arrive at the curved lines.14 And I call
these parallel lines “ lines of order [i.e. ordinates] to the erect diam eter."15 in this translation were diameter mugenib for a transverse diameter and
[6] And when there are two straight lines which are diameters of a curved diametri muzdaguageni for conjugate diameters, the transliterated words
line or of two curved lines and each of them bisects the lines parallel to the originating in the Arabic terms mujanib and muzdawijain respectively.
other, then I call them “ conjugate diameters" (diametros muzdaguageni). The Greek expressions were διάμετρος πλαγία and συζυγείς διάμετροι.
[7] And I call the straight line that is a diameter of a curved line or of The only substantial point concerning the interpretation found in the
two curved lines and cuts all of the parallel lines which are lines of order Latin translation as compared with the Greek text is the move made in
[i.e. ordinate] to it16 at right angles the “ axis" of the curved line or two paragraphs [4] and [5] toward the formal designation of the parallel bisected
curved lines. lines as “lines of order’’ or, as they were later to be called, “ ordinates.”
[8] And I call the two conjugate diameters each of which cuts the lines Minor variations have been remarked on in the preceding footnotes.
parallel to the other at right angles “ the two conjugate axes of the curved Now let us return to the remainder of the Arabic introduction, as para­
line or two curved lines.”
phrased from the early propositions of Book I of Apollonius’ Conics.
I have added arbitrary passage numbers and I shall give brief comments
At this point the literal translation of Apollonius’ text ends.17 Before
going on to examine Gerard’s translation of further definitions that appear to the passages.18
in the Arabic text of the Conics and were based on the early propositions And I now add a preface as an aid to the understanding of what is in
of Book I of Apollonius’ work (but which were not in the Greek text of this book.
[1] When a cone is cut by a plane that does not pass through the apex,
succeeding “ lines of order" (lineas ordinis) is simply ταταγμένως (ordinate-wise) in the then the common section is a surface which a curved line contains; and when
Greek text. the cone is cut by two planes, one of which passes through its apex and the
12 The clause “ which falls between the two curved lines” (quod cadit inter duas lineas center of its base forming a triangular section and the other does not pass
munani) is an expansion of the Greek. See Heiberg, Gr 2, p. LXXII (I p. 8, 2 -3 ) and Nix,
op. cit., Arabic text, p. 5, lines 10-11.
13 This phrase “ and a point on the transverse diameter” (et punctum super diametrum 18 MS Paris, BN lat. 9335, 85v (cf. Heiberg, Gr 2, pp. LXXVIII-LXXIX): “ Et de eo
mugenib) is not in the Greek text. Cf. Heiberg, Gr 2, p. LXXII (I p. 8, 7) and Nix, op. in cuius premissione scitur esse adiutorium ad intelligendum quod in isto existit libro est
cit., Arabic text, p. 5, line 21. The word “ intersecting” is my addition. quod narro. [1] Cum secatur piramis cum superficie plana non transeunte per punctum
14 The phrase “ and bisecting . . . transverse diameter” is much expanded beyond the capitis, tunc differentia communis est superficies quam continet linea munani, et quando
Greek text. See Heiberg, Gr 2, p. LXXII (I p. 8, 8) and Nix, op. cit., Arabic text, p. 5, secatur piramis cum duabus superficiebus planis, quarum una transit per caput eius et per
lines 21-22. The Greek merely says that the erect diameter is the straight line lying between centrum basis et separat eam secundum triangulum et altera non transit per caput ipsius,
the two curved lines which bisects all of the straight lines intercepted between the two immo secat eam cum superficie quam continet linea munani, et stat una duarum super-
curved lines and drawn parallel to some straight line. ficierum planarum ex altera secundum rectos angulos, tunc linea recta que est differentia
15 The expressions “ these parallel lines” (has lineas equidistantes) and “ erect diameter” communis duarum superficierum planarum non evacuantur (/ evacuatur) dispositionibus
{diametrum erectam) in the Greek text are “ each of the parallels” and “ diameter” . See tribus, scilicet aut quin secet unum duorum laterum trianguli et equidistet lateri alteri,
Heiberg, Gr 2, p. LXXII (I p. 8, 10) and Nix, op. cit., Arabic text, p. 5, lines 24, 25. aut quin secet unum duorum laterum trianguli et non equidistet lateri alii (/ alteri), et cum pro­
16 The phrase “ lines of order to it” is not in the Greek. See Heiberg, Gr 2, p. LXXII ducatur ipsa et latus aliud secundum rectitudinem, concurrant in parte in qua est caput
(I p. 8, 16) and Nix, op. cit., Arabic text, p. 6, line 3. piramidis, aut quin secet unum duorum laterum trianguli et non equidistet lateri alii (/
17 The First Definitions of Apollonius were not again translated into Latin from the Arabic alteri), immo concurrant aut intra piramidem aut extra eam cum protrahuntur secundum
until the translation of Christianus Ravius, Apollonii Pergaei Conicarum sectionum libri rectitudinem in parte alia in qua non est caput piramidis.” The Arabie text of all of the
V. VI. et VII (Kilonii, 1669), pp. 8-13, where they are compared with the translations of additional paraphrased definitions translated by Gerard o f Cremona and given in this note and
Federigo Commandino from the Greek as given in the edition of Claude Richard, Apollonii notes 21, 23, 26, 28, 31 accompanied the Arabic translation of the Conics made by Hilal
Pergaei Conicorum libri IV. Cum commentariis R. P. Claudii Richardi (Antwerp, 1655), ibn Abl Hilal al-Himsi found in the Marsh codex noted above in note 5, 6v-7r. It is evident
pp. 1-8. that Gerard made a very close translation of this text.
8 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 9

through the apex of the cone but rather cuts it by a plane which a curved [2] But if the straight line which is the common section of the two surfaces
line contains, and one of the two planes is at right angles to the other, then of the [above-noted cutting] planes is parallel to one side of the triangle, then
the straight line which is the common section of the two planes is in one of the surface in which the cone is cut and which the curved line contains is
the following dispositions: (1) either it cuts one of the two sides of the triangle called a parabola (sectio mukefi).
and is parallel to the other side, or (2) it cuts one of the two sides and is not [3] And if it [the aforementioned common section] is not parallel to the
parallel to the other—and when it and the other side are produced directly side of the triangle but rather meets it (when both are produced directly) in
they meet in the direction in which lies the apex of the cone, or (3) it cuts the direction in which lies the apex of the cone, then the surface in which
one of the two sides of the triangle, is not parallel to the other side but they the cone is cut and which the curved line contains is called a hyperbola
meet either inside or outside of the cone (when they are produced directly) (sectio addita).
in the direction that is not that of the apex. [4] And if it is not parallel to the side of the triangle but rather meets it
[when both are produced directly] in the other direction (that is, that in which
Of course, the “ book” referred to in the preliminary sentence is the the apex of the cone does not lie), then the surface in which the cone is cut,
text itself of Apollonius’ Conics. Now let us examine passage [1], which if not a circle, is called an ellipse (sectio diminuta).
diverges from the Greek text of Apollonius. The paraphraser in this intro­
It will be immediately obvious to the reader that, as I have already
duction attempts as briefly as possible to set out the constructional proce­
suggested, only the constructional instructions for the three sections have
dures for his following definitions of the parabola, hyperbola, and ellipse.
been taken (in a loose fashion) from Apollonius’ propositions; the actual
In a sense, he is merely paraphrasing the conditional clauses of Apollonius’
properties of the sections in terms of the latus rectum and the proofs of
Propositions 1.11, 1.12 and 1.13. It is true that in doing this he vaguely
those propositions have been left out, as is evident from Apollonius’
implies at least two of the earlier propositions of Apollonius, the first
propositions, whose enunciations I now quote:22
of which is Proposition 1.2, which runs:19
11. If a cone is cut by a plane through its axis and is also cut by another
If on either one of the surfaces that are mutually placed with respect to
plane cutting the base of the cone in a straight line perpendicular to the base
the apex two points are taken, and the straight line joining the points does
of the axial triangle, and if further the diameter of the section is parallel to
not verge to the apex, then it will fall within the surface, but if produced
one side of the axial triangle, then any straight line which is drawn from the
it will fall outside.
section of the cone to its diameter parallel to the common section of the
It is this proposition, unspecified at this point of the introduction, that cutting plane and of the cone’s base will equal in square the rectangle con­
lurks in the background to guarantee that the line in the section specified tained by the straight line cut off by it on the diameter beginning from the
in passage [1] is indeed a curved line. The second proposition implied section’s vertex and by another certain straight line which has the ratio to
in the course of this passage is 1.3: “ If a cone is cut by a plane through the straight line between the angle of the cone and the vertex of the section
that the square on the base of the axial triangle has to the rectangle con­
the apex, the section is a triangle.” 20 The paraphrase of this proposition
tained by the remaining two sides of the triangle. And let such a section be
is however more restricted since it specifies that the plane passing through called a parabola. . . .
the apex also passes through the center of the base of the cone. Missing 12. If a cone is cut by a plane through its axis and is also cut by another
from the paraphrase (but of course to follow in the Arabic text) is the plane cutting the base of the cone in a straight line perpendicular to the base
elegant march of Propositions 1.1- 1.10 with their proofs which Apollonius of the axial triangle, and if the diameter of the section produced meets one
gives before presenting the key Propositions 1.11- 1 .13. These latter prop­ side of the axial triangle beyond the apex of the cone, then any straight
ositions the paraphraser then palely reflects in his passages [2]-[4].21 line which is drawn from the section to its diameter parallel to the common
section of the cutting plane and of the cone’s base will equal in square some
19 Heiberg, Gr 1, p. 10. See the translation of the Conics by R. C. Taliaferro in Great area applied to a straight line to which the straight line added along the
Books of the Western World, Vol. 11 (Chicago, 1952), p. 605, which I have used (but changed diameter of the section and subtending the exterior angle of the triangle has
as I have seen fit) in this and the succeeding quotations from the Conics. the ratio that the square on the straight line drawn from the cone’s apex to
20 Heiberg, Gr 1, p. 12; Taliaferro, op. cit., p. 606. the triangle’s base parallel to the section’s diameter has to the rectangle
21 BN lat. 9335, 85v-86r (cf. Heiberg, Gr 2, p. LXXIX): “ [2] Quod si linea recta que est contained by the segments of the base which this straight line makes when
differentia communis duarum superficierum planarum equidistat lateri trianguli, tunc super­ drawn, this area having as breadth the straight line cut off on the diameter
ficies super quam secatur piramis et quam continet linea munani nominatur sectio mukefi.
beginning from the section’s vertex by this straight line from the section to
[3] Et si non equidistat lateri trianguli, immo concurrit ei quando protrahuntur secundum
rectitudinem in parte in qua est caput piramidis, tunc superficies super quam secatur piramis
the diameter and exceeding by a figure similar to and similarly situated to
et quam continet linea munani nominatur sectio addita. [4] Et si non equidistat lateri trianguli, the rectangle contained by the straight line subtending the exterior angle
immo occurrit ei in parte alia in qua non est caput piramidis, tunc superficies super quam
secatur piramis, si non est circulus, nominatur sectio diminuta.” 22 Heiberg, Gr 1, pp. 36-48: Taliaferro, op. cit., pp. 615-18.
10 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 11

of the triangle and by the parameter. And let such a section be called a [5] When there are two hyperbolas with a common diameter and the gib­
hyperbola. . . . bosity of one of them follows the gibbosity of the other, then they are called
13. If a cone is cut by a plane through its axis and it is also cut by another two “ opposite sections’’.
plane that on the one hand meets both the sides of the axial triangle and on
the other is extended neither parallel to the base nor subcontrariwise, and Again this is but a pale reflection of Apollonius (Proposition I.14):24
if the plane in which the base of the cone lies and the cutting plane meet If the surfaces that are mutually placed with respect to the apex are cut
in a straight line perpendicular either to the base of the axial triangle or to by a plane not through the apex, the section on each of the two surfaces
it produced, then any straight line which is drawn from the section of the will be that which is called the hyperbola; and the diameter of the two sections
cone to the diameter of the section parallel to the common section of the will be the same straight line; and the parameters drawn to the diameter
planes will equal in square some area applied to a certain straight line to parallel to the straight line in the cone’s base are equal; and the transverse
which the diameter of the section has the ratio that the square on the straight side of the figure, that between the vertices of the sections, is common.
line drawn from the cone’s apex to the triangle’s base parallel to the section’s And let such sections be called opposite.
diameter has to the rectangle contained by the intercepts of this straight line
(on the base) from the sides of the triangle, an area having as breadth the Note that rather than following Apollonius’ statement that the param­
straight line cut off from the diameter beginning from the section’s vertex eters of the two branches of the hyperbola are equal the paraphrase merely
by this straight line from the section to the diameter, and deficient by a says that the gibbosity (i.e. convexity or curvature) of one is the same as
figure similar to and similarly situated to the rectangle contained by the that of the other, without specifying the measure of that gibbosity. Also
diameter and parameter. And let such a section be called an ellipse. . . . obscured by the paraphraser is the fact that the branches are formed by
By translating only the introductory definitions and not the actual prop­ passing a single plane through the opposite surfaces of a double cone
ositions of Apollonius, Gerard has failed to present the parabola in terms and that indeed the two branches constitute the curve. This consideration
of the equality of the square of an ordinate and the rectangle composed of the two branches as forming one curve (with one center) is usually
of the latus rectum and the ordinate’s abscissa. It is, of course, this key regarded as having originated with Apollonius; at any rate it seems likely
property of the parabola that is fundamental to the proof of Proposition 1 that he was the first to investigate their properties completely.25 Our
in the De speculis comburentibus, as we shall see. Its inclusion from paraphraser follows this definition of opposite branches with a definition
Proposition 1.11 of the Conics would have greatly aided the reader. The of the center of a hyperbola and of an ellipse:26
reader of the Arabic text of the introductory definitions would have had [6] And between the two opposite sections there is a point through which
no such difficulty since the succeeding text of Book I of the Conics in­ all the lines that pass are diameters of the two opposite sections, and this
cluded the actual Propositions 1.11-1.13. Hence he would have been point is called the “ center of the two sections.’’ And within the ellipse there
satisfied at this point of the introduction with these preliminary construc­ is a point through which all the lines that pass are diameters to it, and this
tional definitions. point is the “ center of the section’’ [i.e. ellipse].
A word should be added concerning the Latin terminology employed The introduction has thus given the substance of Apollonius, Def. 1
by Gerard for the three conic sections: sectio mukefi, sectio addita and of the Second Definitions: “ 1. Let the midpoint of the diameter of both
sectio diminuta, which rendered respectively qatr mukafi, qafi z a id and the hyperbola and the ellipse be called the center of the section.” 27
qat' naqis. It is evident that the literal meanings of the Arabic terms, namely Following this treatment of the centers of sections the introduction then
sections that are called “ equivalent, augmented and diminished,’’ reflect passes on to diameters:28
the literal meanings of the Greek terms παραβολή, υπερβολή and ελλειψις
as employed by Apollonius in his use of the technique of the application 24 Heiberg, Gr 1, p. 52; Taliaferro, op. cit., p. 620.
of areas for conic sections. It is obvious that mukafi puzzled Gerard of 25 T. L. Heath, Apollonius of Perga. Treatise on Conic Sections (Cambridge, 1896), p.
Cremona, for he settled for the transcription mukefi. As we shall see in lxxxiv.
Chapter 4 below, this transcription was often written as mukesi, the / 26 BN lat. 9335, 86r (cf. Heiberg, Gr 2, p. LXXIX): “ [6] Et inter duas sectiones oppositas
est punctum per quod omnes linee que transeunt sunt diametri duarum sectionum oppo­
being misread as s. sitarum. Et hoc punctum nominatur centrum duarum sectionum. Et intra sectionem dimi­
Passing on from the three basic sections, the Arabic introduction then nutam est punctum per quod omnes linee que transeunt sunt ei diametri. Et hoc punctum
defines opposite branches:23 est centrum sectionis.”
27 Heiberg, Gr 1, p. 66; Taliaferro, op. cit., p. 626.
23 BN lat. 9335, 86r (cf. Heiberg, Gr 2, p. LXX1X): "[5] Et quando sunt due sectiones 28 BN lat. 9335, 86r (cf. Heiberg, Gr 2, pp. LXXIX-LXXX): ” [7] Et cum in sectione
addite quibus est diameter communis, et gibbositas unius earum sequitur gibbositatem al­ diminuta protrahuntur diametri, tunc ille ex illis diametris quarum extremitates perveniunt
terius, tunc ipse nominantur due sectiones opposite.’’ ad circumferentiam sectionis et non pertranseunt eam nec ab ea abbreviantur nominantur
12 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 13

[7] And when in an ellipse diameters are projected, those whose extremities Definitions of Apollonius to the three conic sections. Once more we find
are on the circumference of the section, not extending beyond it and not a promise of later treatment that remains unfulfilled in the Latin text.
falling short of it, are called “ transverse diameters of the ellipse." Now it should be clear that this introductory fragment would hardly
[8] And the diameter whose beginning is from a point of the circumference give the neophyte Latin reader much precise information that would be
of the section and whose other extremity falls short of the circumference of significant use to him when he approached the very specific treatment
of the section or goes beyond it is simply called the "diameter."
of the parabola that followed in theDe speculis comburentibus of Alhazen,
[9] But the diameter which is called "second" is only in the two opposite
sections and it will pass through the center of both of them; and I shall
for, as we shall see, the fragment threw no light on those propositions
describe it at the end of the sixteenth proposition of this tract [On Conics], of Apollonius appealed to in Alhazen’s crucial first proposition. In fact,
I shall limit my discussion of the tract to that first proposition since it
Passages [7] and [8] distinguish between the expressions "transverse contains the basic knowledge of the parabola that is repeated throughout
diameter” and "diameter.” The definition of the transverse diameter of the tract. However, it is worth noting that in his preface Alhazen refers
the ellipse was drawn by the paraphraser from a statement at the end of to the efforts of earlier investigators (“ Archimedes, Anthemius and
Proposition 1.13 where ED , the diameter from which the latus rectum others” ) to find mirrors which would reflect rays "to a point so that
is erected at a vertex, is so labeled as “ transverse” (π λ α γ ία ).29 Passage combustion would be stronger.” 32 It is then remarked that they discovered
[9] concerns the "second diameter” and the paraphraser specifically says that the reflection of the rays to a single point would be achieved from
that he (that is, Apollonius) will describe it at the end of Proposition 16, the surface of a concave paraboloid. But, according to Alhazen, the early
and indeed it is so defined as the third definition among Apollonius’ Second investigators did not sufficiently demonstrate the convergence property
Definitions.30 But Gerard translated no more of the actual text of Apol­ of the paraboloidal surface. Hence Alhazen will undertake this demon­
lonius and so the paraphraser’s promise of later treatment remains unful­ stration. Let us now turn to the first proposition.33
filled in Gerard’s translation.
In the final passage of the introduction the paraphraser shifts from
32 BN lat. 9335, 86r (cf. the text of Heiberg and Wiedemann cited in note 2, p. 219): "Et
diameters to axes of the conic sections:31 ex eis fuerunt quidam qui assumpserunt specula plurima sperica quorum radii converte­
rentur ad punctum unum ut combustio fortior esset, et illi qui invenerunt specula ista famosi
[10] The parabola has only one axis, while the ellipse has two axes within fuerunt, sicut Archimenides et Anthinus (/ Anthemius) et alii ab istis duobus. Deinde accidit
it. But the hyperbola has one transverse axis and it is that which cuts the eis cogitatio in proprietatibus figurarum ex quibus convertitur radius. Aspexerunt ergo in
lines of order [i.e. ordinates] at right angles whether it is within the section proprietatibus sectionum piramidum et invenerunt radios qui cadunt super omnem plani­
or without, or whether it is partly within and partly without, and it has another ciem superficiei concave corporis mukefi converti ad punctum unum eundum. , . . Verum-
"erect axis" (axis erectus), and I shall demonstrate this in what follows. tamen ipsi non exposuerunt demonstrationem super hanc intentionem neque viam qua in­
And conjugate axes occur only in opposite sections [i.e., of the hyperbola] venerunt expositione sufficiente. Sed propter illud quod est in hoc de commoditatibus magnis
and in ellipses. However a line is called an "erect line” on which lines can et utilitatibus communibus vidi ut exponerem illud et explanarem quatinus contineret scien­
be drawn ordinate-wise to the diameter. tiam eius cuius voluntas est in cognitione veritatum et sciret illud cuius sollicitudo est in
velocitatibus rerum. Declaravi ergo in hoc tractatu et abbreviavi demonstrationem super
This is a descriptive passage composed by the paraphraser in order to scientia veritatis eius. . . . ”
apply the general definitions of axis and conjugate axis found in the First 33 Again I have translated the text in Paris, BN lat. 9335, 86r-87r, which differs only
slightly from Heiberg’s text in the article of his and Wiedemann’s, op. cit. in note 2, pp.
221-24. I have followed Heiberg’s practice of capitalizing the letters marking the points
diametri mugenibi sectionis diminute. [8] Et que ex eis est cuius principium est ex puncto and magnitudes. “ In omni sectione mukefi cuius protrahitur sagitta et separatur ex ex­
circumferende sectionis et eius altera extremitas abreviata (/) est a circumferentia sectionis tremitate sagitte quantum est quarta lateris eius recti, omnis linea protracta equidistans
aut pertransit eam nominatur diameter absolute. [9] Diameter vero que nominatur secunda sagitte et perveniens ad sectionem et alia conversa ad punctum quod separat quartam con­
non est nisi in duabus sectionibus oppositis et transit per centrum ambarum, et narrabo tinent cum linea contingente sectionem super illud punctum (mg.\ scilicet, extremitatis) duos
illud in fine sextedecime figure huius tractatus.” angulos equales.
29 Heiberg, Gr 1, p. 52. [Fig. 1.1], “ Verbi gratia, sit sectio ABG sectio mukefi, et sit eius sagitta AD et latus
30 Ibid., p. 66. eius rectum L. Et secabo ex AD lineam AE equalem quarte linee L, et producam lineam
31 BN lat. 9335, 86r (cf. Heiberg, Gr 2, p. LXXX): "[10] Et sectioni quidem mukefi TB equidistantem linee DA, et continuabo BE, et protraham KBH contingentem. Dico
non est nisi unus axis; sectioni vero diminute sunt duo axes intra ipsam; verum addite ergo quod angulus TBK est equalis angulo EBH.
est axis unus mugenib, et est ille qui secat lineas ordinis secundum rectos angulos, sive "Sit itaque in primis angulus BEH acutus. Ergo secundum semitam resolutionis ponam
ipse sit intra sectionem sive extra ipsam, sive pars eius intra sectionem et pars eius extra ut angulus TBK sit equalis angulo EBH. Et quoniam linea TB est equidistans linee DA,
ipsam, et est ei axis alter erectus, et ostendam illud in sequentibus. Et non cadunt axes erit angulus TBK equalis angulo BHE. Sed angulus TBK est equalis angulo HBE per posi­
muzdeguege nisi in sectionibus oppositis et in diminutis. Tamen et nominatur linea erecta tionem; ergo angulus EBH est equalis angulo BHE. Ergo linea BE est equalis linee EH\
linea super quam possunt linee protracte ad diametrum secundum ordinem.” ergo quadratum BE est equale quadrato EH. Et protraham BZ perpendicularem super sagit-
14 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 15

tarn. Ergo duo quadrata EZ. ZB sunt equalia quadrato EH. Sed quadratum BZ est equate
[1] In every parabola whose axis is drawn and from the extremity of which
multiplicationi AZ in L, que est linea recta (7 latus rectum), sicut ostendit Apollonius bonus an abscissa equal to one quarter of the parabola's latus rectum is taken,
in libro de piramidibus. Ergo quadratum EZ et multiplicatio ZA in L sunt equalia quadrato any line drawn parallel to the axis and meeting the section and another line
EH. Sed EA est quarta L. Ergo multiplicatio ZA in AE quater est multiplicatio ZA in L. [thence] reflected to the point which cuts off the quarter [of the latus rectum]
Ergo multiplicatio ZA in AE quater et quadratum EZ sunt equalia quadrato EH. Ergo AH makes two equal angles with the line tangent to the section at that point
est equalis ZA. Sed est ita, quoniam BH est contingens et BZ est secundum ordinem. [where the parallel line meets the section].
“ Et secundum semitam compositionis ponam res omnes secundum suam dispositionem, For example [Fig. 1.1], let section ABG be a parabola with axis AD and
et dico quod angulus TBK est equalis angulo EBH; cuius demonstratio est. Quia protraham latus rectum L, and I shall cut from AD line AE equal to lA L, and I shall
BZ secundum ordinem. Et quoniam BH est contingens sectionem et BZ secundum ordinem,
draw line TB parallel to line DA and I shall join BE, and I shall draw tangent
erit AZ equalis linee AH. Ergo multiplicatio EA in AZ quater cum quadrato EZ est equalis
quadrato EH. Sed EA est quarta L. Ergo multiplicatio EA in ZA quater est multiplicatio
KBH. I say, therefore, that ATBK — AEBH.
And so in the first place let A BEH be acute. Then following the method
L in ZA. Ergo multiplicatio L in ZA cum quadrato EZ est equalis quadrato EH. Verum
multiplicatio L in ZA est quadratum BZ quoniam BZ est secundum ordinem. Ergo quadratum of analysis, I shall posit that ATBK = A EBH. And since line TB is parallel
BZ et quadratum EZ sunt equalia quadrato EH. Sed duo quadrata BZ, EZ sunt quadratum to line DA, ATBK = ABHE. But ATBK = AHBE by assumption. There­
EB; ergo quadratum EB est equale quadrato EH. Ergo EB est equalis EH; ergo angulus fore A EBH = ABHE. Therefore, line BE = line EH. Therefore, BE2 = EH2.
EBH est equalis angulo EHB. Et rursus TB equidistat DA; ergo angulus TBK est equalis And I shall draw BZ perpendicular to the axis. Therefore, EZ2 + ZB2 = EH2.
angulo EBH (! EHB). Ergo angulus EBH est equalis angulo TBK. Et similiter omnis linea But BZ2 = A Z E , L being the latus rectum, as the good Apollonius demon­
que protrahitur equidistans diametro, et convertitur alia ad punctum E et continet cum strates in his Book on Cones [Prop. 1.11]. Therefore, EZ2 +ZA L = EH2.
EA angulum acutum. Et illud est quod declarare voluimus. But EA — l/4 L. Therefore, 4Z A A E = Z A L. Therefore 4Z A A E + EZ2
[Fig. 1.2] “ Et firmemus quod diximus secundum dispositionem suam, et sit linea BE = EH2. Therefore, AH = ZA. But this is so [as Apollonius proves in Prop.
continens cum linea EA angulum rectum. Dico ergo quod angulus TBK est equalis angulo
1.35] since BH is a tangent and BZ is drawn ordinate-wise.
EBH. Per resolutionem ergo ponam duos angulos equales. Et quoniam linea TB equidistat
linee AD, erit angulus TBK equalis angulo EHB. Sed angulus TBK per positionem est
AM; ergo ZA est equalis AH. Sed est ita, quoniam BH est contingens sectionem et BZ
equalis angulo EBH. Ergo angulus EBH est equalis EHB. Ergo linea BE est equalis linee
EH. Ergo quadratum EB est equale quadrato EH. Sed quadratum EB est equale multiplica­ est secundum ordinem.
“ Et secundum partem compositionis ponam res omnes secundum suam dispositionem.
tioni EA in L, quod est latus rectum, quoniam BE est secundum rectos angulos. Ergo
Dico ergo quod angulus TBK est equalis angulo EBH; cuius demonstratio est ista. Protraham
multiplicatio EA in L est equalis quadrato EH. Sed multiplicatio EA in L est quarta quadrati
BZ secundum ordinem. Et quoniam BH contingit sectionem et BZ est secundum ordinem,
E, quoniam EA est quarta E; ergo quadratum EH est quarta quadrati E. Ergo linea EH
erit linea ZA equalis linee AH. Et ponam AM equalem AE. Remanet ergo ZE equalis
est medietas linee E. At EA est quarta E; ergo AH est quarta E. Ergo linea EA est equalis
MH. Ergo ZM est equalis EH. Ergo quadratum ZM est equale quadrato EH. Sed multipli­
linee AH. Sed est ita, quoniam BH est contingens et BE est secundum ordinem.
catio ZA in AE quater cum quadrato ZE est equalis quadrato ZM. Ergo multiplicatio ZA
“ Et per compositionem ponam res omnes secundum dispositionem suam. Dico ergo quod
in AE quater cum quadrato ZE est equalis quadrato EH. Sed multiplicatio ZA in AE quater
angulus TBK est equalis angulo EBH; cuius demonstratio est hec. Quoniam linea BH est
est multiplicatio ZA in E, quoniam EA est quarta E. Ergo multiplicatio ZA in E cum quadrato
contingens sectionem et BE est secundum ordinem, tunc linea AE est equalis linee AH.
ZE est equalis quadrato EH. Sed multiplicatio ZA in E est quadratum BZ, quoniam BZ
Sed EA est quarta E; ergo EH est medietas E. Ergo quadratum EH est quarta quadrati
est secundum ordinem. Ergo quadratum BZ et quadratum ZE sunt equalia quadrato EH.
E. Sed multiplicatio EA in E est quarta quadrati E, quoniam EA est quarta E. Ergo multi­
Sed quadratum BZ et quadratum ZE sunt quadratum BE. Ergo quadratum BE est equale
plicatio EA in E est equalis quadrato EH. Sed multiplicatio EA in E est equalis quadrato
quadrato EH. Ergo BE est equalis EH. Ergo angulus EBH est equalis angulo EHB. Sed
EB, quoniam EB est secundum ordinem. Ergo quadratum EB est equale quadrato EH.
angulus EHB est equalis angulo TBK, quoniam linea TB equidistat linee DH. Ergo angulus
Ergo linea BE est equalis linee EH. Ergo angulus EBH est equalis angulo EHB. Et quoniam
EBH est equalis angulo TBK. Et similiter contingit in omni linea que producitur in sectione
linea TB equidistat linee DH, erit angulus TBK equalis angulo EHB. Et iam fuit angulus
EBH equalis angulo EHB; ergo angulus TBK est equalis angulo EBH. Et illud est quod et continet cum linea EH ab eo quod sequitur caput eius angulum expansum.
“ Itaque {ed ., Nam MS) omnis linea que protrahitur in sectione equidistans sagitte eius
demonstrare voluimus.
et convertitur ad punctum E continet cum linea contingente super illud punctum duos angulos
[Fig. 1.3] “ Et figamus res quas diximus secundum dispositionem suam, et sit angulus
equales. Et illud est quod declarare voluimus."
BEH expansus. Dico ergo quod angulus TBK est equalis angulo EBH. Secundum partem
We should note in passing that the Winter and 'Arafat translation cited in note 2 incorrectly
ergo resolutionis ponam ut illud sit ita. Et quoniam linea TB equidistat linee DH, erit
adds (p. 28) the first part of the specification of the proposition to the end of the enunciation.
angulus TBK equalis angulo EHB. Sed angulus TBK per positionem est equalis angulo
Furthermore, the restatement of the conclusion of the enunciation in the specification (“ I
EBH; ergo angulus EBH est equalis angulo EHB. Ergo linea EB est equalis linee EH; ergo
say, therefore, that ATBK = A EBH [AHBC in Winter’s text]” ) is incorrectly rendered
quadratumBE est equale quadrato EH. Et protraham BZ secundum ordinem; ergo quadratum
as “ Suppose that ATBK = AHBC." I suspect that the translators inadvertently anticipated
BZ et quadratum ZE sunt equalia quadrato EH. Sed quadratum BZ est equale multiplicationi
that supposition when it was made as the starting point for the proof by analysis. At any
AZ in E. Ergo multiplicatio AZ in E cum quadrato ZE est equalis quadrato EH. Sed EA
rate it is correctly given in both the Arabic text and Gerard’s translation. In Diodes, On
est equalis quarte E; ergo multiplicatio ZA in AE est equalis quarte multiplicationis ZA
Burning Mirrors. The Arabic Translation of the Lost Greek Original, ed. and transl. of G. J.
in E. Ergo multiplicatio ZA in AE quater cum quadrato ZE est equalis quadrato EH. Et
Toomer, (Heidelberg, New York, 1976), pp. 203-204, the editor has epitomized this proof.
ponam AM equalem AE; ergo multiplicatio ZA in AE quater cum quadrato EZ est equalis
Note that earlier, on p. 22, Toomer has tentatively argued that Alhazen’s proof of this
quadrato ZM. Ergo quadratum ZM est equale quadrato EH; ergo ZM est equalis EH.
Proposition 1 “ was inspired by D iodes.”
Prohiciamus ergo EM communem; ergo remanebit ZE equalis MH. Sed EA est equalis
16 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 17

And following the method of synthesis, I shall pose all things [the same] And following the method of synthesis I shall pose all things [the same]
regarding its disposition, and I say that LTBK = LEBH, whose demon­ regarding its disposition. I say, therefore, that LTBK = LEBH, of which the
stration follows. demonstration is this.
For I shall draw BZ ordinate-wise; and since BH is a tangent to the section I shall draw BZ ordinate-wise. And since BH is tangent to the section and
and BZ is drawn ordinate-wise, AZ = AH. Therefore [by Euclid, II.8] 4EA AZ BZ is drawn ordinate-wise, line ZA = line AH. And I shall posit AM = AE.
+ EZ2 = EH2. But EA = A L. Therefore, 4EAZA = L Z A . Therefore, Therefore, ZE = MH. Therefore, ZM — EH. Therefore, Z M 2 = EH2. But
L-ZA + EZ2 — EH2. But L ZA = BZ2 since BZ is drawn ordinate-wise. There­ 4ZA AE + ZE2 = Z M 2. Therefore, 4ZA ZE + ZE2 = EH2. But 4ZA AE
fore, BZ2 + EZ2 = EH2. But BZ2 + EZ2 = EB2. Therefore, EB2 = EH2. There­ = ZA L since EA = 14 L. Therefore, ZA L + ZE2 = EH2. But ZA L —BZ2
fore EB =EH, and therefore LEBH = LEHB. And further, TB is parallel since BZ is drawn ordinate-wise; therefore, BZ2 + ZE2 = EH2. But BZ2 + ZE2
to DA. Therefore, LTBK — LEHB. Therefore, LEBH = LTBK. And [it = BE2. Therefore BE2 = EH2. Therefore, BE = EH. Therefore, LEBH
is] similarly [proved for] every line which is drawn parallel to the diameter = LEHB .But LEHB = L TBK since line TB is parallel to line DH. Therefore,
and [for] another line [that] is reflected [therefrom] to point E and forms with LEBH = LTBK. And it happens in the same way for every line which is
EA an acute angle. Q.E.D. drawn in the section and whose reflection from its extremity meets line EH
And let us aver everything which we have said [before] regarding its disposi­ in an obtuse angle.
tion. And let line BE form a right angle with line EA [Fig. 1.2]. I say, And so every line projected in the [parabolic] section parallel to its axis
therefore, that LTBK = LEBH. and reflected to point E forms with the tangent at the point [of reflection
Following the method of analysis, I shall posit that the two angles are on the section] two equal angles. Q.E.D.
equal. And since line TB is parallel to line AD, then LTBK = LEHB. But
4 TBK was posited to be equal to LEBH; therefore, LEBH = LEHB. There­ It should be immediately obvious that Alhazen wants to show that when
fore, line BE = line EH. Therefore EB2 = EH2. But EB2 = EA L, L being he assumes that all reflected lines converge at a point in the axis 14-the-
the latus rectum, since BE is at right angles [to HD]. Therefore, EA L = EH2. latus-rectum distant from the vertex, the angle of incidence of each line
But EA ·L = 14 L 2 since EA = !4 L. Therefore, EH2 = !4 L 2. Therefore equals the angle of reflection, a universal condition for physical reflection.
EH = Vi L. But EA — A L. Therefore AH = A L. Therefore, line EA — line The proposition is established for all possible cases, namely, when the
AH. But this is so, since BH is a tangent and BE is drawn ordinate-wise. reflected line BE forms with the axis HD an acute angle, or a right angle,
And following the method of synthesis, I shall pose all things [the same] or an obtuse angle. Each case is proved in the same way, first by analysis
regarding its disposition. I say, therefore, that LTBK = LEBH, whose
and then by synthesis (the Latin terms are resolutio and compositio).
demonstration is this.
The analysis in each case assumes the equality of the angles of incidence
Since line BH is a tangent to the section and BE is drawn ordinate-wise,
then line AE = line AH. But EA = A L ; therefore, EH = A L. Therefore, and reflection and then deduces the tangential property enunciated and
EH2 = lA L 2. But EA ■L = !4 L 2 since EA = lA L. Therefore, EA ■L = EH2. proved in Proposition 1.35 of Apollonius’ Conics, namely that AH = ZA
But EA L = EB2 since EB is drawn ordinate-wise. Therefore, EB2 = EH2. where BH is a tangent andBZ is drawn ordinate-wise. The synthesis in each
Therefore, line BE = line EH. Therefore, LEBH = LEHB. And since line case merely reverses the steps of the analysis, proceeding from the tangen­
TB is parallel to line DH, then LTBK = LEHB. And LEBH = LEHB; tial property to the equality of angles. The proof then gives the reader
therefore, LTBK = LEBH. Q.E.D. some idea of two important propositions of Apollonius: 1.11 and 1.35.
Now let us affix the things which we have said [before] regarding its disposi­ However, without Apollonius’ text of Proposition 1.11 (see above note 22)
tion. And let L = BEH be obtuse [Fig. 1.3].34 I say, therefore, that LTBK and its proof, how the section was formed from a given latus rectum would
= LEBH.
have been somewhat of a mystery to the reader, as indeed the author of the
Therefore, following the method of analysis, I shall posit that it be so
Speculi almukefi compositio was later to confess (see Chapter 4 below).
[namely, that LTBK = LEBH]. And since line TB is parallel to line DH,
From the paraphrased passage [2] of the Arabic introduction to the Conics
then LTBK = LEHB. But LTBK was posited to be equal to LEBH; there­
fore, LEBH = LEHB. Therefore, line EB = line EH; therefore BE2 = EH2. the reader would only know that the parabola was formed by cutting the
And I shall draw BZ ordinate-wise. Therefore, BZ2 + ZE2 = EH2. But BZ2 cone with a plane perpendicular to the axial triangle and parallel to one
= A Z L . Therefore, AZ-L -t- ZE2 = EH2. But EA — A L. Therefore, ZA AE of its sides. Needless to say, the full text of Proposition 1.1135 would have
= A ZA L. Therefore, 4ZA AE + ZE2 = EH2. And I shall posit AM = AE. shown him that when parabolic section ABG was formed in the cutting
Therefore, 4ZA AE + EZ2 = ZM 2 = EH2. Therefore, ZM = EH. Therefore, of cone EFC [Fig. 1.4] the latus rectum L was such a length that EF2I
we subtract the common line EM and therefore ZE will remain equal to EC CF = LIAC, from which Apollonius showed that BZ2 = AZ L, that
MH. But EA = AM; therefore, ZA = AH. But such is [in fact] the case since is, the fundamental property of the parabola employed in Alhazen’s proof.
BH is a tangent to the section and BZ is drawn ordinate-wise.
34 Note that the Latin word used by Gerard for “ obtuse” is expansus. 35 Heiberg, Gr 1, pp. 36-42.
18 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 19

Of course, the reader could at least have reasoned backward from Alhazen’s tiva was cited as early as the time of Jordanus de Nemore’s De triangulis.39
statement of the equation of the parabola that the latus rectum was the There are only two propositions of the Perspectiva in which Alhazen
quotient arising from the division of the square of any ordinate of the acknowledges his use of Apollonius’ Conics, namely V.33 and V.34, and
parabola by the ordinate’s abscissa on the axis. Incidentally, we should I shall confine myself to an analysis of these propositions while noting
note that Gerard of Cremona’s rendering of this parameter as “latus that there are other propositions of the Perspectiva where citation of
rectum” is the first use of that expression, which was to become the con­ elementary propositions of the Conics (e.g. Propositions 1.3, 1.4 and 1.5)
ventional one in the Latin West. As for Proposition 1.35, while the reader might have been appropriate,40 but where the reader would have been
would not have known its proposition number, its proof, or even that it was ignorant of the possibility that the Conics was involved.
in Apollonius’ work, he would have been able to deduce its substance. In speaking of Propositions V.33 and V.34, I use the proposition num­
It runs:36 bers added by F. Risner in his edition of 1572, the original numbers used
35. If a straight line touches a parabola, meeting the diameter outside in the pristine Latin version being difficult to determine.41 I have also
of the section, the straight line drawn from the point of contact ordinate-wise
to the diameter will cut off on the diameter beginning from the vertex of deserves a thorough investigation and we should be cautious in identifying the translator
the section a straight line equal to the straight line between the vertex and until that investigation has been made.
the tangent. . . . 39 See below, note 55.
40 The possible dependence on Apollonius is suggested by Risner in bracketed additions
to the following propositions of Alhazen’s Perspectiva: IV.41 (ed. cit., p. 122; citing Conics
No further knowledge of conic sections was to be deduced from the re­
1.4); IV.43 (p. 122— Conics 1.4); V.50 (p. 156— Conics 1.4 and 1.3); V.52 (p. 157— Conics
maining propositions of De speculis comburentibus, which elaborate on 1.4 and 1.3); V.53 (p. 158— Conics 1.4); V.54 (p. 158— Conics 1.4); V.55 (p. 159— Conics
the formation of a paraboloidal body by the rotation of a parabolic section 1.4 twice and 1.3); V.56 (p. 160— Conics 1.4 and I.Def. 1 twice); V.57 (p. 160— Conics
and the construction of paraboloidal mirrors. Needless to say, the focal 1.4) ; V.59 (p. 161— Conics 1.4); V.100 (p. 186— Conics 1.4); VI.30 (p. 210— Conics 1.4);
convergence demonstrated in the first proposition for a parabola holds VI.31 (p. 211— Conics 1.5); VI.32 (pp. 211-212— Conics 1.4 and 1.5 twice); VI.56(p. 230—
for each and every section formed by passing each and every plane through Conics 1.5). The reader will recall that Conics 1.3 indicates that a section through the apex of a
cone is a triangle, while 1.4 tells us that sections through either part of a double cone that
the surface of the paraboloid and its axis and thus for the whole parab­ are parallel to the circle on which the generator moves are circles. 1.5 defines the conditions
oloidal surface. Alhazen’s concentration on the focal property of the necessary for a sub-contrary circular section. For the first definition of the First Definitions
parabola was a significant beginning, in its Latin form, of a trend that of the Conics, see note 5 above, passage [1].
would make the focus an important element in the generation of a parabola 41 I have read the following manuscript copies of Propositions V.33-V.34; 1. Brugge,
Stadsbibl. 512, 65r-66r, 13c; 2. Edinburgh, Royal Observatory 9-11-3 (20), 112r-113v, 1269;
without reference to a cone (see below, Chap. 5, Johannes Fusoris’
3. London, Royal College of Physicians 383 (36), 70r-71r, 13c; 4. London, British Library
Libellus de seccione mukefi), a trend that would culminate in the complete Royal 12.G.VII, 62r-63r, 14c; 5. Sloane 306, 107r-108v, 14c; 6. Erfurt, Wissensch. Bibl.
dependence on the focus-directrix descriptions of conics. Amplon. F.392, 83r-84r, 13c; 7. Vatican, Bibl. Apost. Pal. lat. 1355, 88r-89r; 8. Cambridge,
The translation of a second work of Alhazen produced for the Latin Peterhouse Coll. Libr. 209, 62r-v, 14c; 9. Trinity Coll. Libr. 0.5.30(1311), 98r-100r, 13c;
reader a further meager harvest of knowledge of conic sections, namely 10. Saint-Omer, Bibl. Munic. 605, 89r-90r, 14c; 11. Paris, Bibl. Nat. lat. 7247, 119v-121r,
his De aspectibus (or, as it was often called in the Middle Ages and as 14c; 12. BN lat. 16199, 144v-146v, 16c; 13. Munich, Bay. Staatsbibl. 10269, 94r-95v, 14c;
14. Oxford, Bodl. Libr. Corp. Christ. Coll. 150, 62r-63r, 13c; 15. Florence, Bibl. Naz.
I shall designate it here, Perspectiva).37 We cannot be sure that Gerard II.III.324 (Magliab. XX.52), 63v-65r, 15c; 16. Vienna Nat.-bibl. 5322, 147v-150r, 15c. There
of Cremona was the translator but this seems not unlikely. At least the are no proposition numbers as such in the texts of these copies, although the figures for
terminology and style used by the translator bear some similarity to those the propositions are almost always marked and so in effect serve as proposition numbers.
used by Gerard of Cremona.38 Furthermore, we know that the Perspec- There is great variety in the figure numbering. Take for example the proposition we have
called after Risner V.34. One set of MSS (which I can call the “ Edinburgh tradition” )
tends to number each of the two figures included in my Fig. 1.8 (covering the first case to
36 Ibid., p. 104; Taliaferro, op. cit., p. 643. be proved) as “ 12a” (see MSS nos. 2, 4, 7, 8, 9, 12, 13, 16). Of these MSS only MS no. 9
37 In addition to the edition of F. Risner, Opticae thesaurus. Alhazeni Arabis libri septem has the two figures contained in my Fig. 1.9 (covering the second case to be proved— but
etc. (Basel, 1572), there are many manuscripts (see Lindberg, A Catalogue, pp. 17-18). in fact not proved), and MS no. 9 labels them also as “ 12a.” A pair of MSS— those num­
There is also an Italian translation in a unique copy of 1341 (ibid., p. 19). For Arabic bered 3 and 14 (which I can designate the “ Royal College tradition )— labels the figures
manuscripts, see M. Krause, “ Stambuler Handschriften islamischer Mathematiker,” Quel- of my Fig. 1.8 as “ 19a” and the figures for my Fig. 1.9 as “ 20a” (except that the lower
len und Studien zur Geschichte der Mathematik, Astronomie und Physik, Abt. B; Studien, fig. of Fig. 1.8 is carelessly labelled as “ 10a” though “ 19a” is clearly written in the margin
Vol. 3 (1936), p. 476. A. I. Sabra “ Ibn al-Haytham,” Dictionary o f Scientific Biography, at the beginning of the proof of Prop. V.34). There are no figures given for Props. V .32-34
Vol. 6 (1972), p. 205, gives the proper order of the folios of Koprulu MS 952. in MSS nos. 1 and 10, while MSS nos. 5 and 6 contain all the figures of my Figs. 1.8 and 1.9
38 I have seen in this translation some of the characteristics of Gerard's translations but no numbers. MS no. 11 has aberrantly marked both of the figures included in my Fig. 1.8
noted in my Archimedes in the Middle Ages, Vol. 1, p. 30, n. 2. But the whole question with “ 13” and omitted the figures of my Fig. 1.9. Finally it should be noticed that Jordanus
20 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 21

included within brackets in my translation of these propositions the enun­ g linea aequidistans ab, quae sit gn, et sumatur linea, quaecunque sit, z t ;et [per 23 p 1] super
ciations added by Risner under the influence of Witelo’s Perspectiva, punctum z fiat angulus aequalis angulo agd per lineam z j , et [per 31 p 1] ducatur a puncto t linea
aequidistans z /, et sit tm ; et [per 23 p 1] ex angulo t z f secetur angulus aequalis angulo dgn per
where the longer propositions of Alhazen are often broken up into series lineam zm . Haec igitur linea necessario concurret cum tm [per lemma Procli ad 29 p 1], cum sit
of shorter propositions. In the case of Proposition V.33 the addition of inter aequidistantes. Sit punctum concursus m; restat ergo [per 3 ax] angulus m z f aequalis
Risner creates a wrong impression. For in fact Alhazen’s text earlier angulo agn. Et a puncto t ducatur linea aequidistans lineae zm [per 31 p 1], quae sit to, quae
includes the general enunciation of a neusis that embraces both those quidem necessario concurret cum/z [per lemma Procli ad 29 p 1] et sit concursus in puncto k ,
of V.32 and V.33. This enunciation of Alhazen is stated at the beginning et sumatur [per 12 p 6] linea cuius proportio ad lineam zt sicut bg ad qe lineam datam, et sit i .
Deinde fiat super punctum m sectio pyramidalis, quemadmodum docet Apollonius in libro
of what was to become Proposition V.32 for Risner and thus produces secundo de pyramidalibus, propositione quarta, et sit ucm, quae quidem sectio non secat
a rather curious situation because the enunciation added by Risner to lineas ko, kf; et in hac sectione ducatur linea aequalis lineae i , scilicet m c, et producatur usque
V.32 is more restricted than Alhazen’s general enunciation with which ad lineas kt, kf, et sint puncta sectionum ο, l. Igitur, sicut ibidem [8 th 2 coni (/) conicorum]
the proposition begins.42 In my translation I have not included Proposition probatur, erit om aequalis cl. Et a puncto t ducatur linea aequidistans cm [per 31 p l],quae sit
V.32, since it does not involve conic sections, but I remind the reader tf; et [per 23 p 1] super punctum a fiat angulus aequalis angulo zft per lineam and. Palam quod
haec linea concurret cum gd, cum angulus agn sit aequalis/zm angulo [per conclusionem] et
that Proposition V.32 takes up the first case of the neusis, the case in angulus gan angulo z f t [per fabricationem, et totus angulus/z/ aequatus sit toti angulo dga, et
which ab = ag (see my comments below following the translation of V.33 per 32 p 1 anguli ad z e t /s in t minores duobus rectis. Ergo anguli ad g et a ipsis aequales
and V.34). And since, for Alhazen, what was later to become Proposition minores erunt duobus rectis. Itaque per II ax. gd, ad concurrent]. Igitur ad linea aut tanget
V.33 was in fact another part of the general proposition whose enuncia­ circulum aut secabit ipsum. Quoniam si non tetigerit, et arcus ab fuerit maior arcu a g , secabit
tion was given earlier, Alhazen commences the proof of V.33 without arcum ab\ et si ab fuerit minor, secabit arcum ag. Tangat igitur in puncto a. Cum igitur [per
fabricationem] angulus gan sit aequalis angulo zft, et angulus agn angulo/zy, erit [per 32 p 1]
further enunciation but merely with the statement “ Now if ab and ag tertius tertio aequalis, et erit triangulum agn simile triangulo z f y ■ Similiter cum [per
are not equal. . . . ” Finally I note that I have included in my translation fabricationem] agd sit aequalis angulo fzt, erit [per 32 p 1; 4 p, 1 d 6] triangulum agd simile
of V.33 and V.34 the bracketed additions from Risner’s text that include triangulo/ ζ ί . Igitur quae est proportio an ad ag ea est proportio^ ad /z , et quae est proportio
expanded geometrical steps and citations of Euclid, Proclus, Campanus ag ad gd ea est /z ad zt. Quare [per 22 p 5] quae est proportio an ad gd ea est fy ad z t . Verum
and Apollonius. To Risner’s citations of Euclid I have added the name cum [per fabricationem] tm sit aequidistans fl, et f t sit aequidistans Im, est [per 34 p 1] f t
aequalis Im. Quare [per 2 ax] erit aequalis co, cum [per 8 th 2 conicorum Apollonii] mo sit
of Euclid. I have also added a few bracketed phrases of my own (especially aequalis /c; sed [per 34 p 1] mo est aequalis yt, cum [per fabricationem] sit ipsi aequidistans et
in V.34) that make the text clearer. Here, then, are Propositions V.33 ym aequidistans to. Restat ergo [per 3 a\ ] f y aequalis cm; sed [per fabricationem] cm est
and V.34:43 aequalis i. Quare [per 1 ax]/y est aequalis i ; sed [per fabricationem] proportio i [id est, per 7 p 5
fy] ad zt sicut bg ad eq. Igitur [per 11 p 5] proportio an ad gd sicut bg ad eq. Verum angulus gan
de Nemore used a manuscript of the “ Royal College tradition" in citing V.34 as V.19 est aequalis angulo gba, sicut probat Euclides in tertio [32 propositione]; sed [per 29 p 1]
(see below, note 55). Because of the diversity of numbering here noted and also because angulus ngd est aequalis angulo abg, cum [per fabricationem] ng sit aequidistans ab. Igitur
of the ready availability of the Risner edition, I have decided to use Risner’s numbers. I [per 1 ax] angulus ngd aequalis est angulo nag, et angulus ndg communis. Quare [per 32 p 1]
shall also use his text, since, save for his bracketed additions and his added enunciations tertius tertio est aequalis. Quare [per 4 p, 1 d 6] triangulum ndg simile triangulo adg. Igitur
presented in italics, it agrees quite well with the text of these propositions found in the proportio ad ad dg sicut gd ad dn. Quare [per 17 p 6] quod fit ex ductu ad in dn est aequale
manuscripts. quadrato dg. Verum quadratum ad est aequale ei quod fit ex ductu bd in dg, sicut probat
42 Ed. of Risner, p. 142: “ [J2. A puncto dimidiatae peripheriae medio, ducere lineam Euclides, 36 propositione [libri tertii], et quadratum ad est aequale ei quod fit ex ductu ad in dn
rectam ut segmentum eius conterminum continuatae diametro aequetur datae lineae rectae. et ei quod fit ex ductu ad in na [per 2 p 2], et illud quod fit ex ductu bd in dg est aequale quadrato
128 p. 1.] Sumpta circuli diametro, et sumpto in circumferentia puncto: est ducere ab eo dg et ei quod fit ex ductu bg in gd, sicut probat Euclides [3 p 2]. Ablatis ergo aequalibus
ad diametrum extra productam lineam quae a puncto in quo secat circulum usque ad con­ [quadrato nempe dg et rectangulo adn] restat [per 3 ax] ut quod fit ex ductu ad in an sit aequale
cursum cum diametro sit aequalis lineae datae. Verbi gratia, sit qe data linea, gb diameter ei quod fit ex [ductu] bg in gd. Igitur [per 16 p 6] proportio primae lineae ad secundam est sicut
circuli a bg, a punctum datum. Dico quod a puncto a ducam lineam quae a puncto in quo tertiae ad quartam [nempe ut ad ad gd sic bg ad an, et alterne per 16 p 5 ut ad ad bg sic gd ad
secuerit circulum usque ad diametrum gb sit aequalis lineae q e , quod sic constabit. Ducantur an]. Quare [per consectarium 4 p 5] proportio an ad gd sicut bg ad ad. Sed iam dictum est quod
duae lineae ab, ag, quae aut erunt aequales aut inaequales. Sint aequales. . . (Punctua­ proportio an ad gd est sicut bg ad e q . Igitur [per 9 p 5] eq est aequalis a d . Quod est propositum.
tion altered.) The proposition number and the enunciation in brackets are Risner’s. Quod si ad non tetigerit circulum sed secuerit, et fuerit ag maior ab, secabit quidem arcum ag
After Risner’s enunciation we see Alhazen’s general enunciation, which embraces the neuseis [Fig. 1.6]. Secet ergo in puncto h , et ducatur linea hg. Palam [per 22 p 3] quod duo anguli ahg,
of both Propositions V.32 and V.33, his specification of that enunciation and his statement abg valent duos rectos; sed angulus ngd aequalis est abg [per 29 p 1 quia ng parallela ducta est
of the alternatives, i.e. where lines ab and ag are equal and where they are unequal. ipsi at]. Igitur angulus ahg et angulus ngd sunt aequales duobus rectis. Quare [per 13 p 1; 3 ax]
Then follows the solution for the case where lines ab and ag are equal. As I have noted angulus ngd est aequalis angulo nhg, et angulus ndg communis. Quare [per 32 p 1] tertius
above, I have not given that solution since it does not involve conic sections. angulus tertio angulo est aequalis, et triangulum hgd simile triangulo ndg [per 4 p 1; 1 d 6],
43 Ed. of Risner, pp. 143-45: “ [33. A puncto dimidiatae peripheriae non medio, ducere Igitur proportio hd ad dg est sicut proportio dg ad dn. Quare [per 17 p 6] illud quod fit ex ductu
lineam rectam ut segmentum eius conterminum continuatae diametro aequetur datae lineae hd in dn est aequale quadrato d g ; sed quod fit ex ductu ad in hd est aequale ei quod fit ex ductu
rectae. 130p. 1.] Si vero ab et ag non sint aequales [Fig. 1.5] protrahatur [per 31 p 1] a puncto bd in d g , sicut probat Euclides [consectario 36 p 3], et [per 1 p 2] illud quod fit ex ductu ad in dh
22 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 23

est aequale ei quod fit ex ductu dh in dn et dh in an, et [per 3 p 2] quod fit ex ductu bd in dg est [33. From a point that is not in the middle of a semicircumference to draw a
aequale ei quod fit ex ductu bg in gd et quadrato dg . Ablatis igitur aequalibus, scilicet quadrato straight line so that its segment coterminous with the diameter produced is equal
dg et eo quod fit ex ductu dh in dn, restat [per 3 ax] ut illud quod fit ex ductu dh in an sit aequale to a given straight line (=Witelo, Perspectiva, 1.130)].
ei quod fit ex ductu bg in d g . Quare proportio secundae lineae ad quartam, id est an ad gd, sicut Now if ab and ag are not equal [Fig. 1.5], let there be drawn [by Euclid 1.31]
tertiae ad primam, id est bg ad dh [est enim per 16p6ut<//z ad dg sic bg ad an, et per 16 p 5 ut dh from point# line#« parallel to ab; and let any line zt be taken, and [by Euclid 1.23]
ad bg sic dg ad an, et per consectarium 4 p 5 ut an ad dg sic bg ad dh]. Sed iam probatum est let there be constructed at point z an angle equal to Z. agd by means of line z /; and
quod proportio an ad dg sicut bg ad eq. Igitur [per 9 p 5] eq est aequalis dh. Et ita est [by Euclid 1.31] let there be drawn from point t line tm parallel to z f, and [by
propositum. Si veroa# sit minor ab, et secetad arcumab, sit sectionis punctum/z [Fig. 1.7], et Euclid 1.23] let there be cut from A t z f an angle equal to /Ldgn by means of line
ducatur linea hg. Palam [per fabricationem primam et 29 p I] quod angulus ngd est aequalis
zm. This line necessarily meets tm [by the lemma of Proclus to Euclid 1.29] since
anguloabg \ sed [per 27 p 3] anguliabg,ahg sunt aequales quia cadunt in eundem arcum. Igitur
[per 1 ax] angulus ngd est aequalis angulo ahg, et angulus ndg communis. Quare [per 32 p 1]
it is drawn between parallels. Let the point ofjuncture be m . It remains therefore
tertius tertio aequalis, et triangula simila [per 4 p, 1 d 6], Igitur proportio hd ad dg sicut dg ad [by the third axiom of Euclid] that 2 m z /= Lagn. And from point t [by Euclid
dn. Quare [per 17 p 6] quod fit ex ductu hd in dn est aequale quadrato d g ; sed quod fit ex ductu 1.31] let line to be drawn parallel to line zm . This line [to] necessarily meetsfz [by
hd in da est aequale ei quod fit ex ductu bd in dg [per consectarium Campani ad 36 p 3], et [per 1 the lemma of Proclus to Euclid 1.29] and let the juncture be in point k, and let
p 2] quod fit ex ductu hd in da est aequale ei quod fit ex ductu dn in hd et an in hd, et [per 3 p 2] there be assumed [by Euclid VI. 12] a line i whose ratio to line zt is as bg to the
ductus bd in dg valet quadratum dg et ductum bg in dg. Igitur remotis aequalibus [rectangulo given line qe. Then let there be drawn through point m a conic section [i.e. a
nimirum hdn et quadrato dg] erit [per 3 ax] ductus hd in an sicut bg in dg. Igitur [per 16 p 6; 16 hyperbola], as Apollonius teaches in the second book On Conics, Proposition 4,
p, 13 d 5] proportio an ad dg sicut bg ad hd. Sed iam dictum est quod proportio an ad dg est and let it be ucm, which section does not cut lines ko and k f [the asymptotes], and
sicut bg ad eq. Igitur [per 9p 5]eq est aequalis hd. Quod est propositum. Quare a puncto a dato in this section let line me be drawn equal to line i and let it be produced [in both
duximus lineam secantem circulum, et a puncto sectionis ad diametrum est aequalis lineae
directions] to lines kt and kf, which it meets in points o and /. Therefore, as
datae.
proved in the same work [Conics, Proposition II.8], om will be equal to cl. And
“ [34. A puncto peripheriae circuli extra datam diametrum dato, ducere lineam rectam ita
sectam data diametro ut segmentum inter diametrum et punctum peripheriae dato puncto from point t let line t f be drawn parallel to cm [by Euclid 1.31], and [by Euclid
oppositum aequetur datae rectae minori circuli diametro. 133 p. 1] 1.23] let an angle equal to /Lzft be constructed at point a by means oilineand. It is
“ Amplius a puncto dato in circulo extra diametrum eius, est ducere lineam per diametrum clear that this line will meet gd. [For] since Aagn = Cfzm, then Agan = Δζ ft
ad circulum ut pars eius a diametro ad circulum sit aequalis lineae datae. Verbi gratia, abg sit [by construction, and the whole Cfzt = whole zLdga, and by Euclid 1.32 the
datus circulus, bg diameter, a punctum datum, hz linea data [Fig. 1.8]. Dico quod a puncto α angles at z and/ are less than two right angles. Therefore, the angles at g and a
est ducere lineam transeuntem per diametrum bg cuius pars a diametro ad circulum sit equal to them are less than two right angles. Thus by the second axiom of Euclid
aequalis lineae hz. Ducantur lineae a b , a g ; et [per 23 p 1] super punctum Λ fiat angulus aequalis gd and ad will meet.] Therefore, line ad will either touch the circle or will cut it.
angulo agb per lineam mh, et super idem punctum fiat angulus aequalis angulo abg per lineam For if it does not touch it and arc ab > arc ag, it will cut arc ab; and if arc ab is less
hl, et [per 32 p 1] a puncto z ducatur aequidistans lineae hm, quae sit zn, quae quidem secabit hl
[than arc ag], it will cut arc ag. Therefore, let it touch in point a . Therefore, since
[per lemma Procli ad 29 p 1]; et a puncto z ducatur aequidistans hi, quae sit zt, et secet hm in
[by construction] Agan = L z ft and /Lagn = Cfzy, then [by Euclid 1.32] the
puncto i; et a puncto t ducatur sectio pyramidis tp, quam assignavit Apollonius in libro
pyramidum [4 th 2], quae quidem sectio non continget aliquam linearum zn, hl inter quas iacet. third angle will be equal to the third angle, and triangle agn will be similar to
Similiter fiat sectio pyramidis ei opposita inter easdem lineas, quae sit cu. Cum igitur linea triangle z /y . Similarly, since [by construction] Lagd — Lfzt, [by Euclid 1.32, VI.
minima ex lineis a puncto t ad sectionem cu ductis fuerit aequalis diametro b g , circulus factus Def. 1, and VI.4] triangle agd will be similar to triangle fzt. Therefore, an /
secundum hanc minimam lineam, posito pede circini super punctum t, continget sectionem ag = f y / fz and a g ! gd = f z ! zt. Wherefore [by Euclid W.22]an / gd = fy /z t. But
cu. Si vero minima ex lineis a puncto t ad sectionem cu ductis fuerit minor diametro bg,
circulus factus modo predicto secundum quantitatem bg secabit sectionem [cu] in duobus
punctis. Sit ergoic minima et aequalis diametro bg, quae quidem secabit zn et hl cum ducatur quod, cum fuerint duae sectiones oppositae [piramidales inter duas lineas] et producatur linea
ad sectionem quae inter eas interiacet; et [per 31 p 1] ducatur a puncto z aequidistans huic, a [una] sectione ad aliam, pars eius quae interiacet inter unam sectionem et unam ex lineis est
quae quidem secabit hm, hl [per lemma Procli ad 29 p 1], sicut sua aequidistans t c . Secet ergo aequalis alii parti quae interiacet inter aliam sectionem et aliam lineam. Quare qt aequalis est
in punctis m, l, et sit mzl·, et punctum sectionis in quo tc secat zn sit q. Et [per 23 p 1] super cf. Sed [per 34 p 1] tq est aequalis mz cum sit illi aequidistans, inter duas aequidistantes. Igitur
diametrum gb fiat angulus aequalis angulo him, qui sit dgb, et ducantur lineae duae ad, db. [per 1 ax] mz aequalis fc, et [per 34 p 1] zl aequalis tf. Igitur [per 2 ax] ml aequalis tc. Quare
Palam ergo, cum [per 31 p 3] angulus gab sit rectus, alii duo anguli trianguli agb valent rectum proportio tc ad hz sicut ml ad hz [per 7 p 5]. Quare proportio gb ad ed sicut tc ad hz
[per 32 p 1]. Quare angulus Ihm est rectus [constat enim e duobus angulis per fabricationem [demonstratum enim est ut gb ad bd sic Im ad mh, item ut bd ad de sic mh ad hz; ergo per 22 p 5
aequalibus angulis agb, abg rectum aequantibus] et est aequalis angulo gdb, et [per fabrica­ ut gb ad de sic Im ad hz, sed ut Im ad hz sic tc ad/zz; quare per 11 p 5 ut gb adde sic tc ad hz.] Et
tionem] angulus him est aequalis angulo dgb. Igitur [per 32 p 1] tertius tertio, et triangulum cum/c sit aequalis bg [ex thesi], erit [per 14 p 5] ed aequalis hz. Quod est propositum. Si autem
simile triangulo [per4p, 1 d 6], Quare proportio bg ad bd est sicut Im ad mh. Sed quoniam [per lineate ad sectionem cu ducta et minima fuerit minor diametro bg, producatur ultra sectionem
27 p 3] angulus adb aequalis est angulo bga quia cadunt in eundem arcum, et angulus bga donec sit aequalis, et secundum quantitatem eius fiat circulus, qui quidem circulus secabit
aequalis angulo mhz [per fabricationem], est ergo [per 1 ax] angulus adb aequalis angulo mhz. sectionem in duobus punctis, a quibus lineae ductae ad t erunt aequales bg [per 15 d 1], et a
Et iam habemus quod angulus dbg est aequalis angulo hmz. Igitur [per 32 p 1] tertius tertio, et puncto z ducatur aequidistans utrique. Et tunc erit ducere a puncto a modo predicto duas
triangulum deb simile triangulo mhz [per 4 p, 1 d 6]. Sit autem e punctum in quo linea ad secat lineas aequales lineae datae; eritque idem penitus probandi modus.” I have considerably
diametrum bg. Igitur proportio bd ad de sicut mh ad hz. Verum Apollonius [16 th 2] probat altered the punctuation and have italicized the letters designating magnitudes.
24 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 25

since [by construction] tm is parallel to ft and f t is parallel to Im , then [by Euclid dg2], hd-an - bg-dg [by the third axiom of Euclid]. Therefore [by Euclid
1.34] f t = lm. Wherefore [by the second axiom of Euclid] f t = co since [by VI. 16, V.16 and V. Def. 13] an I dg = bg I hd. But it has already been stated
Apollonius, Conics, II.8] m o = l c . But [by Euclid 1.34] m o = y t since [by that an / dg = bg / eq. Therefore [by Euclid V.9] eq - hd, which has been
construction] mo is parallel toy/ and ym is parallel to to . It remains therefore [by proposed.
the third axiom of Euclid] thatf y = cm . But [by construction] cm = /. Wherefore Wherefore from a given point a we have drawn a line cutting the circle
[by the first axiom of Euclid] f y = i ; but [by construction] i I zt = bg I eq, or [by [at h], and [the line] from the point of section to the diameter is equal to a
Euclid V.7] f y / zt = bg I eq. Therefore [by Euclid V. 11] an / g d = bg I eq. But given line.
Z-gcin = L g b a , as Euclid proves in [Book] III [Proposition 32], But [by Euclid [34. From a point given in the circumference of a circle outside of a given
1.29] C n g d = C a b g since [by construction] ng is parallel to a b . Therefore [by diameter, to draw a straight line that is cut by the given diameter so that the
the first axiom of Euclid] L n g d = L n a g , and L n d g is common. Wherefore [by segment between the diameter and the point in the circumference opposite
Euclid 1.32] the third angle is equal to the third angle. Wherefore [by Euclid VI. the given point is equal to a given line that is less than the diameter of the
Def. 1 and VI.4] triangle ndg is similar to triangle adg. Therefore a d / dg circle (= Witelo, Perspectiva, 1.133).]
= g d Idn. Wherefore [by Euclid VI. 17] a d -d n = d g 2. But a d 2 = b d - d g , as Euclid Further, from a point given in [the circumference of] a circle and outside
proves in [Book III] Proposition 36, and a d 2 = ad -dn + ad -n a [by Euclid II.2], and of the diameter, to draw a line through the diameter to the [circumference
b d - d g = d g 2 + bg -gd, as Euclid [II.3] proves. Therefore, with equals subtracted of the] circle so that its segment from the diameter to the [circumference
[i.e. d g 2 and a d-d n ] it remains [by the third axiom of Euclid] that ad -a n = bg -gd. of the] circle is equal to a given line [less than the diameter of the circle].
Therefore [by Euclid VI. 16] the ratio of the first line to the second is as the third to For example, let abg be the given circle, its diameter bg, a the given point,
the fourth [i.e. a d I g d = bg / an, and alternately by Euclid \ A 6 a d l b g = g d Ian]. hz the given line [Fig. 1.8]. I say that from point a we are to draw a line
Wherefore [by deduction from Euclid V.4] an / g d = bg / ad. But it was already intersecting diameter bg whose segment from the diameter to the [circum­
stated that an I g d = bg I eq. Therefore [by Euclid V.9] eq = a d , which has been ference of the] circle is equal to line hz.
proposed. Let lines ab and ag be drawn, and [by Euclid 1.23] let an angle equal to
But if ad is not tangent to the circle but rather cuts it, and if a g > a b , Lagb be constructed at point h by means of line mh, and at the same point
then it will cut the arc ag [Fig. 1.6]. Therefore, let it cut it in point h and let there be constructed an angle equal to Labg by means of line hl·, and
let line hg be drawn. It is evident [by Euclid III.22] that the two angles [by Euclid 1.31] let a line be drawn from point z parallel to line hm, the
ahg and abg [together] equal two right angles. But Lngd = Labg [by Euclid parallel being zn, which parallel will cut hi [by the lemma of Proclus to
1.29 because ng has been drawn parallel to ab]. Therefore, Lahg + Lngd Euclid 1.29]; and let a line be drawn from point z parallel to hi, the parallel
= two right angles. Wherefore [by Euclid 1.13 and the third axiom of Euclid] being zt, which parallel will cut hm in point f, and from point t let there be
Lngd = Lnhg, and Lndg is common. Wherefore [by Euclid 1.32] the third drawn the conic section [hyperbola] tp, which Apollonius has described in
angle is equal to the third angle, and triangle hgd is similar to triangle ndg the Book on Cones [i.e. Conics, II.4], which section does not touch [ever]
[by Euclid 1.4 and VI. Def. 1]. Therefore hd / dg = dg / dn. Wherefore [by either of the lines zn and hi between which it lies [i.e. zn and hi are its
Euclid V I.17] hd-dn = dg2. But ad-hd = bd-dg, as Euclid proves [by a asymptotes]. Similarly let the conic section opposite it [i.e. the opposite branch]
deduction from III.36], and [by Euclid II. 1] ad-dh =dh-dn +dh-an. And be constructed between the same lines [asymptotes] [by Apollonius 1.14 and
[by Euclid II.3] bd-dg = bg-gd + dg2. Therefore with equals subtracted, 11.15], which section we let be cu. Therefore, when the least of the lines
i.e. dg2 and dh-dn, it remains [by the third axiom of Euclid] that dh-an drawn from point t to section cu is equal to diameter bg, the circle drawn
= bg-dg. Wherefore, the ratio of the second line to the fourth, i.e. an / gd, with the foot of the compass at point t and the least line [as its radius] will
is as the third to the first, i.e. bg / dh [for by Euclid VI. 16 dh / dg = bg / an, touch section cu. But if the least of the lines drawn from point t to section
and by Euclid V.16 dh I bg = dg I an, and by a deduction from Euclid V.4 cu is less than diameter bg, the circle described in the aforesaid way with
an I dg = bg I dh]. But it has already been proved that an I dg = bg I eq. diameter bg [as its radius] will cut the section [cu] in two points. Therefore,
Therefore [by Euclid V.9] eq = dh. And it has been thus proposed. let the least line tc be equal to diameter bg, which line will cut zn and hi as
But if ag < ab and ad cuts arc ab, let the point of cutting be h [Fig. 1.7] it is drawn to the section which lies between them [i.e. the asymptotes zn
and let line hg be drawn. It is evident [by the first construction and by Euclid and hi]. And [by Euclid 1.31] from point z let a line parallel to it [tc] be
1.29] that Lngd = Labg; but [by Euclid III.27] angles abg and ahg are equal drawn, which parallel to tc will cut hm and hi [by the lemma of Proclus
because they fall in the same arc. Therefore [by the first axiom of Euclid] to 1.29], just as its parallel tc [cuts them]. Therefore, let it cut [them] in
Lngd = Lahg, and Lndg is common. Wherefore [by Euclid 1.32] the third points m and / and so let it be mzl. And let the point in which tc cuts zn
angle is equal to the third angle, and the triangles are similar [by Euclid VI.4 be q, and [by Euclid 1.23] let an angle equal to Lhlm be constructed on
and VI. Def. 1]. Therefore, hd I dg = dg / dn. Wherefore [by Euclid VI. 17] diameter gb, the angle being dgb, and let the two lines ad and db be drawn.
hd-dn = dg2\ but hd-da = bd-dg [by a deduction of Campanus in Euclid It is evident, therefore, that since Lgab is a right angle [by Euclid III.31],
III.36], and [by Euclid II. 1] hd-da = dn -hd + an -hd; and [by Euclid II.3] the other two angles of triangle agb [together] equal a right angle [by Euclid
bd-dg = dg2 + bg-dg. Therefore, with equals subtracted [i.e. hd-dn and 1.32]. Wherefore Llhm is a right angle [for it is obvious from the fact that by
26 ARCHIMEDES IN THE MIDDLE AGES
THE LATIN WORKS OF ALHAZEN 27

construction the two equal angles agb and abg together equal a right angle]. and the circumference abg, to insert between the two lines a given straight
And [ziIhm] is equal zLgdb, and [by construction] zLhlm = Z-dgb. Therefore line that verges on given point a .” In V.33 Alhazen solves the neusis for
[by Euclid 1.32] the third angle is equal to the third angle, and the triangle three cases: (1) when a is a point of tangency for the inserted line ad (see
is similar to the triangle [by Euclid VI.4 and VI. Def. 1]. Wherefore bg / bd
Fig. 1.5), i.e. when the terminus of the inserted line that lies on the
= Im / mh. But since [by Euclid III.27] Ladb = Lbga because they fall in
the same arc and Lbga ~ zimhz [by construction], therefore [by the first circumference is identical with point a; (2) when a is a point on the
axiom of Euclid] /Ladb = /Lmhz- And we already have it that ZLdbg = Lhmz. semicircumference closer to b than to g and the insertion is thus dh (Fig.
Therefore [by Euclid 1.32] the third angle is equal to the third angle, and 1.6); and (3) when a is closer tog and the insertionis again dh (Fig. 1.7). The
triangle deb is similar to triangle mhz [by Euclid VI.4 and VI. Def. I]. More­ case where a bisects the semicircumference, i.e. where arc ag = arc ab,
over let e be the point in which line ad cuts diameter bg. Therefore, bd / de was solved in Alhazen’s V.32, but, as I have said, without conic sections
= mh / hz. But Apollonius [Conics, 11.16] proves that when there are two (see note 42). Now let us recapitulate the construction and its proof for the
opposite [conic] sections [between two lines] and a line is drawn from one second case of Proposition V.33 (Figs. 1.5a and 1.6) as a kind of paradigm
section to the other, the segment of it which lies between one section and for all three cases. Given are circle abg, diameter bg (indefinitely extended
one of the lines [asymptotes] is equal to the segment of it which lies between as bgd), a point a on the circumference closer to b than to g, and a straight
the other section and the other line [asymptote] [i.e. t f = qc; and with fq
line qe to be inserted between bgd and the circumference and verging on a .
added to each] wherefore qt = cf. But [by Euclid 1.34] tq = mz since they
Alhazen’s construction has as its objective the construction at a of an angle
are parallel and are between parallels. Therefore [by the first axiom of Euclid]
mz = fc and [by Euclid 1.34] zl = tf. Therefore [by the second axiom of one of whose sides cuts the circumference at h so that hd is equal to q e .
Euclid] ml = tc. Wherefore, tc i hz - ml / hz [by Euclid V.7]. Wherefore, Draw straight lines ag and ab\ draw indefinite line gn parallel to ab. Take
gb / ed - tc I hz [for it has been demonstrated that gb / bd = Im / mh and any definite line zt and join it to indefinite line z f at the same angle as zLagd.
that bd I de - mh / hz; therefore by Euclid V.22 gb / de = Im l hz; but Now draw indefinite line tm parallel to line z f . By means of indefinite line
Im ! hz - tc l hz; wherefore by Euclid V .ll gb I de = tc I hz]· And since zm construct at z an zLtzm equal to zLdgn. Line zm and tm meet at point m.
tc = bg [by supposition], then [by Euclid V.14] ed = hz, which has been Hence L m z f = zLagn. Draw through t an indefinite line to parallel to line zm ,
proposed. which to, extended, meets the extension of z / at k. Then by Conics II.4
But if the least line tc drawn to section cu is less than diameter bg, let it be draw through point m a hyperbola cm within asymptotes ko and kf. Take a
produced beyond the section until it equals [bg] and let a circle be drawn with line i such that i / zt = bg I qe. Then starting from m insert a chord in the
this [produced line] as its radius, which circle will cut the section in two points hyperbola equal to i ; let the other terminus of this chord be c . Let chord me
from which lines drawn to t will [each] equal bg [by Euclid I. Def. 15], and from
be produced in both directions until it meets the asymptotes respectively at
point z a parallel is drawn to each line. And thus it will be possible to draw from
0 and l (this last step is necessary for the later proof of the equality of hd and
the point in the aforesaid way two lines equal to a given line, and the method of
demonstration is completely the same [as before; see Fig. 1.9]. qe but not for finding the desired angle to be constructed at a). Hence, by
Conics II.8, mo = cl. Then from t line tf is drawn parallel to mo. Now
These propositions of Alhazen present and solve two neuseis or “ verg­ construct at a an angle gad equal to L z ft and the neusis is solved, for the line
ing” problems like those found in Archimedes’ On Spiral Lines. The ad, verging as it does on a, cuts the circumference at h and meets the
history of these problems is briefly but lucidly treated by Heath.44 We can diameter extended at d so that hd = q e .
begin our discussion by repeating Pappus’ definition of the neusis-type With the construction thus shown, Alhazen then proceeds to the demon­
construction: “ Two lines being given in position, to place between them a stration that hd does in fact equal q e . He first shows that triangle hgd is
straight line given in length and verging towards a given point.’’ The similar to triangle ndg. Hence hd / dg = dg / dn, or hd-dn = dg2. But
neuseis of Alhazen’s Propositions V.33 (together with the particular case of ad-hd = bd-dg, and ad-hd = dh-dn + dh-an. And bd-dg = bg-gd
it in V.32) and V.34 are actually special cases of a more generally expressed + dg2. Hence, with equals subtracted, dh -an = bg -dg,or(l)an I dg = bg I
neusis; “To insert a given straight line between a given diameter (V.34) or dh. But it was shown in the first case that (2) an I dg = bg / eq as follows. It
that diameter extended (V.32, V.33) and the diameter’s circumference that was shown that triangle agd is similar to triangle/ζ ί, that an / ag = fy / fz,
verges on a given point.” More specifically we can express the neusis of that ag / gd = fz I zt, and that an I gd = fy I zt. But tm is parallel to fl, and ft
V.33 in this way: “ Given two lines, the indefinitely produced diameter bgd is parallel to Im. And since, by Conics II.8, it was shown that mo = cl,
hence ft = co; and since mo = yt, hence fy = cm. But by construction
44 T. L. Heath, The Works o f Archimedes (Cambridge, 1897; reprint, New York, 1950). 1 = cm. Hence fy - i. But by assumption i / zt = bg I eq. Thus (2) an /
pp. c-cxxii. Cf. the recent article by W. R. Knorr, “ Archimedes’ Neusis-Constructions dg = b g /eq . Hence by (1) and (2) together dh = eq. Q.E.D. The proofs for
in Spiral Lines,’’ Centaurus, Vol. 22 (1978), pp. 77-98.
the other locations of a are similar.
28 ARCHIMEDES IN THE MIDDLE AGES THE LATIN WORKS OF ALHAZEN 29

Now turning to the neusis of V.34 we can recapitulate its construction Alhazen’s Proposition V.32 (the case not requiring conic sections for its
and proof in a like fashion (see Fig. 1.8). Given are circle abg, its diameter solution) was used in Archimedes’ Proposition 5. Neuseis like that of
bg, a point a in the circumference, and the straight line hz (less than the Alhazen V.33 appear in Archimedes’ Propositions 6 -7 and those like that
diameter) to be inserted between bg and the circumference and verging on of Alhazen V.34 in Archimedes’ Propositions 8 -9 . The neusis of Alhazen
a . Alhazen’s objective is to find the angle to apply at g that will determine d V.33 was also used (without solution) in the eighth lemma of the Liber
and thus yield the intercept ed equal to hz and verging on a . Draw lines ag assumptorum {or Lemmata) attributed to Archimedes for the trisection of
and ab. At terminus h of given line hz construct angles zhm and zhl equal an angle.46 And further, the neusis of Alhazen V.34 was employed for the
respectively to angles agb and abg by drawing indefinite lines hm and hi. same purpose in Proposition XVIII of the Verba filiorum of the Banii Musa,
Draw indefinite line zn parallel to hm, and indefinite line tz parallel to hi. but with a mechanical solution given.47 Alhazen himself uses one or more of
Line tz cuts hm in point t. Then, by Conics II.4, draw through point t the neuseis of V.32, V.33 and V.34 in the series of propositions that lead up
hyperbola tp with asymptotes zn and hi. Similarly construct opposite to V.39 (and indeed also in V.39).48 The latter proposition is the celebrated
branch cu with the same asymptotes, a point being easily determined “ Problem of Alhazen” which is expressed in Risner’s enunciation as
through the necessarily symmetrical placement of the branches between follows: “ With the eye and the visible object at points unequally distant
the asymptotes. Then from t draw a minimum line to the branch cu (by from the center of a convex spherical mirror, to find the point of reflection
finding the midpoint of a chord of cu whose termini are determined by [on the surface of the mirror].”49
placing a compass at t and marking off two points on cu). Then consider One more consideration remains, to review the relationship between the
successively the cases where the minimum line tc is equal to diameter bg Latin texts of Alhazen’s propositions and Apollonius’ Conics and to see
and less than bg (the possibility that tc is greater than bg need not be what knowledge of conics could have been deduced by a contemporary
considered since hz must be less than bg). So first take tc = bg, and with its reader. In the course of solving Propositions V.33 Alhazen specifically
center at t draw a circle with radius bg that touches branch cu at c . Then cites Propositions II.4 of the Conics of Apollonius when he instructs us to
draw through z a line mzl parallel to tc, which line mzl intersects lines hm construct by that proposition a hyperbola (actually he merely says sectio
and hi respectively at m and /. (Alhazen also notes here that tc cuts hi a t / pyramidalis) through point m within asymptotes ko and kf. The enunciation
and zn at q , facts necessary for his later demonstration of the equality of hz of Proposition II.4 of Apollonius says:50
and ed). Now construct at g an Z-dgb = Lhim, the latter angle having been 4. Given two straight lines containing an angle and a point within the angle,
determined by drawing mzl. Points has thus been determined and when ad to describe through the point the section of a cone called the hyperbola so that
is drawn it cuts bg at e and the problem is solved, ed being the desired the given straight lines are its asymptotes.
intercept equal to hz and verging on a .
Later in the same proof Alhazen asserts the equality of om and cl by
With the construction completed, Alhazen then proceeds to prove that in
mentioning Apollonius but not the proposition number. But his authority
fact ed is equal to hz. First he proves, by similar triangles, that (1) bg /
was clearly Proposition II.8 of the Conics, which asserts:51
bd = Im / mh. He then proves, again by similar triangles, that (2) bd /
de = mh / hz. Then qt = cf, since, by Conics II. 16, tf = qc, andfq = qf. 46 Ibid., pp. ci-cii. Cf. my Archimedes in the Middle A ges, Vol. 1, pp. 667-68.
But tq = mz since they are parallel and between parallels, and mz = fc, 47Ibid., pp. 344-49, 668.
and zl = tf. Hence ml = tc. Therefore, tc / hz = ml I hz, and hence gb / 48 Ed. of Risner, pp. 145-51. Propositions V.33-V.38 are presented as lemmas in specific
terms by Sabra, op. cit. in note 37, p. 200 but without sufficient diagrams or proofs for
ed = tc / hz by (1) and (2). But since tc = bg by assumption, then ed = hz. the reader to judge their import. However, the texts of these propositions as given in the
Q.E.D. Risner edition are quite lucid and easily followed.
Now if tc is less than bg, extend it until it equals bg. Once more draw a 49 Ed. of Risner, p. 150: ‘‘[59. Visu et visibili a centro speculi sphaerici convexi inaequabi­
circle with center at t and radius equal to b g . It will cut the hyperbola cu in liter distantibus, punctum reflexionis invenire. 22 p 6.] Predictis habitis, dato speculo sphae­
two points, which are joined to t by lines equal to bg. One then constructs rico, erit invenire punctum reflexionis. . . .” Some students of the problem have, because
of its mathematical equivalence to V.39, designated Proposition V.36 as Alhazen’s Problem
parallels to these lines through point z and each of these lines will provide
(pp. 146-48; “ [56. Duobus punctis extra circuli peripheriam, vel uno extra, reliquo intra
an angle that gives a solution. The proof is not given, though some datis, invenire in peripheria punctum in quo recta linea ipsam tangens bifariam secet angu­
manuscripts provide a diagram for the solution (see Fig. 1.9). lum comprehensum duabus rectis a dictis punctis ad punctum tactus ductis. 135 p. I.] Amplius:
Now that we have examined the solutions of these neuseis in some detail duobus punctis datis, scilicet e, d , et dato circulo, est invenire punctum in eo ut angulum
the reader should be reminded that the neuseis themselves were used by contentum a lineis a punctis predictis ad illud punctum ductis dividat per aequalia linea
circulum contingens in illo puncto. . . .” ) See S. Unguru, Witelonis Perspectivae liber
Archimedes (without solution) in his On Spiral Lines.45 The special case of primus. Book I o f Witelo’s Perspectiva (Wroclaw, Krakow, 1977), pp. 209-12.
50 Heiberg, Gr 1, p. 198; Taliaferro, op. cit., p. 685.
45 Heath, The Works o f Archimedes, pp. c -c i. al Heiberg, Gr 1, p. 204; Taliaferro, op. cit., p. 687.
 THE LATIN WORKS OF ALHAZEN 31
30 ARCHIMEDES IN THE MIDDLE AGES

8. If a straight line meets an hyperbola in two points, produced in both find the mechanical solution lacking in certainty and the solution based
directions it will meet the asymptotes, and the straight lines cut off on it by the on Alhazen’s proposition more sufficient. But before the availability of the
section from the asymptotes will be equal. text of Apollonius’ Conics itself, the Latin reader would have had little
understanding of the Apollonian conics that lay behind the solutions of
In Proposition V.34 Alhazen again appeals to Conics, II.4, but without V.33 and V.34. From the terminology used in the translation (namely
citing the specific proposition number, when once more he instructs us to sectio pyramidalis or sectio pyramidis instead of sectio addita as used
construct a hyperbola (this time called sectio pyramidis in the edition, in the fragmentary introduction to De speculis comburentibus) he would
but sectio pyramidalis in some manuscripts) through point t within asymp­ scarcely have known that he was dealing with a hyperbola. True, the
totes zn and hi. He follows this by instructing us in the same way to con­ vague description of the lines that the section did not meet (i.e. the asymp­
struct the opposite branch within the same asymptotes. While no mention totes) might have given him a hint if, for example, he had read the tract
of Apollonius is included, Alhazen no doubt also had in mind Proposition on the hyperbola that will be discussed in Chapter 2, though even there
1.14 (see above note 24) as well as Proposition 11.15, the latter of which the “ curved line” is never identified by a specific name related to the
tells us that “The asymptotes of opposite sections are common.” 52 When sectio addita or hyperbola and so perhaps he would have had to wait until
he then draws the minimum line from point t to the branch cu, he would the appearance of Witeio's Perspectiva for the identification of the curve as
perhaps have considered that Conics V.34 guaranteed that it was a mini­ a hyperbola. Furthermore, the reader depending solely on Alhazen’s Per­
mum, though the Latin text has no such reference to Apollonius.53 Then spectiva would not have known the form of the enunciations of Conics
further along in the solution of Alhazen’s Proposition V.34 he mentions II.4 and II.8 or their proofs. As for Propositions 11.15 and V.34, the
Apollonius once more and gives the substance of the enunciation of Propo­ reader would have had no way (without the text of the Conics) to know
sition II. 16 of the Conics without however giving the proposition number. that they were being silently appealed to. It is true that he would have
This proposition says:54 known the substance of the enunciation of Proposition II. 16 but not which
16. If in opposite sections some straight line is drawn cutting each of the
proposition it was or its proof. The first author to read Alhazen’s solu­
straight lines containing the angle adjacent to the angles containing the tions in conjunction with the text of Apollonius appears to have been
sections, it will meet each of these opposite sections in one point only, and the Polish optician Witelo, as we shall see in Chapter 3 below.
the straight lines cut off on it by the sections from the asymptotes will be
equal.
Now what would a Latin reader make of these two neuseis of Alhazen
and their solutions? At the least he would consider the problems solved
and use the neuseis for the solution of other problems. Such was the case
with Jordanus de Nemore, who in Proposition IV.20 of the De triangulis
repeats the solution by the BanO Musa of the problem of the trisection
of any angle and then adds a reference to Alhazen’s solution of the second
neusis (our V.34) to supplement the mechanical solution given by the
Banu Musa.55 Incidentally, Jordanus was an acute enough geometer to

52 Heiberg, Gr 1, p. 218; Taliaferro, op. cit., p. 692.

53 Sabra op. cit. in note 37, p. 201, seems to say that the Arabic text has a reference to
Conics, V. 34. For V. 34 see the edition of E. Halley, (separate title and pagination for
Books V, VI and VII), Apollonii Pergaei Conicorum libri tres posteriores (sc. V(us. VItus.
Vllmus) (Oxford, 1710), p. 23: "'Si sumatur punctum in aliqua vel e Maximis vel Minimis,
extra Sectionem Conicam productis: erit portio ejus, quae interjacet punctum illud et Sec­
tionem, Minima rectarum de puncto illo ad utrumvis latus Sectionis egredientium, modo faciam, hoc idem sic demonstro. . . . et a puncto E per BZ semidiametrum ducatur linea
non produci sed in uno tantum puncto Sectioni occurrere concipiantur: e caeteris vero per figuram 19 quinti perspective, ut TS sit equalis semidiametro BL." All the copies of the
quae eidem propinquior minor erit remotiore. . . .” text that I have seen cite V. 19 as the proper number for the neusis that appears in Alhazen’s
54 Heiberg, Gr 1, p. 220; Taliaferro, op. cit., pp. 692-93. Perspectiva (in the Risner edition) as V.34. Though I have not been able to straighten
55 Clagett, Archimedes in the Middle Ages, Vol. 1, for Jordanus' proof, pp. 672-77, and out the difficulties in connection with the proposition numbers (see note 41 above), it is evi­
see particularly (p. 674): “ De divisione anguli in tres partes equales mihi nequaquam suf­ dent that Jordanus used a manuscript in the tradition of the one found at the Royal College
ficit dicta demonstracio, eo quod nihil in ea certum reperio. Ut autem mihi me sufficientem of Physicians (ibid.).
 CHAPTER 2

John o f Palerm o’s Translation o f a

Short Arabic Tract on the H yperbola
Following the translations of the two tracts of Alhazen discussed in
Chapter 1, the next step in the development of a knowledge of conic sec­
tions in the West was taken when a tract entitled De duabus lineis semper
approximantibus sibi invicem et numquam concurrentibus was translated
from the Arabic by John of Palermo, a court philosopher, “ faithful” and
notarius of Frederick II. John’s role as a court philosopher is twice referred
to by Leonardo Fibonacci as he mentions in 1225 that John had proposed to
him mathematical questions.1 We can trace John’s notarial career from
1221 to 1240 in the diplomatic papers of Frederick.2

1 See Leonardo's Flos in Leonardo Pisano, Scritti, ed. B. Boncompagni, Vol. 2 (Rome,
1862), p. 227: “ Cum coram maiestate vestra, gloriosissime princeps, Frederice, magister
Iohannes panormitanus, phylosophus vester, pisis mecum multa de numeris contulisset,
interque duas questiones, que non minus ad geometriam quam ad numerum pertinent,
proposuit. Quarum prima fuit ut inveniretur quadratus numerus aliquis, cui addito vel
diminuto quinario numero, egrediatur quadratus numerus; quem quadratum numerum, ut
eidem magistro Iohanni retuli, inveni esse hunc numerum, undecim et duas tertias et cen­
tesimam quadragesimam quartam unius. . . . (p. 228) Altera vero questio a predicto
magistro Iohanne proposita fuit, ut inveniretur quidam cubus numerus, qui cum suis duobus
quadratis et decem radicibus in unum collectis essent viginti. . . The first problem is
again mentioned in the Liber quadratorum of 1225 (ibid., p. 253): “ Cum Magister dominicus
pedibus celsitudinis vestre, princeps gloriosissime domine F., me pisis duceret presentan-
dum, occurrens Magister Johannes panormitanus, questionem mihi proposuit infrascriptam,
non minus ad geometriam quam ad numerum pertinentem; ut invenirem numerum quadratum,
cui quinque additis vel diminutis, semper inde quadratus numerus oriretur.” The two prob­
lems proposed by John Palermo can be expressed in modern terms: (1) x2 + 5 = y2 while
x2 - 5 = z2; (2) x3 + 2x2 + lOx = 20. K. Vogel, “ Fibonacci, Leonardo,” Dictionary of
Scientific Biography, Vol. 4 (New York, 1971), p. 610, after noting that Leonardo presents
a series of indeterminate linear problems in the Flos, raises the possibility that if the first of
them was suggested by John of Palermo, “ he must have taken it from the algebra of al-
Karajl.” In view of John’s obvious command of Arabic and access to Arabic mathematical
treatises as exemplified by his translation of the tract we are here considering, Vogel’s sug­
gestion is plausible. Vogel further notes that John’s second problem “ appears in the algebra
of al-Khayyami, where it is solved by means of the intersection of a circle and a hyperbola.”
Thus if John had indeed found the problem in al-Khayyami’s tract and had read the latter
thoroughly he had a more considerable knowledge of conic sections than any other Latin
scholar of his time. We can also note that there was at least one other Arabist in Frederick’s
circle, namely, the Magister Theodorus to whom Leonardo addressed a mathematical let­
ter (Scritti, Vol. 2, pp. 247-52) addressing him as “ imperialis aule . . . phylosophe,” for as
I remark in the next footnote he was to prepare a letter in Arabic to the King of Tunis.
2 J. L. A. Huillard-Breholles, Historia diplomatica Friderici Secundi, Vol. 2 (Paris, 1852),
p. 185, where we see a document executed by John for Frederick at Catania in the month of

33
34 ARCHIMEDES IN THE MIDDLE AGES JOHN OF PALERMO’S TRANSLATION 35

It was my good fortune to have discovered this tract— the only medieval The overall objective of our tract is to show that a curved line (in fact a
treatise on conic sections without an optical context— and to have pub­ hyperbolic curved line) and a straight line (in fact an asymptote of the
lished it in a preliminary version over twenty years ago.3 Since that time hyperbola) have the desired relationship of lines that continually come
other copies of a paraphrased version of the tract (Version B below) have closer together but never meet. Thus it is precisely the asymptotic property
turned up, thus demanding an appraisal of all of this material as a part of my of the hyperbola that is being demonstrated, but without our author ever
treatment of conic sections in the Middle Ages. mentioning the terms hyperbola and asymptote. The construction of the
Though the colophon of our tract mentions its title and the fact that John hyperbola and its asymptote and the proof that they have the desired
of Palermo translated it from Arabic into Latin, no mention is made of the relationship are accomplished in Proposition 5, with the four preceding
author’s name. A tract with a similar title is listed among Alhazen’s works propositions considered as antecedent propositions.
(On the Refutation o f the Demonstration that the Hyperbola and its The author commences in the proem by noting that the curved line in
Asymptotes Continually Approach Each Other Without Ever Meeting),4 question is one drawn on [the surface of] a cone. The cone is then defined
but I have found no trace of a manuscript of it and the disputative nature of by reference to the beginning of Book XI of Euclid’s Elements. It is desig­
Alhazen’s tract suggested by its title makes it an unlikely source of John of nated as a figura rotunda solida piramidea. The quotation from Euclid is
Palermo’s translation. A tract on the same subject by Ahmad ibn Muham­ close to the definition as given in the most popular of the Latin versions
mad ibn Abdu’l-Jalil al-Sijzi is extant but has no relationship with our tract.5 from the Arabic, the version designated Adelard II.6 Note that the defini­
tion is that of a right circular cone rather than of the unrestricted circular
May, 1221 (“ . . . presens privilegium sibi exinde fieri jussimus per manus Joannis de cone understood in the definition of Apollonius’ Conics. No doubt our
Panormo, notarii et fidelis nostri. . . . Datum Catane, anno incarnationis dominice MCCXXI, author had in mind throughout the treatise that the cone was to be con­
mense maii, indictione IX.” ). The next document was prepared nineteen years later when
Frederick wrote to the Abbot Henry and to John, ordering them to undertake a mission to sidered a right cone since he mentions in Proposition 1 the right triangle
the King of Tunis, where John’s Arabic would certainly have been useful. John was sick at that describes the cone and he would hardly have altered the form of the
the time, for Frederick gives instructions for choosing another envoy in case John was pre­ cone in succeeding propositions. We note further that in all of the diagrams
vented by illness from going. John’s letter has not been printed, but it is described by the the cones are drawn as right cones. In point of logic the enunciations of
editor as similar to Henry’s, which is given as follows {Ibid., Vol. 5 [Paris, 1859], p. 726): Propositions 3 -5 do not require that the cone be a right cone, i.e. that the
“ Fridericus, etc., Henrico Abbati, etc. Fidelitati tue precipiendo mandamus quatenus una
cum magistro Johanne de Panormo notario et fideli nostro vel, si ipse infirmitate gravatus non
axis be perpendicular to the base, but only that the axial triangle be perpen­
poterit, cum alio quem R. de Amicis justitiarius Sicilie ultra flumen Salsum juxta quod sibi dicular to the base. This is all of no great moment since we have to realize
damus per nostras licteras in mandatis, duxerit eligendum, conferas te ad regem Tunisi cum that the author is not trying to establish in a general or universal way the
licteris nostris quas sibi duximus destinandas, processurus cum eodem notario vel alio quem asymptotic property of any hyperbola but merely to show how to construct
justiciarius eliget, in negotio quod tibi et eidem commictimus, juxta quod Alberius de
an example of a curved line (hyperbola) and a straight line (its asymptote)
Pontetremulo presentium portitor, fidelis noster, qui causam novit, ex parte nostra oretenus
tibi referet et exinde te instruet diligenter. Datum Fuligini, VI februarii, XIII indictione.” that never meet.
[1240] On Feb. 10, G. de Tocco wrote a letter to Magister Theodorus requesting him to write a After defining the right cone the author continues in the proem by noting
letter to the King of Tunis in Arabic on the abovenoted matter (ibid., pp. 727, 745). The final that the curved line in question that falls on the cone is that which he says
notice concerning John of Palermo is a letter of 27 April, 1240, written by Frederick at Orta is described in the proem of the first book of the Conics of Apollonius.
and requesting John’s presence before him post-haste to undertake unspecified services However, no such exact definition exists in Apollonius’ Greek text. In fact
(ibid., p. 928): “ Fredericus, etc., magistro Johanni de Panormo, etc. Fidelitati tue districte
precipiendo mandamus quatenus visis hys licteris, omni mora et occasione cessante, ad
the author of the treatise is not quoting Apollonius but rather the intro­
nos per mare te conferas festinanter, cavens ne inde aliud facias. Te enim volumus pro ductory paraphrase accompanying the Arabic text of Apollonius which
quibusdam nostris servitiis destinare, et ecce quod mictimus tibi barcettam in qua propere was translated by Gerard of Cremona and which we have treated in Chap­
venire valeas. Datum Orte, XXVII aprelis, XIII indictionis.” [1240], It is of interest that ter 1. For if we look at the first numbered definition in footnote 18 of that
Frederick considered John’s presence important enough to send a barque for him.
chapter, we shall see the complete identity of that definition with the one
3 M. Clagett, “ A Medieval Latin Translation of a Short Arabic Tract on the Hyperbola,”
Osiris, Vol. 11 (1954), pp. 359-85. quoted here as being from Apollonius.
4 F. Woepcke, L ’Algebre d ’Omar Alkhayyami (Paris, 1851), p. 74n, quoting Ibn Abi
Usaybi'a, who in turn has given Alhazen’s own list of mathematical works. Cf. G. Nebbia, 6 Its most popular form was in the version prepared by Campanus, which employed the
“ Ibn al-Haytham nel millesimo anniversario della nascita,” Physis, Anno 9 (1967), p. 175 (I- definitions and enunciations of Adelard II (see ed. of Basel, 1546, p. 346): “ Pyramis rotunda
a-18). est figura solida, estque transitus trianguli rectanguli alterutro suorum laterum rectum
5 See Leiden, Bibl. der Rijksuniversiteit, Cod. Or 14(6), pp. 226-31, where the title is given angulum continentium fixo, donec usque ad locum unde moveri coepit, redeat, triangulo ipso
as “ On the matter of two lines— one of which is a straight line and the other the perimeter of a circunducto.” Cf. the Arabie text attributed to al-Tusi, Elementorum geometricorum libri
hyperbolic section— that always approach each other on being extended indefinitely but with­ tredecim ex traditione doctissimi Nasiri-Dini Tusini: nunc primum Arabice impressi (Rome,
out meeting.” 1594), p. 327.
36 ARCHIMEDES IN THE MIDDLE AGES JOHN OF PALERMO’S TRANSLATION 37

In Proposition 1 of De duabus lineis the author proves that the straight triangle abg, then dz must be perpendicular to that diameter and hence
line joining any two points on the surface of a cone that are not in line with tangent to the circular section. Thus no point on dz except d is in the surface
the apex of the cone must lie within the cone. This is Proposition 1.2 of the of the cone. Hence if any other plane is passed through dz and the cone,
Conics of Apollonius (see Chapter 1, n. 19) but without Apollonius’ con­ there is still no point of dz except d that will be on the surface of the cone
cluding phrase (“but if produced it will fall outside”) since that phrase is not and hence dz will be tangent at d to the section formed in the cone by that
germane to our author’s objective. The proof is, needless to say, taken al­ plane.
most verbatim from Apollonius, and the letters marking the magnitudes At this point Version B, which I shall describe in more detail later, notes
are the same as Apollonius’. The comment to the effect that the cone has that the actual demonstration follows when the axis of the cone is per­
been described by a right triangle is not in the Conics and is not necessary. pendicular to the base. Its author adds that it also follows whenever the
Proposition 2, which asserts that the section of a cone made by a cutting axial triangle is perpendicular to the base (regardless of how the axis meets
plane parallel to the base circle is also a circle, has its source (in a round­ the base). But, according to the author of Version B, the demonstration
about way) in the Conics of Apollonius, namely, Proposition 1.4, which runs:7 would simply not apply when the axial triangle is not perpendicular to the
If either one of the vertically opposite surfaces is cut by some plane parallel
base, for then a cutting plane through dz could not be drawn parallel to the
to the circle along which the straight line generating the [conic] surface is base circle as the demonstration requires. But all of this is beside the point
moved, the plane cut off within the surface will be a circle having its center since our author at least tacitly assumed a right cone for all of the proposi­
on the axis, and the figure contained by the circle and the conic surface inter­ tions, as I have said earlier.
cepted by the cutting plane on the side of the vertex will be a cone. . . . The author is ready to present his key antecedent theorem, namely,
Proposition 4 (see Fig. 2.4). Again an axial triangle abg is formed by a plane
The proof in the tract on the hyperbola is essentially the same as Apol­
through the apex and the axis. Following this, he produces side ab beyond
lonius’, except that in the former tract the cone is considered as already
the apex to point d. From d, in the plane of the axial triangle, he draws a
constructed with az as its axis and az intersecting the cutting plane at h (see
straight line dez through the second side of the triangle. At points h and t
Fig. 2.2), d and e are two points on the surface of the cone and a section cut on line dez lines hk and tl are drawn perpendicular to the axial triangle,
off by the plane parallel to base circle b g , and these points are taken to be on points k and / being their respective termini on the surface of the cone.
lines ab and ag. Thus our author proves that the curved line on the surface
Planes parallel to the base circle are then drawn through the perpendiculars
of the cone and passing through points d and e is a circle. This is easily
hk and tl, forming by Proposition 2 circular sections nkm and ols . Then, by
done, for, by similar triangles, a z/a h = bz / dh and bz = zg (since az is the the property of a circle, (\)mh -hn = hk2and (2) st to = tl2. But, by similar
axis and bg is a circle). Thus dh = he, and this would be true for all other
triangles, the author shows that (6) dh-he I mh-hn = dt te I st-to. Hence
points on curved line d e . Hence curved line de is a circle with center at h . (7) dh-he / hk2 = dt-te / tl2, which the fourth proposition enunciates. He
Proposition 3 shows that if we draw a plane through a point on the surface also shows immediately from (1) and (2) that since mh < st and hn < to, so
of a cone and through the axis of the cone and then erect on that point a per­ hk2 < tl2', and hence hk < tl. But the same could be demonstrated for any
pendicular to the previously drawn plane, then that perpendicular is two perpendiculars similarly situated on line ez with respect to external
tangent to any section formed on the surface of the cone by passing a plane point d. Hence, the farther removed a perpendicular is from the external
through that perpendicular and the cone.
point the longer it is. Now so far the author has not mentioned anything
He first constructs axial triangle abg through point d on the surface of the about the section (i.e. the hyperbola) formed in the cone by the plane per­
cone (see Fig. 2.3). He then constructs dz perpendicular to axial triangle pendicular to the axial triangle and drawn through line dez . This he leaves
abg. He then passes through dz a plane parallel to the base of the cone, for Proposition 5. But he has in effect, by demonstrating equation (7), re­
which, according to Proposition 2, forms circular section dh. If straight vealed the fundamental property of a hyperbola, as related to its axis, the
line dh is taken as the diameter of the circle and it falls in the plane of axial property assumed by Archimedes and expressed as the second (italicized)
part of Proposition 1.21 of Apollonius’ Conics:8
7 Heiberg, Gr 1, p. 14 (Cf. above, Chap. 1, n. 6); Taliaferro, op. cit. in Chap. 1, n. 19, p. 606.
In giving a single proof for this proposition, Apollonius includes three diagrams for the three If in a hyperbola or an ellipse or the circumference of a circle straight lines are
cases to which it applies, namely, when the cutting plane that is parallel to the circle along drawn ordinatewise to the diameter their squares are to the areas contained by
which the line generating the conic surface moves is (1) between the apex and the circle, or the straight lines cut off by them beginning from the ends of the transverse
(2), below the circle, or (3) beyond the apex and thus cuts the upper conic surface. In the last
diagram, therefore, we have a full double cone, with the diameter of the bottom side desig­ 8 Heiberg, Gr 1, pp. 72-74; Taliaferro, op. cit., p. 628. For the axial property as under­
nated as BC , that of the upper circle as DE and vertex as A (for which detail compare Fig. 3.11 stood by Archimedes, see T. L. Heath, The Works of Archimedes (Cambridge, 1897; Dover
below). reprint, New York, 1950), p. liii.
38 ARCHIMEDES IN THE MIDDLE AGES JOHN OF PALERMO’S TRANSLATION 39

side of the figure as the latus rectum of the figure is to the transverse side, and To prove this, the author supposes that the two lines meet in point o. If
to each other as the areas contained by the straight lines cut off in the said we draw a perpendicular from o to / ( o n dez), then, by Proposition 4, (2)
manner. dn-ne / nm2 = df-fe / f o 2. And, by (1) and (2) together,
But since neither the curve nor its latus rectum plays any part in Proposi­ (3) df-fe / f o 2 = se2 / eh2
tion 4 of our tract, it is not surprising that the proof diverges completely
from that found in the Conics. Indeed the hyperbolic curve ekl that I have = s f 2 / f o 2 (by similar triangles).
added to Fig. 2.4 as a broken line is properly missing from all the manu­ And so (4) df-fe = s f 2, which is not possible, for in fact df-fe - s f 2 - es2.
scripts of both Versions A and B. Therefore, since o is any point of supposed meeting and it has been shown
At this point the author is fully prepared in Proposition 5 to draw the to be impossible that the meeting occurs, straight lines sh and curve kel will
hyperbola and its asymptote and to prove that the two lines have the de­ never meet.
sired relationship of always approaching but never meeting as they are By an easy march of reason which the reader may examine in Text A
produced indefinitely, a proposition that may be compared with Proposi­ below and which we need not repeat the author also shows that the lines
tion II. 1 of the Conics:9 come always closer to each other as they are indefinitely produced. The
If a straight line touches a hyperbola at its vertex, and from it on both sides comparable theorem in the Conics is Proposition II. 14:10
of the diameter a straight line is cut off equal to the straight line which in square The asymptotes and the section, if produced indefinitely, draw nearer to
is equal to one quarter of the figure [contained by the transverse diameter and
each other and they will come within a distance less than any given distance.
the latus rectum], then the straight lines drawn from the center of the section to
the extremities of the tangent so assumed will not meet the section. Apollonius’ proof, given in an entirely different context to prior proposi­
As we proceed with our discussion of Proposition 5 of the tract on the tions, differs rather completely from the proof found here in Proposition 5.
hyperbola we shall see that it is markedly different from Apollonius’ prop­ It is now clear that the medieval reader of our tract has acquired some
osition, primarily because it does not have behind it the full range of additional knowledge of conic sections from De duabus lineis. He has
Apollonian propositions and it is not concerned with presenting in a general propositions and proof similar to the elementary propositions of the Conics,
way an essential property of a hyperbola. namely, 1.2 and 1.4, but without knowing that they were suggested by the
As before, our author commences by fashioning axial triangle abg (see Conics. He has, in addition, learned something of the hyperbola. He could
Fig. 2.5). As in Proposition 4, he produces side ab beyond the apex to point easily see that the curve formed in Proposition 5 was constructed in the
d and line dez is again drawn in the plane of triangle abg. At point e a line manner that the sectio addita was to be constructed according to the
eh (this time indefinite in length) is drawn perpendicular to triangle abg. paraphrase whose definitions were translated by Gerard of Cremona and
Then a plane deh is drawn that cuts the cone’s surface in line kel (which is, which was discussed in Chapter 1 (see n. 21, def. [3]). Furthermore the
of course, the hyperbolic curve, but which is never so named by the author). reader of De duabus would now know, as the result of studying Proposi­
This line can be produced indefinitely by extending the cone and the cutting tions 4 and 5, the basic axial property of this sectio addita, namely, that
plane indefinitely, with line ez indefinitely produced inside the cone. By the squares of the ordinates are to each other as the areas contained by the
Proposition 3, straight line eh is tangent to section kel. The author then products of the lines cut off by the ordinates from the termini of the trans­
proves by Proposition 1 that kel is a curved line (linea obliqua). verse side. He would also have known by steps (7)-(9) of Proposition 5 (see
Now the author must construct a straight line that is an asymptote to this the translation of Text A below) that an essential property of the curved
curved line (but the author never labels the straight line as an asymptote). line (that is, hyperbola), in relationship to an asymptote, was that the sum
He first takes any point m on kel and draws from m a perpendicular to tri­ of the square of the extension of any ordinate of the curve up to the straight
angle abg that must fall on dez at some point n inside the cone. Then the line (that is, the segment between the ordinate’s terminal point on the
author bisects de at point s (which in Apollonian terms would be desig­ curve and the asymptote) and twice the rectangle having one side equal to
nated the center of the hyperbola). He then determines the length eh by the that extension and the other equal to the ordinate is a constant for all or­
proportion {\)dn -en! nm2 = se2/eh 2. Point h is thus determined and so we dinates. It was this property (along with the axial property mentioned
draw sh and this is in fact the asymptote to k e f or, as the author puts it, is earlier) that Johannes Werner was to take from the De duabus lineis and
the line which if extended indefinitely as ek is extended indefinitely will al­ make fundamental to his treatment of the hyperbola (see Chap. 6 below).
ways approach ek but never meet it. And finally the medieval reader would have known that there is a straight

9 Heiberg, Gr 1, p. 192; Taliaferro, op. cit., p. 682. 10 Heiberg, Gr 1, p. 216; Taliaferro, op. cit., p. 691.
40 ARCHIMEDES IN THE MIDDLE AGES JOHN OF PALERMO’S TRANSLATION 41

line (he possibly would have been able to deduce that there are two such Archimenidis of Johannes de Tinemue.12 This expression is used by the
lines) which the curve of the sectio addita continually approaches without author of B to distinguish that surface from the “ superficies secans” (“ cut­
ever meeting. Thus he would have had some partial knowledge of Proposi­ ting surface”). The common section of both of the surfaces is more pre­
tions 1.21, II. 1 and 11.14 of the Conics of Apollonius, without, of course, cisely designated by the author of B as the “ linea continens circulum”
knowing the actual texts of these propositions. rather than A ’s “circulus” . Note also in line 7 of B that the author has ex­
Version A of our tract exists in only one copy, that of MS Oxford, Bod­ panded A’s “ lineam de" to “ lineam curvam de." The rest of the proof has
leian Library, D ’Orville 70, 61v-62v, a careful and excellent copy that may been rewritten with no change of substance.
have been written in the thirteenth century not long after the tract was As we turn to Proposition 3 we again see that the author of B has rewrit­
translated, although the manuscript as a whole has been dated to the ten the enunciation and proof, expanding some of A’s expressions. For
fourteenth century.111 have designated it as O in the variant readings that example, instead of A ’s “ superficiem equidistantem basi” in line 13 of that
accompany the text. The only infelicity I have detected is the designation version, B has in line 16 “ superficiem secantem piramidem equidistanter
of two different points by the letter z both in Fig. 2.5 and the text of Proposi­ basi.” B’s author refers back to Proposition 2 in line 18 when he says
tion 5 .1 have changed the upper z into z ' (while the author of Version B has “per premissam.” There is no such reference in A. I have already dis­
changed it into*, or perhaps he saw a copy of Version A that hadx). I have cussed above the interesting remarks of the author of B as to the insuf­
added the proposition numbers missing from the text and have capitalized ficiency of the demonstration of Proposition 3. Of course, rtothing reveals
the enunciation of the propositions in order that they might stand out. I also more clearly than these remarks that the author of Version B has assumed
present an English translation of O in which I have used some simple the role of a commentator on Version A.
modern notation that does not, I believe, misrepresent the text so long as In Proposition 4 the author of Version B rewrites the enunciation in such
we realize that my presentation of ratios in fractional form is only for con­ a way as to make the terms of the proportion clearer. For example, instead
venience and does not imply any modern theory of real numbers. Finally of A’s statement (lines 10-13) designating the product of the lines cut off by
the reader should realize that this unique copy of Version A is the only copy the perpendiculars as “ superficiei existentis ex linea que est inter punctum
of either version to contain the colophon with its indication that John of extra piramidem positum et quamlibet perpendicularem ducta in partem
Palermo has translated the tract from the Arabic. huius linee que est infra piramidem” the author of B writes (lines 11-14)
Version B constitutes the first reaction to De duabus lineis so far as I can “ superficiei contente sub linea a puncto extra piramidem usque ad primam
tell. It varies little from Version A in the proemium. There is one marginal perpendicularem et sub linea que est intra piramidem usque ad primam
gloss in the hand of the scribe of its best manuscript (P ): “ sagitta et axis perpendicularem.” Notice also that in both enunciation and proof the
eius.” (See the variant readings, Version B, Proemium, line 9.) In fact, the author has specified that the axial triangle is perpendicular to the base. No
scribe of P, who I believe to have been the author of Version B, has through­ such specification was made in Version A, I suppose because the author
out the text added a few explanatory glosses, as we readily see by examin­ was tacitly assuming throughout that the axis itself was perpendicular to
ing the variant readings to Version B. the base. The author of B no doubt thought that, in line with his earlier
Proposition 1 has been slightly rewritten in Version B and reveals criticism of Proposition 3, he should indicate in Proposition 4 as well that
evidence of scholastic adaptation, especially in such phrases (missing from the proper limitation was not that the axis was perpendicular but only that
Version A) as “ Ista consequenda est manifesta” and “ Ista consequencia the axial triangle was perpendicular. This he felt important because Prop­
patet.” The term “ consequencia” (“ implication” ) is a regular term in the osition 5 employs both Proposition 3 and Proposition 4. Again, note that
logical form adopted by the schoolmen in the late thirteenth and fourteenth in Proposition 4 (line 35) the author of B has used the expression curva
centuries. Also the reader will see references to Euclid’s Elements missing superficies for lateral surface of the cone. Also remark that our later author
from Version A. in line 35 specifically refers back to Proposition 2 when indicating that the
Equally interesting are the changes made in Proposition 2 by the author common section of the cutting plane and the curved surface is the circum­
of Version B. Note that instead of using A ’s expression “ Omnis rotunde ference of a circle. No such reference appears in Version A. Finally, notice
piramidis” he writes “ Omnis piramidis cuius basis est circulus.” And in­ that the author specifies the two parts of the enunciation as “ primum” and
stead of speaking merely of the “ superficies” of the cone he speaks of the “ secundum” (see line 48).
“ curva superficies.” The latter expression became popular after the ap­ Proposition 5 in Version B stays quite close to the text of Version A. In
pearance in the early thirteenth century of the Liber de curvis superficiebus
11 F. Madan, A Summary Catalogue o f Western Manuscripts in the Bodleian Library at 12 See M. Clagett, Archimedes in the Middle Ages, Vol. 3 (Philadelphia, 1978), p. 216. For
Oxford, Vol. 4 (1897), p. 55, where no mention is made of our tract. the text of this tract, see Vol. 1, pp. 439-530.
42 ARCHIMEDES IN THE MIDDLE AGES JOHN OF PALERMO’S TRANSLATION 43

establishing that line eok (designated as line kel in Version A) is a linea Cod. 5277 is grossly incomplete. It omits the proem, gives the enuncia­
obliqua, the author of Version B (line 26) makes specific reference to tions but not the proofs for Propositions 1-4 . Proposition 5 is complete (but
Proposition 1. No such reference appears in Version A. The same is true of without the added lemma). It has a title: De lineis semper appropinquanti­
the reference to Proposition 4 in Version B (see line 38) where it is neces­ bus et numquam concurrentibus, but no colophon. Also lacking are prop­
sary to use the relationship established in that proposition between the osition numbers and diagrams. It is evident that the author of this copy was
squares of the perpendiculars and the products of the cut-offlines indicated only interested in the final construction; hence his reduction of the first four
in the enunciation. There is also specific appeal to the authority of Euclid’s propositions to their mere enunciations to serve as premises for Proposi­
Elements which is not present in Version A. One final remark is necessary tion 5. This codex was no doubt also copied from MS P or a close copy of
concerning Version B. In Version A, it was, as I indicated above, simply P. But Cod. 5277 was not copied from Cod. 5176 since it contains parts of
asserted that we can construct a perpendicular eh such that dn-en / nm2 the text in P that were omitted from Cod. 5176. Like Cod. 5176, Cod. 5277
= se2 / eh2. The author of Version B felt that the method of finding such a is of no use in the establishment of the text of Version B and I have not in­
line needed elaboration and so he showed how this was to be done in an ad­ cluded its variant readings.
dendum to Proposition 5 (see Lemma added to Prop. 5). He had already Finally, we should note that the De duabus lineis influenced to some
indicated at appropriate points in Proposition 5 (at which I have placed degree the important treatise Speculi almukefi compositio (see Chap. 4 be­
asterisks) that he would make such an addition (see variant readings, Prop­ low) and even more markedly the Libellus super vigintiduobus elementis
osition 5, lines 32-33 and 53-54). conicis of Johann Werner (see Chap. 6 below). It is somewhat surprising
A word is necessary concerning the manuscripts of Version B. The best that Leonardo Fibonacci appears not to have used the tract.16 But this may
copy (and indeed, I believe, the autograph copy of the version) appears in merely mean that the tract was translated after the completion of the lat­
MS Paris, BN lat. 7434, 79v-81r, a manuscript of the fourteenth century ter’s Practica geometrie in 1220.
(but perhaps this part is of the thirteenth),13 which I designate as P. It is
excellently written with marginal (and occasionally interlinear) notes in the
hand of the author (=m. 1) and with other notes that are in a different hand
(=m. 2). These notes have all been given in the variant readings to Version
B below. They are largely explanatory or give references to the Elements.
There is, however, one reference (see variant readings, Prop. 4, lines
42-44) to a Libellus de proporcionibus thebith a work attributed in at
least one manuscript to Thabit ibn Qurra and in another to Jordanus.14
There is a third hand (m. 3) that attempts to explain and redraw Fig. 2.4 (see
the end of the variant readings to Proposition 4). MS P contains no colophon
or title. It numbers the propositions in the margins. The marginal folio
references in my text of Version B below are to MS P on which the text is
exclusively based.
Two other manuscripts of Version B are extant, both in the National-
bibliothek in Vienna: Cod. 5176,143v-46r and 5277, 276v-77r, both MSS of
the sixteenth century.15 The copy in Cod. 5176 is fairly close to the text in
MS P and probably was made from it. It, however, was very carelessly
made with many omissions and errors. Further, it is without originality and
its variants are useless for the establishment of the text of Version B. I
have, therefore, not reported these readings. Cod. 5176 has no title,
colophon or proposition numbers and omits the lemma to Proposition 5.

13 Catalogus codicum manuscriptorum Bibliothecae Regiae, Pars tertia, Vol. 4 (Paris,

1794), p. 358, dates it as 14c. 1 believe this part could be thirteenth century.
14 M. Clagett, Archimedes in the Middle Ages, Vol. 2, (Philadelphia, 1976), p. 19. 16 In the Practica geometrie, (Scritti, Vol. 2, pp. 180-81), Leonardo has the equivalent of
15 Tabulae codicum manu scriptorum praeter graecos et orientales in Bibliotheca Palatina Proposition 2 of the De duabus lineis but he drew the proposition and proof from the Verba
Vindobonensi asservatorum, Vol. 4 (1870), pp. 49, 82. filiorum of the Banu MQsa(see my Archimedes in the Middle Ages, Vol. 1, pp. 298-301,363).
 DE DUABUS LINEIS (VERSION A) 45

A. non opponitur puncto a . Dico quod linea de in piramide intus incidit, quod
sic probatur.
The Treatise on Two Lines Always Approaching Trahamus duas lineas ad, ae, et protrahamus eas in rectum. Incidunt
Each Other but Never Meeting io ergo super circumferentiam circuli, et incidant super puncta b, g. Linee
ergo ab, ag sunt in superficie piramidis, quam scilicet trigoni orthogonii
latus rectum angulum cordans descripsit cum idem trigonus volveretur.
[Tractatus de duabus lineis semper approximantibus sibi Et protrahamus lineam bg. Sit ergo linea bg intus in piramide. Et ponamus
invicem et numquam concurrentibus] supra lineam de punctum z. Et protrahamus lineam az. Incidit ergo supra
15 lineam bg, eo quod triangulus abg in una superficie est. Incidat ergo supra
[Proemium] punctum h. Quia ergo punctum h est intus in piramide, et sit linea ah intus
in piramide. Similiter demonstramus in omni puncto quod positum fuerit
6iv / Demonstrare volumus qualiter protrahamus duas lineas inter quas in supra lineam de. Linea ergo de intus est in piramide. Et hoc demonstrare
initio sui exitus sit aliqua distantia et quanto magis protracte fuerint eadem voluimus.
distantia minuitur et alter alteri appropinquat; nec unquam contingi pos­ [2.] OMNIS ROTUNDE PIRAMIDIS QUE SUPERFICIE SECATUR
sunt, etiam si protracte fuerint in infinitum. Sit autem altera linearum recta EQUIDISTANTE SUE BASI EADEM SUPERFICIES FACIT IN
5 et altera obliqua. Trahimus autem huiusmodi lineam obliquam in piramide SUPERFICIE [PIRAMIDIS] CIRCULUM CUIUS CENTRUM EST
rotunda quam exposuit Euclides in undecimi libri prohemio ubi dixit: SUPER SAGITTAM.
Figura rotunda solida piramidea est quam continet triangulus rectangulus, 5 Exempli causa. Piramis abg [Fig. 2.2], cuius basis sit circulus bg et
fixo altero latere rectum angulum continente, et revoluto trigono eodem sagitta linea az, secta est superficie circulo bg equidistante, et facit in
quousque redeat ad locum unde incepit. Huius autem figure sagitta est superficie piramidis lineam d e . Dico quod linea de est circulus cuius centrum
10 fixum latus, et basis est circulus. Cum igitur fuerit in hac piramide quam est supra sagittam, quod sic probatur.
Euclides exposuit latus rectum angulum cordans protractum in infinitum, Quoniam linea az incidit supra superficiem que secat piramidem, incidat
piramidem describit infinitam. Unde peractis hiis propositionibus demon­ 10 ergo in ipsa supra punctum h. Et ponamus supra lineam de duo puncta,
stramus qualiter sit in ipsa piramide obliqua linea infinita. Linea autem d, e, et protrahamus lineas ad, ae, et incidant supra circulum bg supra
obliqua que incidit in piramide est quam exposuit Apollonius in prohemio puncta b, g. Et protrahamus lineas zb, zg, hd, he. Sit ergo linea dh equi-
15 primi libri piramidum ubi dixit: Si piramis recta secetur superficie non per distans linee bz, et he equidistans zg. Proportio ergo az ad ah est tanquam
punctum sui capitis transeunte, communis sectio sit superficies quam proportio bz ad dh, et tanquam proportio zg ad he, et linea bz est equalis
obliqua continet linea. Deinde probamus qualiter protrahimus rectam 15 linee zg. Linea ergo dh est equalis linee he. Et similiter demonstramus in
lineam inter quam et lineam curvam aliqua sit distantia et quanto magis he omni puncto quod positum fuerit supra lineam de quod linea protracta a
protracte fuerint linee sibi invicem appropinquant, nec unquam coniungen- puncto h ad ipsum sit equalis linee dh. Linea ergo de est circulus cuius
20 tur, etiam si protracte fuerint in infinitum. centrum est punctum h, quod est supra sagittam, quod demonstrare
[1.] OMNIUM DUORUM PUNCTORUM IN SUPERFICIE ROTUNDE voluimus.
PIRAMIDIS EXISTENTIUM, NEC IN OPPOSITIONE CAPITIS [3.] CUM PROTRACTA FUERIT A PUNCTO EXISTENTE IN
PIRAMIDIS ENTIUM, LINEA CONIUNCTIVA INTUS IN PIRAMIDEM SUPERFICIE ROTUNDE PIRAMIDIS PERPENDICULARIS SUPRA
INCIDIT. SUPERFICIEM TRANSEUNTEM PER ILLUD PUNCTUM ET PER
5 Exempli causa. Esto piramis abg [Fig. 2 .1], cuius basis circulus bg, et in SAGITTAM PIRAMIDIS, EADEM PERPENDICULARIS TANGIT
cuius superficie duo puncta d , e. Linea autem coniungens eadem puncta 5 OMNEM SECTIONEM QUAM FECERIT IN SUPERFICIE PIRAMIDIS
OMNIS SUPERFICIES PROTRACTA SUPER IPSAM.
Tit. [Tractus . . . concurrentibus] addidi·, cf. Prop. 5, lin. 65-66
Exempli causa. In piramide abg punctum d [Fig. 2.3]. Piramidis autem
Proemium sagitta sit linea ae, et superficies que transierit per punctum d et per sagit-
[Proemium] addidi
12 ante piramidem del. O pro / Unde correxi ex ubi in O 11 quam correxi ex quas in O
14 post obliqua dei. O infinita 14 Incidit: Indcidit O
Prop. 2
Prop. 1 11 ante ad dei. O d
1 [1] addidi; cf. Vers. B, et similiter pro numeris sequentium propositionum 14 bz' bis O
44
46 ARCHIMEDES IN THE MIDDLE AGES DE DUABUS LINEIS (VERSION A) 47

tam piramidis faciat in ipsa triangulum abg. Et protrahamus a puncto d ad quadratum hk est tanquam proportio superficiei existentis ex dt in te ad
io lineam dz perpendicularem supra superficiem trigoni abg. Dico quod linea quadratum tl, quod sic probatur.
dz tangit omnem sectionem quam fecerit in superficie piramidis omnis 30 Trahamus a punctis t, h duas lineas equidistantes linee bg, et sint linee
superficies que per ipsam protrahatur, quod sic probatur. mhn ,sto. Superficies ergo in qua sunt h k ,m n tst equidistans basi piramidis.
Protrahamus super lineam dz superficiem equidistantem basi. Ipsa igitur Facit ergo in piramide circulum cuius diametrum (/) est linea mhn. Sit ergo
facit in superficie piramidis circulum. Sit ergo circulus dh. Sit ergo diameter circulus mkn. Similiter facit superficies in qua sunt tl, so in piramide circu­
15 linea dh, et linea dz est perpendicularis super ipsam. Linea ergo zd tangit lum cuius diameter est linea s to . Sit ergo circulus sio. Linea igitur mn est
circulum. Non est ergo de linea zd in superficie piramidis quicquam preter 35 diametrum (/) circuli mkn, et linea hk perpendicularis super ipsam. Ductus
punctum d. Cum igitur protrahatur alia superficies quocunque modo con­ ergo mh in hn est equalis quadrato hk. Similiter ductus st in to est equalis
tingat supra lineam zd et fecerit in superficie piramidis sectionem, lineam quadrato tl. Et proportio superficiei existentis ex dh in he ad superficiem
dz tanget illa sectio, quod demonstrare voluimus. existentem ex mh in hn est tanquam proportio composita ex dh ad mh et ex
eh ad hn. Tamen proportio dh ad mh est tanquam proportio dt ad ts, et
[4.] OMNIS ROTUNDE PIRAMIDIS QUAM SECUERIT SUPER­ 40 proportio eh ad hn est tanquam proportio et ad to. Proportio ergo super­
FICIES SUPRA TRIANGULUM TRANSEUNTEM PER EIUS SAGIT­ ficiei existentis ex dh in he ad superficiem existentem ex mh in hn est tan­
TAM, ET PROTRACTA FUERIT LINEA AB EIUS SUMITATE IN quam proportio composita ex dt ad ts et ex et ad t o , que est tanquam pro­
RECTITUDINEM ALTERIUS LATERUM TRIANGULI AD ALIQUOD portio superficiei existentis ex dt in te ad superficiem existentem ex st in to .
5 PUNCTUM EXTRA PIRAMIDEM EXISTENS, ET PROTRACTA Proportio igitur superficiei existentis ex dh in he ad superficiem existentem
FUERIT AB ILLO PUNCTO LINEA AD BASIM TRIGONI, POSITIS 45 ex mh in hn est tanquam proportio superficiei existentis ex dt in te ad
PUNCTIS SUPER EANDEM LINEAM INTRA TRIGONUM, ET EX­ superficiem existentem ex st in to. Iam autem demonstratum est quod
TRACTIS INDE PERPENDICULARIBUS SUPRA SUPERFICIEM ductus mh in hn est equalis quadrato hk, et quod ductus st in to est equalis
TRIGONI PERVENIENTIBUS AD SUPERFICIEM PIRAMIDIS, SIT quadrato tl. Proportio igitur superficiei existentis ex dh in he ad quadratum
io PROPORTIO SUPERFICIEI EXISTENTIS EX LINEA QUE EST hk est tanquam proportio superficiei existentis ex dt in te ad quadratum
INTER PUNCTUM EXTRA PIRAMIDEM POSITUM ET QUAMLIBET 50 tl. Et similiter demonstramus in omni puncto quod ponitur supra lineam
PERPENDICULAREM DUCTA IN PARTEM HUIUS LINEE QUE ez, etiam si protrahatur in infinitum; quod demonstrare voluimus.
EST INFRA PIRAMIDEM AD QUADRATUM ILLIUS PERPENDICU­ Item dicimus quod linea hk est minor linea tl, eo quod quadratum hk est
LARIS TANQUAM PROPORTIO SUPERFICIEI EXISTENTIS EX equale superficiei existenti ex mh in hn, et quadratum tl iam probatum est
15 LINEA QUE EST INTER EXTRINSECUM PUNCTUM ET ALTERAM esse equale superficiei existenti ex st in to , et linea mh est minor linea s t , et
PERPENDICULAREM DUCTA IN PARTEM HUIUS LINEE QUE 55 linea hn est minor linea to . Quadratum ergo hk est minus quadrato tl. Linea
EST INFRA PIRAMIDEM AD QUADRATUM EIUSDEM ALTERIUS ergo hk est minor linea tl. Et similiter demonstramus in omnibus
PERPENDICULARIS, ET QUOD QUECUNQUE EX HIS PERPEN­ perpendicularibus protractis a quolibet puncto posito supra lineam ez,
DICULARIBUS FUERIT REMOTIOR A PUNCTO EXTRINSECO etiam si protrahatur in infinitum, quod quanto magis remote fuerint
20 A PIRAMIDE (EADEM EST MAIOR EA) EX PERPENDICULARI­ perpendiculares, tanto maiores; quod demonstrare voluimus.
BUS EIDEM EXTRINSECO PUNCTO PROPINQUIOR.
62r Exempli causa. / Piramis abg secta est superficie per eius sagittam [5.] HIS ITAQUE DEMONSTRATIS DEMONSTRANDUM EST
transeunte [Fig. 2.4], que facit in ea [triangulum] abg. Deinde extractum QUALITER PROTRAHIMUS DUAS LINEAS QUARUM ALTERA
est latus ab ad punctum d, et extracta est a puncto d linea dez ad basim bg, EST RECTA ET ALTERA OBLIQUA, QUAS DESCRIPSIMUS IN
25 que secuit lineam ag supra punctum e. Et posita sunt supra lineam ez duo PREMISSIS.
puncta, h, t, et extracte sunt ab ipsis perpendiculares supra triangulum 5 Esto igitur piramis abg [Fig. 2.5], cuius basis circulus b g , et secetur super­
abg, scilicet hk,tl. Dico quod proportio superficiei existentis ex dh in he ficie transeunte per eius sagittam, faciente in ipsa triangulum abg. Et
protrahamus in rectum lineam ab ad punctum d. Et protrahamus a puncto
Prop. 3 d ad basim bg lineam dez■ Et protrahamus a puncto e perpendicularem
13 post Ipsa del. O ii (?)
19 sectio correxi ex sectionem in O
Prop. 4 32 Sit correxi ex si in O
6 ante ad dei. Ο 1 49 superficiei existentis correxi ex existentis superficiei in O
10 existentis correxi ex existens in O 51 etiam correxi ex et in O
23 facit correxi ex fecit in O 52 dicimus corr. ex dicamus in O
48 ARCHIMEDES IN THE MIDDLE AGES DE DUABUS LINEIS (VERSION A)

supra triangulum abg, et sit perpendicularis he. Et protrahamus super- protrahamus perpendicularem z'cu. Proportio igitur superficiei existentis
10 ficiem deh, que secet piramidem supra lineam kel. Sit ergo perpendicularis ex dn in ne ad quadratum nm est tanquam proportio quadrati se ad quadra­
eh tangens sectionem kel, et sit perpendicularis deh stans perpendiculariter tum eh, et tanquam proportio quadrati sn ad quadratum nq. Proportio igitur
supra triangulum abg. Sectio autem communis inter ipsam et triangulum adiectionis quadrati sn supra superficiem dn in ne, que est quadratum s e , ad
abg est linea ez, et hec linea cum eo quod iungitur ei in infinitum sit inter adiectionem quadrati nq supra quadratum nm, que est quadratum mq et
lineas ab, ag cum protracte fuerint linee ab,ag in infinitum. Linea ergo ez, ductus mq in mn bis, est tanquam proportio quadrati se ad quadratum eh.
15 cum protracta fuerit in infinitum, sit infra piramidem abg cum protracta Quadratum igitur eh equatur quadrato mq et ductui mq in mn bis. Similiter
fuerit, et iam protracta est super ipsam lineam superficies, videlicet que etiam demonstramus quod quadratum eh equatur quadrato uc et ductu[i]
iungitur superficiei deh in infinitum. Hec igitur superficies secat super­ uc in z'c bis. Quadratum igitur mq et ductus mq in mn bis equatur quadrato
ficiem piramidis supra lineam infinite. Secet igitur ipsam supra lineam kel uc et ductui uc in z'c bis; et linea mn est minor linea z'c. Linea igitur qm est
cum eo quod iungitur ei in infinitum. Et omnia duo puncta que posita fuerint maior linea uc. Similiter etiam perpendicularis protracta ab m ad lineam
20 supra lineam kel non opponuntur supra linea kel, eo quod omnia duo sh est maior perpendiculare protracta a puncto c . Punctum igitur c propin­
puncta que posita fuerint in ipsa sunt in superficie deh, et caput piramidis, quius est linee sh quam punctum m . Similiter autem demonstramus in omni
quod est punctum a, est in sumitate superficiei. Duo ergo puncta non puncto existente super obliquam lineam ek in infinitum. Et hoc est quod
fuerint in oppositione capitis piramidis. Linea ergo coniungens utrumque demonstrare voluimus.
incidit infra piramidem. Non igitur kel, existens in superficie piramidis, Explicit tractatus de duabus lineis semper approximantibus sibi invicem
25 est protracta in oppositione quorumlibet duorum punctorum que posita et nunquam concurrentibus, translatus a magistro Ihoanne (Iohanne)
fuerint in ipsa. Linea igitur kel est obliqua. Panormitano de arabico in latinum.
Ponamus igitur in ipsa punctum m , et protrahamus ab ipso perpendicu­
larem supra triangulum abg, et sit perpendicularis mn. Manifestum est
igitur quod ipsa incidit supra distinctionem communem superficiei deh et
30 trianguli abg, que est dz, et incidit supra portionem que est ex hac linea The Treatise on Two Lines Always Approaching Each Other
infra piramidem, eo quod incidit post punctum e , a quo protracta est per­ but Never Meeting
pendicularis eh contingens. Et dividamus lineam de in duo media supra
punctum s, et ponamus proportionem superficiei existentis ex dn in en ad
Proem
quadratum nm tanquam proportionem quadrati se ad quadratum eh. Et
35 protrahamus lineam sh. Dico igitur quod linea sh et obliqua linea ek, si
We wish to demonstrate how to draw two lines such that in the beginning
protrahantur in infinitum, non coniungentur; quod sic probatur.
of their extension there is some distance between them, but the more they
Quoniam impossibile est eas coniungi, quod, si possibile est, coniungan-
are protracted, this distance is diminished and the closer they approach
tur; igitur supra punctum o. Et protrahamus a puncto o perpendicularem
each other, without, however, ever meeting, even if they are extended indefi­
supra lineam d e , et sit perpendicularis of. Sit igitur perpendicularis supra
nitely. Further, let one of the lines be a straight line and the other a curve.
40 triangulum abg. Proportio igitur superficiei existentis ex dn in ne ad qua­
The curved line we draw on a cone of the sort that Euclid described in the
dratum nm est tanquam proportio superficiei d f in fe ad quadratum fo.
beginning of the eleventh book [of the Elements], where he said: “ A cone
< Proportio igitur superficiei d f infe / ad quadratumfo est tanquam proportio
is the solid figure which a right triangle describes when one of the sides in­
quadrati se ad quadratum eh, et tanquam proportio quadrati s f ad
cluding the right angle remains fixed and the triangle is rotated until it re­
quadratumfo. Superficies d f infe est equalis quadrato sf, quod impossibile
turns to the place whence it began.” The axis of this figure is the fixed side
45 est; adicit (/) enim quadrato es. Linea igitur sh non iungitur oblique linee
and the base is a circle. When, therefore, in this cone which Euclid de­
ek, etiam si protrahatur in infinitum.
scribed the side opposite the right angle is protracted indefinitely, it de­
Item dicimus quod, quanto magis prolongate fuerint, tanto sibi in in­
scribes an indefinitely large cone. And so with these propositions deter­
vicem appropinquant, eo quod trahamus lineam nm ad punctum q\ et
mined, we demonstrate how there can be an indefinitely long curved line
ponamus aliud punctum super obliquam lineam ek, scilicet punctum c; et
in this cone. The curved line which falls on the core is that which Apol­
lonius described in the proem of the first book of the Conics, where he said:
33 en correxi ex de in O ” If a cone is cut by a plane not passing through its apex, the common sec­
39 ante of del. O os (?) tion is a surface which a curved line contains. ’’ Then we prove how to draw
48 ante ad del. O ab (?) a straight line such that between the straight line and the curved line there
50 ARCHIMEDES IN THE MIDDLE AGES ON TWO LINES 51

is some distance, but the more these lines are protracted, the closer they 3. WHEN THERE IS DRAWN FROM A POINT ON THE SURFACE
approach each other, without ever meeting, even if they are protracted in­ OF A CONE A PERPENDICULAR TO A PLANE PASSING THROUGH
definitely. THAT POINT AND THE AXIS OF THE CONE, THAT SAME PER­
PENDICULAR IS TANGENT TO EVERY SECTION FORMED ON
1. IN THE CASE OF ANY TWO POINTS LYING ON THE SUR­
THE SURFACE OF THE CONE BY ANY PLANE DRAWN THROUGH
FACE OF A CONE WHICH ARE NOT IN LINE WITH THE APEX OF
THE CONE, THE STRAIGHT LINE JOINING THESE POINTS THAT PERPENDICULAR.
Exemplification. Let point d be on the cone abg [see Fig. 2.3]. Let the
FALLS INSIDE OF THE CONE.
axis of the cone be line ae, and let the plane which will have passed through
Exemplification. Let there be a cone abg, whose base is the circle bg,
point d and the axis of the cone form triangle abg in the cone. And let us
and on whose surface are two points, d and e [see Fig. 2.1]. Further, the
draw from point d the line dz perpendicular to the plane of the triangle abg.
line joining these points is not in line with point a . I say that the line de falls
I say that line dz is tangent to every section formed on the surface of the
inside of the cone, which is proved as follows.
cone by every plane drawn through that perpendicular, which is proved as
Let us draw two lines ad, ae, and extend them in a straight line. There­
fore, they fall upon the circumference of the circle. Let them fall on points follows.
Let us draw through line dz a plane parallel to the base. Hence it makes
b and g. Therefore, the lines ab and ag are in the surface of the cone which
a circle in the surface of the cone [by Proposition 2]. Let it be circle dh.
the side of the right triangle opposite the right angle described when the
Therefore, let the diameter be line dh, and line dz is perpendicular to the
triangle was rotated. And let us draw the line bg, letting it be within the
diameter. Therefore, line zd is tangent to the circle. Hence no part at all of
cone, and let us posit point z on straight line d e . Then let us draw line az. It,
line zd is in the surface of the cone except point d. Therefore, when another
therefore, falls upon line bg because the triangle abg lies in one plane.
plane is drawn in any way at all through line zd and forms a section on the
Therefore, let it fall upon point h. Since, therefore, point h is inside the
surface of the cone, that section will be tangent to line dz. Q.E.D.
cone, the line ah is inside the cone. We similarly demonstrate [the same
thing] for every point which is placed on line de. Line de, therefore, is in­ 4. IN THE CASE OF EVERY CONE CUT BY A PLANE TO FORM
side of the cone. Q.E.D. A TRIANGLE PASSING THROUGH THE AXIS OF THE CONE, IF A
LINE IS DRAWN— IN THE STRAIGHT LINE DIRECTION OF ONE
2. IN THE CASE OF EVERY CONE CUT BY A PLANE PARAL­
OF THE SIDES OF THE TRIANGLE— FROM THE APEX OF THE
LEL TO ITS BASE, THE CUTTING PLANE FORMS A CIRCLE ON
CONE TO SOME POINT OUTSIDE OF THE CONE, AND IF A LINE
THE SURFACE [OF THE CONE]. THE CENTER OF THIS CIRCLE
LIES ON THE AXIS [OF THE CONE]. IS DRAWN FROM THAT POINT TO THE BASE OF THE TRIANGLE,
Exemplification. Cone abg, whose base is the circle bg and whose axis THEN, ASSUMING POINTS ON THAT LINE WITHIN THE TRI­
ANGLE AND PERPENDICULARS ERECTED TO THE SURFACE OF
is the line az, is cut by a plane parallel to circle bg [see Fig. 2.2], It forms
THE TRIANGLE AT THOSE POINTS, PERPENDICULARS WHICH
line de on the surface of the cone. I say that line de is a circle whose center
CONTINUE UNTIL THEY MEET THE SURFACE OF THE CONE,
lies on the axis, which is proved as follows.
THE RATIO OF (1) THE PRODUCT OF THE LINE BETWEEN THE
Since line az intersects the plane which cuts the cone, let it intersect the
POINT OUTSIDE OF THE CONE AND ANY ONE OF THE PER­
plane at point h . And let us place two points, d and e , on line d e , and let us
PENDICULARS AND THE PART OF THAT LINE WITHIN THE
draw lines ad and a e , [extending them] until they fall on the circle bg at
CONE TO (2) THE SQUARE OF THAT PERPENDICULAR IS AS THE
points/? andg. Then let us draw lines zb, zg, hd, and he. And so let linedh
RATIO OF (3) THE PRODUCT OF THE LINE BETWEEN THE OUT­
be parallel to line bz and he parallel to zg. Therefore,
SIDE POINT AND ANOTHER PERPENDICULAR AND THE PART
az I ah = bz / dh OF THAT LINE WHICH IS WITHIN THE CONE TO (4) THE SQUARE
OF THE SECOND PERPENDICULAR. FURTHER, WHICHEVER OF
—zg / he, THESE PERPENDICULARS IS FARTHER FROM THE POINT
OUTSIDE THE CONE IS LONGER THAN ANY PERPENDICULAR
and line bz is equal to line zg. Therefore, line dh is equal to line he. We CLOSER TO THE SAME EXTRINSIC POINT.
similarly demonstrate for every point placed on line de that a line pro­ Exemplification. Cone abg is cut by a plane passing through its axis,
tracted from point h to the assumed point is equal to line dh. Line d e , there­ which forms triangle abg in the cone [see Fig. 2.4]. Then the side ab is ex­
fore, is a circle whose center is point h, which is on the axis. Q.E.D. tended to point d, and the line dez is drawn from point d to base bg, cut-
52 ARCHIMEDES IN THE MIDDLE AGES ON TWO LINES 53

ting line ag at point e. And two points, h and t, are posited upon line ez, — ONE STRAIGHT AND THE OTHER CURVED— WHICH WE DE­
and at these points are erected perpendiculars to triangle abg, namely, SCRIBED IN THE PREMISES.
[perpendiculars] hk and tl. I say, that Let there be a cone abg, whose base is the circle bg, and let it be cut by
a plane passing through its axis to form triangle abg in the cone [see Fig.
dh-he / hk2 = dt-te / tl2, 2.5]. And let us extend line ab to point d and let us draw line dez from point
which is proved thus. d to the base bg, and erect at point e a perpendicular to triangle abg. Let
Let us draw from points t and h two lines parallel to line b g , and let these the perpendicular be he. Then let us draw the plane deh, which cuts the
be the lines mhn and sto. Hence the plane in which hk and mn lie is paral­ cone in line kel. Hence let the perpendicular eh be tangent to the section
lel to the base of the cone. Therefore [by Prop. 2] this plane forms in the kel [by Proposition 3]. Let the plane deh be perpendicular to the triangle
cone a circle whose diameter is line mhn. Hence let the circle be mkn. abg. Then the common section of the plane and the triangle abg is line ez,
Similarly the plane in which tl and so lie forms a circle in the cone whose and this line when continued indefinitely is between the lines ab and ag,
diameter is line sto. Hence let the circle be slo. Therefore, line mn is the when they are extended indefinitely. Hence line ez protracted indefinitely
diameter of circle mkn; and line hk is perpendicular to mn. Therefore [from is inside of the cone abg protracted [indefinitely]. Now a plane is erected
the property of a circle], on this line, a plane which is the indefinite continuation of plane deh. Hence
(1) mh ■hn = hk2, this plane cuts the surface of the cone in an infinitely long line, which we let
and similarly, be the indefinite continuation of kel. Now any two points taken on kel are
(2) st -to = tl2. not in a straight line on kel, since any two points on kel are in the plane of
And deh, and the apex of the cone, point a, is at the summit of the surface [of
(3) dh -he / mh -hn — {dh / mh)-(eh / hn). the cone]. Therefore the two points will not be in line with the apex of the
But cone. Hence the straight line joining these points falls inside the cone [by
(4) dh / mh = dt / ts [by similar triangles], Proposition 1]. Therefore, kel, being on the surface of the cone, is not ex­
and tended in the straight line of any two points posited on kel. Hence kel is a
(5) eh / hn = et / to [by similar triangles]. curved line.
Therefore, And so let us posit point m on kel, drawing from it a perpendicular to
(6) dh -he / mh -hn = (dt / ts)-(et / to) triangle abg, and let it be the perpendicular mn. It is clear, therefore, that
=dt-te / st-to. it falls upon the common section of plane deh and triangle abg, which is dz,
and it falls upon the segment of this line inside the cone, because it falls be­
But with (1) and (2) already demonstrated, then [by substitution in (6)], yond points, from which the tangent perpendicular eh is drawn. Now let us
(7) dh-he / hk2 = dt-te / tl2. divide line de into two halves at point 5, and let us posit that

And we similarly demonstrate [the same thing] for every point posited on (1) dn -en / nm2 = se2 / eh2
line ez, even if it is protracted indefinitely. Q.E.D.
We also say that line hk is less than line tl, because, assuming (1) and (2) [since we can construct eh any length we wish]. Then let us draw line sh. I
and the fact that line mh is less than line st and that line hn is less than line say, therefore, that if line sh and the curved line ek are indefinitely ex­
to, hence tended, they will not meet; which is proved as follows.
It is impossible that they meet, for suppose that it were possible for them
(8) hk2 < tl2, and therefore,
to meet, and that they do so at point o. Then let us draw a perpendicular
(9) hk < tl. from point o to line de[z). Let this perpendicular be of. It is then a perpen­
dicular to triangle abg. Therefore,
And similarly we demonstrate for all perpendiculars protracted from any
point on line e z , even if it is protracted indefinitely, that the farther removed (2) d n -ne / nm2 = df-fe / f o 2 [by Proposition 4].
the perpendiculars are [from the external point], the longer they are.
Q.E.D. Therefore,
(3) df-fe / f o 2 = se2 / eh2 [by (1) and (2)]
5. AND SO WITH THESE [PREVIOUS PROPOSITIONS] DEMON­
STRATED, IT IS TO BE SHOWN HOW WE DRAW THE TWO LINES = s P If o 2
54 ARCHIMEDES IN THE MIDDLE AGES

[since by similar triangles, s f / fo = se / eh]. [Therefore,] B.

(4) df-fe = s f 2, Another Version of the De duabus lineis
which is impossible, for [actually] df-fe = s f 2 - es2 [since d f = s f + se
andfe = s f - se]. Therefore, linesh does not meet the curved lineek, even
if it is protracted indefinitely.
We also say that the more the lines are prolonged, the closer to each [Tractatus de duabus lineis semper approximantibus sibi
other they come. For let us extend line nm to point q, and let us posit invicem et nunquam concurrentibus]
another point on the curved line ek, namely, point c. Let us draw the per­
pendicular z'cu. Then
(5) dn -ne / nm2 —se2 / eh2 [by construction] [Proemium]
= sn2 / nq2 [since se I eh = sn I nq].
79v / Demonstrare volumus qualiter protrahamus duas lineas inter quas in
Therefore, inicio sui exitus aliqua distancia sit et quanto magis protracte fuerint eadem
(6) {sn2—dn -ne) / {nq2—nm2) = se2 / eh2, and distancia minuitur et altera alteri appropinquat; nec umquam coniungi
sn2—dn -ne = se2, [since dn = sn + se, ne —sn—se], and possunt, etiam si protracte fuerint in infinitum. Est autem altera earum
nq2—nm2 = mq2 + 2mq-mn [since nq —mq + mn]. 5 recta et altera obliqua. Trahimus autem huiusmodi lineam obliquam in
piramide rotunda quam exposuit Euclides in proemio libri 11 ubi dixit:
Hence Figura rotunda solida piramidea est quam continet triangulus rectangulus,
(7) eh2 = mq2 + 2mq-mn, and similarly we demonstrate that fixo altero latere rectum angulum continente, et revoluto trigono eodem
quousque redeat ad locum unde cepit. Huius autem figure sagitta est fixum
(8) eh2 = uc2 + luc-z'c. 10 latus, et basis eius est circulus. Cum igitur fuerit in hac piramide quam
Therefore, Euclides exposuit latus rectum angulum cordans protractum in infinitum,
(9) mq2 + 2mq-mn = uc2 + 2uc-z'c. piramidem describit infinitam, in qua, peractis hiis propositionibus que ad
hoc antecedunt, demonstrabimus qualiter fiat in ipsa piramide linea obliqua
But [by Proposition 4] line mn is less than line z'c. Hence [to satisfy (9)], infinita que alicui recte semper appropinquat nec cum ea unquam concurrat
line qm is greater than uc. Similarly the perpendicular drawn from m to 15 quantumcunque protrahantur. Linea autem obliqua que incidit in pira­
line sh is greater than the perpendicular drawn from point c. Therefore, mide est quam exposuit Apollonius in proemio libri piramidum ubi dixit: Si
point c is closer to line sh than is point m . We similarly demonstrate [the piramis recta superficie non per punctum sui capitis transeunte secetur,
same thing] for every point lying on the curved line ek [extended] indefi­ communis sectio recte superficiei secantis et curve superficiei piramidis
nitely. Q.E.D. erit linea obliqua. Deinde probabimus qualiter protrahamus lineam rectam
Here ends the treatise on two lines always approaching each other but 20 inter quam et curvam aliqua sit distancia et quanto magis hee linee pro­
never meeting, translated by master John of Palermo from Arabic into Latin. tracte fuerint magis sibi invicem appropinquent nec umquam tamen
coniungantur, etiam si in infinitum protrahantur.
Prima propositio antecedens.
1. OMNIUM DUORUM PUNCTORUM IN SUPERFICIE ROTUNDE
PIRAMIDIS EXISTENCIUM, NEC IN OPPOSICIONE CAPITIS
TRANSEUNCIUM, LINEA RECTA CONIUNGENS EA VADIT
INTUS IN PIRAMIDEM.

Tit. [Tractatus . . . concurrentibus] addidi

Proemium
[Proemium] addidi
9 de sagitta mg. scr. m. 1 sagitta et axis eius
55
56 ARCHIMEDES IN THE MIDDLE AGES DE DUABUS LINEIS (VERSION B) 57

5 Esto piramis abg super circulum b g , et sint duo puncta d, e in superficie AD SUPERFICIEM TRANSEUNTEM PER ILLUD PUNCTUM ET
eius non in opposicione capitis que coniungat lineade recta [Fig. 2.1]. Dico PER SAGITTAM PIRAMIDIS, ILLA PERPENDICULARIS EST
lineam de ire intus in piramidem. 5 CONTINGENS OMNEM SECTIONEM QUAM FECERIT IN SUPER­
Quoniam protraham a puncto a per d et e lineas rectas in superficie FICIE PIRAMIDIS OMNIS SUPERFICIES PROTRACTA PER IPSAM,
piramidis usque ad eius basim sintque adb, aeg et quia super circum- HOC EST DICTU QUOD DE ILLA PERPENDICULARI NON EST
io ferenciam circuli bg signata sunt duo puncta b , g , linea recta coniungens ea IN SUPERFICIE PIRAMIDIS NISI SOLUM PUNCTUM SIGNATUM.
secat circulum. Ista consequencia est manifesta per quamdam proposi- Verbi gratia, sit piramis rotunda abg, cuius basis circulus bg, et ipsius
cionem tercii elementorum et est de 7 primis nescio quota. Protraham io circuli centrum est e, et in eius superficie sit signatus punctus d [Fig. 2.3],
itaque bg que transibit intra circulum, eritque triangulus abg secans Sagitta autem piramidis sit linea a e , sitque superficies secans piramidem
piramidem. Ista consequencia patet ex prima vel secunda vel tercia l l mi 80r super punctum d, et super sagittam eius triangulus abg ad / quam super­
15 elementorum. Quare linea de que erit intra triangulum illum in eius super­ ficiem protraham a puncto d perpendicularem dz. Dico quod linea dz
ficie est intra piramidem. Hoc etiam patet, signato puncto h in linea d e , et contingit omnem sectionem quam fecerit in superficie piramidis omnis
protracta linea ah usque ad bg, que ei occurrat in puncto z, hec enim linea 15 superficies que transit per ipsam, quod sic probatur.
tota est intus in piramide. Protraham super lineam dz superficiem secantem piramidem equi-
distanter basi. Ipsa igitur facit in superficie piramidis lineam continentem
2. OMNIS PIRAMIDIS CUIUS BASIS EST CIRCULUS SECTE
circulum per premissam, que sit dmh, cuius diameter est linea dh. Et quia
SUPERFICIE EQUIDISTANTE BASI COMMUNIS SECTIO CURVE
linea dz est perpendiculariter extracta a termino diametri dh, ipsa erit
SUPERFICIEI PIRAMIDIS ET SUPERFICIEI SECANTIS EST LINEA
20 contingens circulum. Non est igitur ex linea dz in superficie piramidis
CONTINENS CIRCULUM CUIUS CENTRUM EST SUPER SAGIT-
quicquam preter punctum d. Cum igitur protrahitur per punctum d aliqua
5 TAM SIVE AXEM PIRAMIDIS.
superficies secans piramidem quocunque modo contingat, linea dz erit
Sit piramis abg, cuius basis sit circulus bg, seceturque superficie equi-
contingens sectionem quam fecerit in superficie piramidis, et hoc est quod
distante basi que faciat in superficie piramidis lineam curvam de [Fig.
2.2]. Dico lineam de esse circumferenciam circuli et habere centrum super demonstrare volumus.
25 Nota quod ista demonstracio tenet in omni piramide cuius sagitta per­
sagittam piramidis.
pendicularis est ad basim. In aliis etiam tenet cum superficies transiens
io In linea enim de signentur duo puncta quelibet, que sint d et e , et protra­
per punctum d et per sagittam ut est hic triangulus abg secat basim pira­
hantur linee per eaabn ad basim, sintque adb, aeg, sitquez centrum basis,
midis orthogonaliter. Cum autem iste triangulus non secat basim piramidis
et protrahatur a z que erit sagitta piramidis, et ubi transit per superficiem
orthogonaliter, tunc non tenet quamvis sit conclusio vera, non enim potest
secantem piramidem signetur punctum h, et protrahantur linee hd, h e,zb,
30 protrahi super lineam dz, cum triangulus iste non secat basim piramidis
zg', quia igitur trianguli abz et agz secant basim piramidis et superficiem
orthogonaliter, aliqua superficies secans piramidem equidistanter basi.
15 secantem que sunt equidistantes, erunt earum communes sectiones equi-
Quare in tali casu assumitur falsum in demonstratione cum dicitur super
distantes, scilicet dh, b z , he, et z g . Erit igitur proporcio bz ad dh et zg ad he
sicut za ad/ia; quia igitur zb ,zg sunt equales, erunt dh et he equales, sicque lineam dz superficiem et cetera.
probabimus de ceteris venientibus a puncto h ad lineam d e ; ergo de est 4. OMNIS PIRAMIDIS CUIUS BASIS EST CIRCULUS QUAM
linea continens circulum et centrum eius est punctum h, quod est super SECUERIT SUPERFICIES SUPRA TRIANGULUM SECANTEM
20 sagittam, per tercium librum sive per diffinitionem circuli. BASIM PIRAMIDIS ORTHOGONALITER TRANSEUNTEM PER
3. CUM A PUNCTO EXISTENTE IN SUPERFICIE PIRAMIDIS Prop. 3
CUIUS BASIS EST CIRCULUS EXTRAHITUR PERPENDICULARIS 3 de superficiem mg. scr. m. 1 a qua extrahitur
19-20 de ipsa . . . circulum mg. scr. m. 1 per 15 tercii
26-28 de In . . . orthogonaliter mg. scr. m. 1 Istud ultimum probatur per 29 primi
Prop. 1 et diffinitionem superficiei orthogonaliter insistentis super superficiem et
6 de in opposicione capitis mg. scr. m. / puncta illa solum dicuntur esse in quartam peticionem primi illam, scilicet que petit lineas duas concurrere super
opposicione capitis piramidis que sunt in una et eadem linea recta ducta a quas cadit tercia continens 2 angulos cum illis duabus priores ex eadem parte
capite ipso, id est, a cono piramidis ad circumferenciam basis minores duabus (/ duobus) rectis
Prop. 2 28 ante triangulus del. P b
2 supra superficie scr. m. 1 i.e. per superficiem Prop. 4
15 supra equidistantes1 scr. m. I ex ypothesi scilicet 1 supra quam scr. m. 1 scilicet piramidem
16-17 de Erit . . . ha scr. m. 1 per secundam 6li 3 supra transeuntem scr. m. I triangulum dico
58 ARCHIMEDES IN THE MIDDLE AGES DE DUABUS LINEIS (VERSION B) 59

EIUS SAGITTAM, SI UNUM EX DUOBUS LATERIBUS TRIAN- igitur proporcio superficiei contente sub dh et he ad contentam sub mh et
5 GULI AD ALIQUOD PUNCTUM EXTRA PIRAMIDEM SECUNDUM hn composita est ex proporcione dh ad hm et eh ad hn, et proporcio con­
RECTITUDINEM EXTRAHATUR ET AB EODEM PUNCTO AD tente sub dt et te ad contentam sub ts et to est composita ex ea que est dt ad
BASIM TRIGONI LINEA RECTA DUCATUR, SI IPSA SUPRA QUAM ts et ea que est et ad to , est autem eadem dh ad hm que est dt ad ts , et eadem
INFRA TRIANGULUM DUO PUNCTA SIGNENTUR A QUIBUS 45 similiter eh ad hn que est et ad to, sequitur ut proporcio superficiei con­
USQUE AD SUPERFICIEM PIRAMIDIS DUE PERPENDICULARES tente sub dh et he ad contentam sub mh et hn, et ideo ad quadratum hk, sit
10 AD SUPERFICIEM TRIGONI EDUCANTUR, ERIT PROPORCIO sicut proporcio superficiei contente sub dt et te ad contentam sub st et to , et
SUPERFICIEI CONTENTE SUB LINEA A PUNCTO EXTRA PIRA­ ideo ad quadratum tl, quod est primum. Secundum sic. Cum linea st sit
MIDEM USQUE AD PRIMAM PERPENDICULAREM ET SUB maior mh et similiter to maior hn, maior erit superficies contenta sub st et
LINEA QUE EST INTRA PIRAMIDEM USQUE AD PRIMAM PER­ 50 to quam sub mh et hn. Ergo quadratum tl est maius quadrato hk, ideoque
PENDICULAREM AD QUADRATUM PRIME PERPENDICULARIS et linea tl maior linea hk, quod est secundum.
15 SICUT PROPORCIO SUPERFICIEI CONTENTE SUB LINEA A
PUNCTO EXTRA USQUE AD SECUNDAM PERPENDICULAREM 5. HIIS ITAQUE 4 ANTECEDENTIBUS DEMONSTRATIS DE­
ET SUB EA QUE EST INTRA PIRAMIDEM, SCILICET A PUNCTO MONSTRANDUM EST QUALITER PROTRAHANTUR DUE LINEE
SECTIONIS EODEM, USQUE AD SECUNDAM PERPENDICU­ QUARUM ALTERA EST RECTA ET ALTERA OBLIQUA QUE
LAREM AD QUADRATUM SECUNDE PERPENDICULARIS, ET SEMPER QUANTO AMPLIUS PROTRAHUNTUR AMPLIUS AP-
20 QUECUNQUE EX ISTIS PERPENDICULARIBUS REMOCIOR 5 PROPINQUANT NEC TAMEN CONCURRUNT ETIAM SI IN IN­
FUERIT A PUNCTO EXTRA PIRAMIDEM ERIT MAIOR, QUE FINITUM FUERINT PROTRACTE.
AUTEM SIBI PROPINQUIOR ERIT BREVIOR. Sit ergo piramis abg [Fig. 2.5], cuius basis est circulus bg, et secetur
Exempli gratia, sit piramis abg [Fig. 2.4], cuius basis circulus bg, et 80v / superficie transeunte per eius sagittam, que faciat in ipsa piramide tri­
secat eam triangulus abg transiens super eius sagittam et super circulum angulum abg. Et protrahamus latus eius ba usque ad d, et protrahamus a
25 bg orthogonaliter, et protrahatur latus eius ba usque ad d, et a puncto d io puncto d ad basim bg lineam dez, et protrahamus a puncto e perpendicu­
protrahatur linea dz usque ad basim b g , et super eam signentur intra tri­ larem supra triangulum abg, que sit eh, et protrahamus superficiem deh,
angulum ipsum duo puncta h, t, et ab hiis punctis producantur due linee que secet piramidem super duas lineas dez et eh, et sit communis sectio
hk et tl perpendiculares ad superficiem trianguli abg, que protrahantur huius superficiei et superficiei piramidis linea existens in superficie pira­
quousque occurrant superficiei piramidis in duobus punctis k et /. Dico midis, que sit linea eok, eritque per tercium theorema perpendicularis eh
30 tunc quod proporcio superficiei contente sub dh et he ad quadratum hk est 15 contingens sectionem eok, et est superficies deh stans perpendiculariter
sicut proporcio superficiei contente sub dt et te ad quadratum tl et dico supra triangulum abg, et communis earum sectio est linea ez, et hec linea
iterum quod tl est maior hk. ez cum eo quod iungitur ei in infinitum sit inter lineas ab et ag cum ipse
Protrahantur enim in triangulo abg due linee mhn et sto equidistantes protracte fuerint in infinitum. Linea ergo ez cum protracta fuerit in in­
linee bzg, et intelligatur superficies secans piramidem in qua sint due linee finitum existit intra piramidem abg cum protracta fuerit in infinitum, et
35 hk et hmn, eritque per secundam communis sectio curve superficiei pira­ 20 iam protracta est super ipsam lineam superficies, videlicet que iungitur
midis et superficiei secantis linea continens circulum que sit mkn, cuius superficiei deh in infinitum. Hec igitur superficies secat piramidem in in-
diameter erit mn. Similiter intelligatur alia superficies secans piramidem in
qua sint due linee tl et sto, eritque similiter linea sio continens circulum et 42-44 de et proporcio . . . to mg. scr. m. 1 sicut patet ex 6 elementorum et libello
de proporcionibus thebith
eius diameter so, et erit superficies que continetur sub hm et hn equalis
43 post et1 dei. P et
40 quadrato hk, et superficies contenta sub st et to equalis quadrato tl. Quia 4 4 - 45 supra ts2 . . . similiter scr. m. I per 2 6‘ elementorum
4 5 - 46 de proporcio . . . hn mg. scr. m. I Istud per 7am aut 9 aut 10 5“ patet
51 de Fig. 2.4 scr. mg. m. 3 hk, tl perpendiculares sic quod k et 1 sunt in acie et
7 post basim trigoni mg. add. m. I ita scilicet quod cadat intra duo puncta triangulus equidistans orizonti et piramis iacet (?) et cont. supra triangulus
terminalia ipsius basis et add. m. 2 i.e. in diametro basis inter duo puncta scindit piramidem per medium et inferius est alia figura (cum notis) in m. 3,
finalia ipsius diametri et hoc est in fundo piramidis in diametro basis que non est bona et eam hic omitto
10 de superficiem mg. scr. m. I a qua educuntur Prop. 5
11 supra superficiei scr. m. I scilicet rectangule quadrangule 1 sub 5 mg. scr. m. 1 disposicio
13 supra intra piramidem scr. m. I scilicet a puncto sectionis 10 post dez mg. scr. m. 2 que est in superficie trianguli abg ita quod e est in
38 ante tl del. P z puncto sectionis et z est diametro basis piramidis inter extremitates diametri
39-40 de et erit . . . tl mg. scr. m. 1 per 30 tercii elementorum et 8am 6l' et a puncto e protrahitur linea eh perpendicularis in puncto e in triangulo abg
60 ARCHIMEDES IN THE MIDDLE AGES DE DUABUS LINEIS (VERSION B) 61

finitum. Secet igitur ipsam super lineam eok cum eo quod iungitur ei in quadratum se, ad [illud quod addit] quadratum [nq super quadratum
infinitum, et quia omnia duo puncta super lineam eok non sunt in op­ mn, et ipsum est] quadratum qm et illud quod fit ex qm in mn bis, sicut
positione capitis piramidis quia omnia sunt in superficie deh que non transit proporcio quadrati se ad quadratum eh. Ergo per secundam partem none
25 per verticem piramidis, linea coniungens quelibet duo puncta in ipsa linea 60 quinti Euclidis quadratum eh est equale quadrato qm et ei quod fit ex qm
eok existencia vadit intus in piramidem per primum theoreuma (/). Ergo in mn bis, nam ad utrumque horum refertur quadratum se in proporcione
linea eok est linea obliqua. Ponatur igitur in ipsa punctum m , et protrahatur una. Similiter quoque demonstrabimus quod quadratum eh est equale
perpendicularis supra triangulum abg, que sit mn. Manifestum igitur quod quadrato uc et ei quod fit ex uc in cx bis. Quare quadratum qm et illud quod
ipsa incidit supra distinctionem communem triangulo abg et superficiei fit ex qm in mn bis est equale quadrato uc et ei quod fit ex uc in cx bis. Quia
30 deh, que est dz, et incidit supra porcionem que est ex hac linea intra pira­ 65 ergo cx est maior mn per secundam partem quarti theoreumatis erit qm
midem post punctum e. Et dividamus lineam de in duo media supra punc­ maior uc. Igitur punctum c propinquius est linee sh quam punctum m,
tum s et ponamus proporcionem superficiei contente sub dn et ne ad quorum utrumque est super obliquam lineam ek. Similiter demonstrabimus
quadratum nm sicut quadrati se ad quadratum eh*, et protrahamus lineam de omni alio puncto existente super eandem obliquam lineam usque in
sh. Dico ergo quod linea sh et obliqua linea ek si protrahantur in infinitum infinitum, et hoc est quod demonstrare volumus.
35 numquam concurrent et tamen quanto magis prolongantur magis ap­
propinquant. Concurrant enim si possibile est in puncto o, et protra­ [LEMMA]
hamus a puncto o perpendicularem supra lineam dez, que sit of, et erit per­ * Modus inveniendi lineam ad cuius quadratum se habeat quecun-
pendicularis supra triangulum abg. Igitur ex quarto theoreumate erit pro- que superficies rectangula secundum proporcionem cuiuscunque dati
porcio superficiei contente sub d f et fe ad quadratum fo sicut superficiei quadrati ad aliud quadratum datum est iste.
40 contente sub dn et ne ad quadratum nm . Et ideo sicut quadratum se ad Inveniatur latus tetragonicum superficiei date, quo invento procedatur
quadratum he, quare et sicut quadrati s f ad quadratum fo . Itaque per pri­ 5 sic. Superficies data est equalis quadrato sui lateris tetragonici; ergo per
mam partem none quinti Euclidis superficies contenta sub d f et fe erit 7a[ml quinti ambo habent eandem proporcionem ad quodlibet tercium, ergo
equalis quadrato sf, utrumque enim ad quadratum fo habet eandem pro­ 8lr ad quadratum linee quam querimus; non habemus igitur aliud / querere
portionem, videlicet eam que est quadrati se ad quadratum eh. Sed hoc est nisi invenire quoddam quadratum ad quod se habeat unum quadratum
45 impossibile per 6 secundi Euclidis. Cum sit enim de divisa per equalia in 5 datum secundum proporcionem cuiuslibet alterius quadrati tercii dati ad
et addita ei linea in longum que est linea ef, necesse est ut quadratum s f 10 quartum quadratum datum. Hoc autem sic fiet. Proponantur duo quadrata
excedat superficiem d f in fe per quantitatem quadrati se. Quare patet pri­ inter que est proporcio data et substituatur tercium et querimus propor-
mum, videlicet quod impossibile est lineam sh et obliquam lineam ek con­ cionale. Deinde sic procedatur. Eadem est proporcio primi quadrati ad
currere.
secundum que lateris primi ad latus secundi proporcio duplicata. Deinde
50 Item dico quod quanto amplius protrahuntur magis appropinquant. Pro­ secundum proporcionem primi lateris ad secundum subiungam lineam
trahemus enim lineam mn usque ad q , et post punctum m signabimus punc­ 15 proporcionalem ad latus tercii quadrati, quod scio facere per 10 sexti. Dico
tum c super lineam obliquam ek, et protrahemus ab eo perpendicularem igitur quia linea inventa est quesita. Probo. Eadem est proporcio prime ad
ad lineam ez, que sit ucx. Et quia proporcio superficiei dn in ne ad secundam que tercie ad quartam; ergo eadem prime ad secundam duplicata
quadratum nm est sicut proporcio quadrati se ad quadratum eh*, et ideo que est tercie ad quartam duplicata, ergo et quadrati primi ad proporcionem
55 quadrati sn ad quadratum ng, erit per 19 quinti Euclidis proporcio eius quod que quadrati tercii ad quartum, et hoc est propositum.
addit quadratum sn super superficiem dn in ne, et ipsum est per 6 secundi
57 quadratum se ad et quadratum2 mg. add. m. 2 / [illud . . . addit] addidi;
22 de ei mg. scr. mg. m. 2 i. e. linee eok et in altero margine Nota quod linea cf. Vers. A
dez et linea eh et linea eok sunt in eadem superficie 57-58 [nq . . . est] addidi; cf. Vers. A
30 de hac linea mg. scr. m. 2 scilicet dz 64-65 de Quia . . . mn mg. inf. scr. m. 2 Nota quod linea ucx est maior quam linea
32-33 ponamus . . . eh mg. scr. m. 2 sicut docetur in fine huius folii qmn, similiter linea eh que comparatur seu correspondet linee ucx est maior
41 de quare . . . fo mg. scr. m. 2 et hoc ideo quia duo trianguli seh et sfo sunt quam linea eh que correspondet linee qmn, sicut enim crescit linea ucx super
equianguli dato quod linea sh recta concurrat cum linea ek in o lineam qmn similiter crescit linea eh correspondens linee ucx super lineam eh
47 supra superficiem df scr. m. 2 per 6 secundi correspondentem linee qmn
53-54 de proporcio . . . ehmg. scr. m. 2 sicut docetur in fine huius foliiet in altero mg. [Lemma]
per quartam secundi 4 supra Inveniatur scr. m. 2 per 14 secundi
54 supra proporcio scr. m. 2 per ypotesim 12-13 de Eadem . . . duplicata mg. scr. m. 2 per 18 sexti
55 qnti (?) P 15 10 corr. ex 12
 CHAPTER 3

William o f M oerbeke and W itelo

The few pittances of information regarding conic sections examined in

the first two chapters were greatly expanded with the appearance of the
translations by William of Moerbeke of the then extant works of Archi­
medes and of two of the three commentaries thereon by Eutocius, trans­
lations made in 1269 in the papal court at Viterbo.1The knowledge of conic
sections revealed in these translations is found in Archimedes’ On the
Equilibrium o f Planes, On the Quadrature of the Parabola, On Conoids
and Spheroids and On Floating Bodies, as well as in Eutocius’ Com­
mentary on the Sphere and the Cylinder o f Archimedes and his Commen­
tary on the Equilibrium o f Planes o f Archimedes. Since all of these works
have been published in my recent edition of the Archimedean transla­
tions of William of Moerbeke and since the Archimedean knowledge of
conic sections has been ably summarized by Heiberg, Zeuthen, Heath
and Dijksterhuis2, I shall in this chapter present only such material from
the translations that was possibly read and used by Moerbeke’s friend,
the Polish optician Witelo, in the treatment of conic sections in his Per­
spectiva, a work composed, I would judge, about 1270 or shortly there­
after. Indeed, except for the possible use of this material by Witelo, we
find little response to the rich Archimedean material on conic sections
found in Moerbeke’s translations until the time of Francesco Maurolico
and Federigo Commandino in the sixteenth century and I reserve treatment
of these scholars to the last chapter. We should also initially observe
that William of Moerbeke may well have provided Witelo with transla­
tions of Apollonius’ Conics and Eutocius’ commentary on it, as I shall
argue later.

1 See M. Clagett, Archimedes in the Middle Ages, Vol. 2: The Translations from the
Greek by William o f Moerbeke (Philadelphia, 1976).
2 Ibid., 17rA-20vD, 20vE-22vH, 34rA-44vP, 45rA-52vO, 53rA-55rW, 55vA-60rA.
Also see below the various passages given in note 6 from Moerbeke's translations of Euto­
cius’ two commentaries. For summaries of Archimedean conics, see J. L. Heiberg, “ Die
Kenntnisse des Archimedes iiber die Kegelschnitte,” Zeitschriftfiir Mathematik undPhysik,
XXV. Jahrg. (1880), Hist.-lit. Abtheilung, pp. 41-67; H. G. Zeuthen, Die Lehre von den
Kegelschnitten im Altertum (Copenhagen, 1886; repr., Hildesheim, 1966), pp. 39-63,
432-55 and passim ', T. L. Heath, Apollonius o f Perga: The Treatise on Conic Sections
(Cambridge, 1896; repr. 1961), pp. xli-lxvii, and E. J. Dijksterhuis, Archimedes (Copen­
hagen, 1956), pp. 55-118.

63
64 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 65

Let us now turn directly to Witelo’s treatment of conic sections in Regiomontanus, were printed in 1544 in a widely popular edition.4 One
the Perspectiva, first observing some generalities. It will be clear as our further, general caution must be given concerning the influence of the so-
analysis unfolds that Witelo most probably had some direct access to the called Greek material. Though Witelo’s work was widely read (and partly
Conics of Apollonius, either through the Greek text itself, which Moerbeke because of that), the Arabic material examined in Chapters 1 and 2 con­
may have helped him read, or, what is more likely, through a Latin trans­ tinued to exert a primary influence on medieval treatments of conic sec­
lation prepared by his friend. But it will also become clear that, if he did tions, as we shall see in Chapters 4 and 5 below.
read or consult the Conics directly, he was probably the only one before In his brief description of the contents of the Perspectiva with which
Regiomontanus to have done so3, and that, despite his reading of the he prefaces his work, Witelo mentions the Conics of Apollonius in the
Conics, Witelo depended significantly on the Apollonian fragment trans­ following terms:5
lated by Gerard of Cremona from the beginning of the Arabic text of the For wishing by mathematical demonstration to draw conclusions concerning
Conics, on Alhazen’sD e speculis comburentibus which accompanied the every visible entity, in so far as its properties affect its visibility, and to
fragment, and, most of all, on Alhazen’s Perspectiva. (He may also have proceed more certainly by this way, we have, to the best of our ability pro­
seen the De duabus lineis which I have edited in Chapter 2.) In fact, duced this book that is self-contained except for the [many] things which
these sources appear to have directed Witelo’s attention to those parts of derive from the Elements of Euclid and the few things that depend on the
the Conics that he deemed useful for his optical tract. Furthermore, I Conic Elements of Apollonius of Perga, of which latter I have used only
would judge that he read little else of the Conics than these pertinent two in this science, as will be evident later in what follows. And so in the
passages, and even in the case of these passages, he did not present first book of this science we premise axioms which are necessary for this
Apollonius’ proofs but satisfied himself with the mere indication of their science but are not in the Elements of Euclid. And in this [book] we make
clear those two [propositions] which have been demonstrated by Apollonius.
existence. But still my analysis will show that Witelo did introduce the
Greek terminology of conic sections (parabola, hyperbola, ellipsis, The two propositions to which Witelo here refers are his own Proposi­
asymptota etc.) that was later to persevere in Western mathematics; or, tions 1.129 and 1.131, which, as we shall see, in fact depend on at least
at least, he co-introduced it with William of Moerbeke, since the latter’s four propositions of Apollonius’ Conics (II.4, II.8, 1.14 and 11.16) and
translation of Eutocius’ Commentary on the Sphere and the Cylinder which Witelo does not undertake to prove. Furthermore, in this context
presented the same vocabulary. Though Witelo’s work was no doubt of a table of contents, he makes no reference to his own Propositions
more widely read by perspectivists and geometers than were Moerbeke’s IX.39 and IX.40, which are also presented without proofs but said by him
translations, it must be remembered that Moerbeke’s translations influ­ to be proved by Apollonius, these propositions being essentially Proposi-
enced the translations of Archimedes and Eutocius made in about 1450
by Jacobus Cremonensis and that the latter translations, as modified by 4 Clagett, Archimedes in the Middle A ges, Vol. 3 (Philadelphia, 1978), Part III, Chap. 2,
Sections I and II. For Moerbeke’s rendering of the Greek terminology of conic sections,
see note 6 below. Like Moerbeke, Jacobus Cremonensis merely transliterates the Greek
3 See the fifth chapter for a discussion of Regiomontanus’ knowledge of conic sections terminology for conic sections (see MS Venice, Bibl. Marc. f.a. 327, 29r et seq.), though
and the text of the Conics, where we show that Regiomontanus at least owned a copy of he prefers hyperbole and often parabole as the nominative cases (with consequent changes
the Conics. It should be noted that Apollonius was mentioned in the introduction by Hypsicles in the accusative and other cases) instead of Moerbeke’s hyperbola and parabola.
to the so-called fourteenth book of Euclid's Elements, available in Latin translation (see 5 Vitellonis Thuringopoloni opticae libri decem, ed. of F. Risner with Opticae thesaurus
J. E. Murdoch, “ Euclides Graeco-Latinus. A Hitherto Unknown Medieval Latin Translation Alhazeni Arabis libri septem, nunc primum editi (Basel, 1572), pag. of Witelo, p. 2; “ Prae­
of the Elements Made Directly from the Greek,” Harvard Studies in Classical Philology, sens itaque negotium decem libris partialibus duximus distinguendum. Volentes enim omne
Vol. 71 [1966], pp. 249-302 [whole article], particularly pp. 283, 285, 300 [n. 108]). Campanus ens visibile, ut suae visibilitati passio accidit, mathematica demonstratione concludere,
picks up this reference to Apollonius in his version of Euclid in Euclidis Megarensis . . . et hac via eatenus (ut nobis est possibile) certius ambulare: librum hunc per se stantem
Elementorum geometricorum libri xv (Basel, 1546), p. 452: “ Nunc ergo explicandum est effecimus, exceptis his, quae ex elementis Euclidis, et paucis, quae ex conicis elementis
quod ait Aristaeus in libro intitulato, Expositio scientiae quinque corporum: necnon et Apollonii Pergaei dependent, quae sunt solum duo, quibus in hac scientia sumus usi, ut
Apollonius in dono secundo in proportionalitate figurae 12 basium ad figuram 20 basium, in processu postmodum patebit. In primo itaque huius scientiae libro axiomata praemit­
dicens quod proportio superficierum figurae habentis 12 bases ad superficies figurae habentis timus, quae praeter elementa Euclidis huic scientiae sunt necessaria: et in hoc ea duo,
20 bases est tanquam proportio corporis 12 basium ad corpus 20 basium, linea etenim quae demonstrata sunt ab Apollonio, declaramus.’’ This text agrees almost exactly with
ducta a centro circuli pentagoni figurae 12 basium dodecedri ad circunferentiam eius est the manuscripts of the Perspectiva that I have checked. Other references to propositions
quasi linea prodiens a centro circuli trianguli figurae viginti basium icosedri ad circun­ found in Book I will be to S. Unguru’s edition (see note 8 below). Incidentally, Unguru
ferentiam eius. Haec sunt ipsius magni Apollonii verba." But this reference to Apollonius discusses the above-noted and other passages concerning conics in a useful but preliminary
is beside the point, since it is to the second edition of a lost work A Comparison o f the article, “ A Very Early Acquaintance with Apollonius of Perga's Treatise on Conic Sections
dodecahedron with the icosahedron and not to the Conics. in the Latin West,” Centaurus, Vol. 20 (1976), pp. 112-28.
 MOERBEKE AND WITELO 67
66 ARCHIMEDES IN THE MIDDLE AGES

est que HL, asymptote autem que ABK, et equedistantes ipsis AHN que MLX, equale est
tions 1.35 and 1.11 of the Conics. But Witelo explains in his Proposition quod A H N ei quod sub MLX per 8. theorema secundi libri elementorum conicorum Apol­
IX.40, at the end, that these latter propositions were not mentioned in lonii. Sed que quidem H N ipsi AB est equalis, que autem L X ipsi MB. Quod ergo sub
his opening description of the contents of the work because they were LMB est equale ei quod sub HAB\ et quia quod sub extremis est equale ei quod sub
needed only for the subsequent theorem (IX.41) and for no other theorems mediis, quatuor recte proportionales sunt. Est ergo ut que LM ad HA ita que AB ad BM.
in the work (see below, note 41). One further point to stress concerning Et ut ergo quod ab LM ad id quod ab HA ita quod ab AB ad id quod a BM. Et quoniam
propter parabolam quod ab LM est equale ei quod sub [Z]M, AH, est ergo ut que Z[A/]
Witelo’s opening description of the contents concerns the title by which ad ML ita que ML ad A H .” 42rH-K (see Fig. 3.5): “ Si igitur fecerimus ut D ad duplam
he refers to Apollonius’ work: [De] conicis elementis. It is the title he uses ipsius G ita ipsam CY ad aliam quandam ut F et circa ipsam CY descripserimus ellipsim
in all of his citations of Apollonius’ treatise. Now this is the first time in ita ut protracte in angulo qui sub XOB, hoc est, in medietate recti, possint [ea] que circa
our study that we have seen conicus used in reference to conic sections, F [deficientia] simili ei quod sub CY, F, transibit per X propter conversionem 20mi the­
orematis primi libri elementorum conicorum Apollonii. Sit descripta et sit ut que YXC.
and it is obvious that Witelo's title comes from the Greek— earlier the
Signum ergo X tangit positione datam ellipsim. Et quoniam diagonalis est que L [K ] ipsius
title had been Liber de pyramidalibus (or piramidalibus) or Liber pyra­ NM parallelogrammi, equale est quod sub NXP ei quod sub ABM. Si igitur per B circa
midum (or piramidum). The title adopted by Witelo must ultimately have asymptotas T[K]M descripserimus yperbolam, transibit per Z et erit positione data propterea
come from Κωνικά στοιχεία rather than from Κωνικά alone. The former quod et signum B positione datum est et utraque ipsarum AB, BM et propter hoc circa
appears many times as the title of Apollonius’ work in Eutocius’ Com­ asymptotas T[K]M. Sit descripta et sit ut que XB. Ergo signum X tangit positione datam
mentary on the Sphere and the Cylinder o f Archimedes and William of yperbolam; tangebat autem et positione datam ellipsim. Datum est ergo X. Et ab ipso per­
pendicularis que XE; datum est ergo E. Et quoniam est ut que MB ad BE ita que ZA ad
Moerbeke always rendered it as Liber elementorum conicorum.6 But the AE, et data est que AE, data est ergo et que AZ. Propter eadem etiam data est et que
HB.” 42rO-P (see Fig. 3.5): “ Quoniam igitur yperbola est que BX, asymptote autem que
T[K], [A]M, equale est quod sub NXP ei quod sub ABM per 8um theorema secundi libri
y Clagett, Archimedes in the Middle Ages, Vol. 2. The following references are taken
elementorum conicorum Apollonii; et propter hoc recta est que [K]EL. Iaceat igitur ipsi
from Moerbeke's translation of Eutocius’ Commentary on the Sphere and the Cylinder
quidem TA equalis que AZ, ipsi autem LB equalis que BH. Quoniam igitur est ut dupla
of Archimedes: 40vB -D (see Fig. 3.1): “ Equale ergo quod sub MHZ ei quod aZ A . Si ergo
ipsius G ad O ita que F ad CY, ut autem que E ad CY ita quod sub COY ad id quod ab XO
circa axem ZH scribatur per H parabola ut producte possint penes HM, veniet per K,
propter 20um theorema primi libri elementorum conicorum Apollonii, ut ergo dupla ipsius
et erit positione data quia data est que HM magnitudine continens cum HG data datum D.
G ad D ita quod sub COY ad id quod ab ΛΌ.” 42vH -K (see Fig. 3.6): “ Qualiter autem
Ipsum ergo K tangit positione datum parabolam. Sit igitur descripta ut dictum est et sit
oportet per datum signum circa datas asymptotas describere yperbolam ostendemus sic,
ut que HK. Rursum, quoniam spatium TL est equale ipsi GB, hoc est, quod sub TKL
quoniam non expresse [ponitur] in conicis elementis. Sint due recte que GA, AB quem­
ei quod sub ABH, si per B circa asymptotas, scilicet TG, GH, [scribatur] yperbola, veniet
cunque angulum continentes eum qui apud A , et datum sit signum aliquod D et adiaceat
per K propter conversionem 8V' theorematis secundi libri elementorum conicorum Apol­
per D circa asymptotas GA, AB describere yperbolam. Copuletur que AD et educatur ad
lonii.’’ 40vK -L (see Fig. 3.2): “ Equale ergo est quod ab HM ei quod sub HZN. Si ergo
E, et iaceat ipsi DA equalis que AE, et per D ipsi AB equedistans ducatur que DZ, et
per Z circa axem ZH descripserimus parabolam ut protracte possint penes ZN , veniet
iaceat ipsi AZ equalis que ZG, et copulata que GD educatur ad B, et quod a GB sit equale
per M. Sit descripta et sit que MXZ. Et quoniam equale est quod TL ipsi AZ, hoc est,
ei quod sub DE, H, et educta ipsa AD describatur circa ipsam per D yperbola ita ut pro­
id quod sub TKL ei quod sub ABZ, si per B circa asymptotas ipsas TG, GZ descripserimus
tracte possint que circa H [excedentia] simili ei quod sub DE, H. Dico quod descripte
yperbolam, veniet per K propter conversionem 8V' theorematis elementorum conicorum
yperbole sunt asymptote que GA, AB. Quoniam enim equedistans est que DZ ipsi BA
Apollonii." 40vQ-U (see Fig. 3.3): “ Sit enim descripta per H circa axem ZH parabola
et equalis que GZ ipsi Z A , equalis ergo et que GD ipsi DB. Quare quod a GB est quadruplum
ita ut protracte possint penes HM\ veniet itaque per A, ut in resolutione demonstratum
eius quod a GD. Et est quod a GB equale ei quod sub DE, H\ utrumque ergo eorum que
est, et concidet educte (!) ipsi TG equedistanti existenti diametro sectionis propter 27.
a GD, DB est quarta pars speciei que sub DE, H. Que ergo GA, AB sunt asymptote ipsius
theorema primi libri elementorum conicorum Apollonii. Sit educta et concidat apud N,
yperbole propter primum theorema secundi libri elementorum conicorum Apollonii." In
et per B circa asymptotas NGH sit descripta yperbola. Veniet ergo per K, ut in resolutione
addition to the foregoing references that gave the longer title to Apollonius’ work, there
dictum est. Si igitur protracta ut que BK et educte ipsi ZH equalis iaceat que HX, et
are some in Eutocius’ Commentary on the Equilibrium o f Planes o f Archimedes, which
copuletur que XK, et educatur ad O. Manifestum ergo quod attingit parabolam propter
give the shorter title: 53rL (see Fig. 3.7): “ In yperbola autem DEZ centrum figure est
conversionem 34. theorematis primi libri elementorum conicorum Apollonii. Quoniam
extra, secundum quod diametri concidunt invicem, sicut se habet T. Dicta sunt enim hec
igitur que BE est dupla ipsius EA, sic enim subponitur, hoc est, que ZK ipsius KT, et est
in secundo libro conicorum Apollonii.” 53vF-G (see Fig. 3.8): “ Quoniam igitur obscurum
simile trigonum OTK trigono XZK, [est] et que XK dupla ipsius KO. Est autem et que
est quod dicitur, necessarium dicere brevia de ipso ex conicis AppoIIonii (/) inventa. Sit
XK dupla ipsius KP, propterea quod et que XZ ipsius XH et quia equedistans est que
figura contenta a parabola ABG et recta AG, cuius diameter sit que BD. Manifestum itaque
PH ipsi KZ. Equalis ergo que EK (/ O K ) ipsi KP. Que ergo OKP contingens yperbolam
quod vertex portionis est signum B; vertices enim linearum vocabat Apollonius que apud
et intermedia existens asymptotarum in duo equa secatur. Attingit ergo yperbolam propter
lineas ultima diametrorum.” 53vR-S: “ Similes portiones sectionum coni Apollonius dif-
conversionem tertii theorematis secundi libri elementorum conicorum Apollonii." 41vB -E
finivit in sexto libro conicorum, in quibus productis in unaquaque equedistantibus basi
(see Fig. 3.4): “ Et circa axem ZB per Z describatur parabola ita ut protracte possint penes
equalibus multitudine equedistantes et bases ad abscisas a diametris apud vertices in eisdem
AH: transibit ergo per T, quoniam quod sub ZAH est equale ei quod ab AT. Sit igitur
proportionibus sunt et abscise ad abscisas, et quod parabole omnes sunt similes.” Note
descripta et sit ut que ZTK, et per B protrahatur equedistanter ipsi AT que BK et secet
two further references citing the short title in the marginalia taken from the scholia (see
parabolam penes K, et per H circa asymptotas ZB, BK sit descripta yperbola. Secabit
the variant readings, 46rC and 47rJ) and one reference to the longer title that intruded into
itaque parabolam inter T et K: secet penes L, et ab L super AB perpendicularis ducatur
the Greek (and thus Latin) text of On Floating Bodies (58rF-G). The first two references
que LM, et per H, L ipsi AB equedistantes ducantur que HN, LX. Quoniam igitur yperbola
68 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 69

shorter title Κωνικά (which appears in Moerbeke’s translation of Eutocius’ erators of a right cone and holds that they make equal, acute angles with
Commentary on the Equilibrium of Planes as Liber conicorum) is the radii of the base circle, that the apex of the cone is a pole of the base cir­
only title found in the text of Apollonius’ Conics and Eutocius’ com­ cle, that any generator of such a cone is in the same plane with the axis
mentary on it (at least it is the only title found in Vat. gr. 203, which of the cone, and finally that the axis of the cone is perpendicular upon the
was in all likelihood the manuscript of these two works that Moerbeke center of the base. Now the reader will remember that the cone under­
and Witelo would have consulted, as I shall note later). I have already stood in the anonymous De duabus lineis of Chapter 2 was always a right
shown elsewhere that Witelo almost certainly read and used Moerbeke’s cone, and so it is here with Witelo, who continually returns to this proposi­
translation of Eutocius’ Commentary on the Sphere and the Cylinder o f tion later when he wishes to employ the fundamental properties of a right
Archimedes.7 Hence I would now suppose that he read that work before cone. In his proof of the various statements that appear in the enuncia­
looking at the Conics and, since the longer title is cited so frequently in tion of 1.89, Witelo depends primarily on the Euclidian definition of a right
Eutocius’ Archimedean commentary (see note 6), Witelo fell into the habit cone that describes the generation of a right cone by the rotation of a
of calling Apollonius’ work by the longer title De conicis elementis and right triangle, the same definition with which the De duabus lineis began
continued to do so even after seeing the shorter title. (see above, Chap. 2, note 6). But Witelo adds the rest of Euclid’s defini­
Though we must soon examine the various propositions in which Witelo tion that is lacking in the De duabus lineis, namely, that if the fixed side of
acknowledges dependence on Apollonius, before we do so we should ex­ the generating triangle is equal to the moving side (these two sides being
amine those Witelonian propositions that reflect elementary propositions the sides containing the right angle), a right-angled cone is described;
of the Conics (but which, for the most part, were probably not drawn but if the fixed side is less than the moving side, an obtuse-angled cone
directly from the Conics) and other non-Apollonian propositions that had is described; and finally if the fixed side is greater than the moving side,
become absorbed into the medieval treatment of conic sections. The first an acute-angled cone is described.9 It was such right cones that the early
of these is Witelo’s Proposition 1.89.8 It asserts the equality of the gen- Greek geometers (the “ Ancients” as Geminus calls them) first began to
study.10 Witelo’s dependence on the Euclidian definition of a right cone
are respectively to Propositions III.20 and 1.20 of the Conics, the last to 1.7 of the Conic separates his treatment from that found in Proposition 1.4 of the Conics
Elements.
7 Clagett, Archimedes in the Middle Ages, Vol. 2, pp. 13-27. of Apollonius (see above, Chap. 2, note 7).
8 S. Unguru, Witelonis perspectivae liber primus. Book I o f Witelo's Perspectiva (Wroc­ Witelo’s next proposition (1.90) tells us that if a plane cuts the vertex
law, Krakow, 1977), pp. 268-69. “ [Propositio] 89. Omnes linee longitudinis unius pyramidis of a [right] cone and contains a generator of the cone, the common sec­
rotunde sunt equales et cum semidyametris basis equales sed acutos angulos continentes; tion is a triangle contained by two generators and a diameter of the base
ex quo patet omnem punctum verticis pyramidis esse polum circuli sue basis; omnemque circle, and that triangle bisects the cone and its base and contains
lineam longitudinis esse in eadem superficie cum axe; ipsum quoque axem centrum circuli
the axis of the cone (we have ignored here and in other propositions the
basis orthogonaliter attingere. Quoniam enim, per principium undecimi Euclidis, pyramis
rotunda fit per transitum trianguli rectanguli alterutro suorum laterum rectum angulum con­ parallel statements about a right pyramid).11 It is just such an axial triangle
tinentium fixo donec ad locum suum unde incepit redeat, triangulo ipso circumducto; qui
triangulus si fuerit duorum laterum equalium, et unum laterum equalium rectum angulum
et prima parte presentis, axe existente communi, omnes anguli ad centrum B constituti
continentium fuerit fixum causabitur pyramis rectangula, ideo quod angulus duplicati sui sunt equales. Patet ergo propositum.” Cf. pp. 110—III for Unguru's English translation.
trianguli ad verticem pyramidis est rectus, per 5am et 32am P. Et si fixum latus fuerit minus See also ed. of Risner, pp. 34-35. In reproducing Unguru’s Latin text, I have italicized
latere moto, erit pyramis ampligonia, quoniam, per 19am I1, angulus ad verticem sit obtusus; the letters marking magnitudes, changed the punctuation slightly, and omitted the para­
et si latus fixum fuerit maius latere moto, erit pyramis oxigonia, quia, per eandem 19am graphing. This is true for all of my succeeding quotations from Unguru’s text. I have also
Γ, angulus eius ad verticem remanet acutus, adiuvante semper 32a 1', Sic ergo diversantur diverged somewhat from Unguru’s translation, translating as I have seen fit.
forme pyramidum secundum diversitatem proportionis lateris fixi ad alterum latus motum
9 See the Campanus version of the Elementa (Basel, 1546), p. 346.
rectum angulum continens cum fixo. Et quia latus subtensum angulo recto causat omnes 10 See below, note 20, and T. L. Heath, A History o f Greek Mathematics, Vol. 2 (Oxford,
lineas longitudinis in qualibet pyramide, palam quod omnes linee longitudinis totius rotunde
1921), p. 111.
pyramidis uni linee sunt equales ei, scilicet que in trigono rectangulo opponitur angulo 11 Ed. of Unguru, p. 269: “ [Propositio] 90. Omnis superficiei plane secantis pyramidem
recto. Ergo et omnes inter se sunt equales. Si ergo trigonum orthogonium causans pyramidem rotundam vel lateratam secundum axis longitudinem et superficiei conice communis sectio
sit ABC [Fig. 3.9], cuius angulus ABC sit rectus, erit, per 32am Γ, angulus ACB acutus; est trigonum duabus lineis longitudinis pyramidis et dyametro basis contentum; ex quo
et est ACB angulus cui omnes anguli contenti a lineis longitudinis et semidyametris basis patet, quoniam illa superficies dividit pyramidem per equalia, et quod superficies que pyra­
sunt equales, et hoc proponebatur. Patet enim ex hiis quoniam punctus verticis pyramidis midem secundum lineam longitudinis per equalia secuerit, secundum axem necessario
cuiusque est polus circuli sue basis, per 65am huius. Et quoniam linea AC est in eadem secabit. Esto pyramis rotunda ABC [Fig. 3.10], cuius vertex A et dyameter basis B C , et
superficie trigona cum linea AB, patet quod omnes linee longitudinis sunt in eadem super­ sit centrum basis D. Et palam, per premissam, quia linea AD est axis illius pyramidis.
ficie cum axe AB. Et quoniam linea BC motu suo describit circulum basis, patet quod Superficies itaque plana secans pyramidem rotundam secundum axis longitudinem pertransit
axis AB centrum circuli basis orthogonaliter attingit, per 8am l·, quia ex circuli diffinitione puncta A et D. Erit itaque illa superficies plana orthogonaliter erecta super basim pyramidis,
70 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 71

that was used repeatedly in the De duabus lineis and that will continue from the generation of cones on either or both sides of a double conic
to appear in the Speculi almukefi compositio edited and analysed in the surface described in Proposition 1.4 of the Conics, except that the figure
next chapter. The fact that the proposition concerns a right cone dis­ for a double cone given by Witelo (see Fig. 3.11) has some of the same
tinguishes it from the more general proposition that was proved by Apol­ letters in the same positions as the similar figure given by Apollonius
lonius in Proposition 1.3 of the Conics·. “ If a cone is cut by a plane in the proof of that proposition (see above, Chap. 2, note 7). If he saw
through the apex the section is a triangle.” 12 Similarly the dependence Apollonius’ proposition and proof, Witelo strayed from it, no doubt be­
in Witelo’s proof on Euclid’s description of the generation of a right cone cause he had started his discussion of cones in 1.89 with a single right
distinctly separates it from Apollonius’. Finally, I should observe that cone, generated in the manner described by Euclid.
Witelo was also familiar with the role played by such an axial triangle The next Witelonian proposition to be considered, 1.98, is so important
in the statements concerning the generation of conic sections found in to my investigation that I must present it fully in translation;14
the fragmentary translation of Gerard of Cremona (see above, Chap. 1,
note 18). 98. The common section of any plane surface that cuts a [right] cone not
In Proposition 1.91 Witelo treats of a double cone.13 The upper cone through the apex and of the surface of the cone cannot be a triangular figure.
he fashions by extending all of the generators of the lower cone beyond Let there be a [right] cone whose apex [is] A [see Fig. 3.12], [whose]
its apex. This distinguishes his treatment from the description of a double diameter of the base [is] BC , [whose] center of the base [is] D , and [whose] axis
conic surface given by Apollonius in his first definition, which Witelo [is] AD, which [cone] is cut by a plane surface through the axis, according to
triangle ABC, by the 90th [prop,] of this [book]. And let another [plane] surface
certainly knew at least from Gerard of Cremona’s fragmentary translation
from the Arabic (see above, Chap. 1, note 5) if not from the text of the
u Ed. of Unguru, pp. 273-74: ‘‘[Propositio] 98. Omnis superficiei plane secantis pyrami­
Conics itself. His treatment in Proposition 1.91 must also be distinguished
dem rotundam non per verticem et superficiei conice pyramidis communem sectionem
figuram triangularem esse impossibile. Esto pyramis cuius vertex A [Fig. 3.12], dyameter
per 18am XI'. Communis itaque sectio basis pyramidis et illius superficiei plane est linea basis BC, centrum basis D, et axis AD, quam secundum axis longitudinem secet superficies
recta, per 3am XI', que est dyameter basis, et sit hec BC. Trigonum itaque ABC est in plana secundum trigonum ABC, per 90am huius. Secetque ipsam alia superficies erecta
superficie secante. Sed et idem trigonum est in superficie conica pyramidis. Et quoniam super trigonum ABC, non per verticem, secundum sectionem que sit EFG, cuius supremus
trigonum orthogonium BAD est illud ex cuius pertransitu describitur pyramis ABC, et punctus sit F, et sit linea EG equidistans alicui dyametro basis pyramidis, cuius medius
trigonum ABC est duplum illi, per l am VT, patet illud quod primo proponitur de pyramide punctus sit H; et ducatur linea FH a supremo puncto sectionis ad medium sue basis. Et
rotunda. Patet etiam quod illa superficies taliter pyramidem secans dividit ipsam per equalia, quia linea EG est linea recta que est equidistans dyametro basis pyramidis, et punctum
quoniam transiens verticem et conclusa dyametro per equalia dividit et basim. In laterata F signatum est in superficie conica, ubi in supremo, superficies EFG secat conicam super­
vero pyramide, aut superficies plana secans transit latus aut angulum; eritque, productis ficiem. Si itaque sectio EFG sit trigonum rectilineum, patet quod (correxi ex quoniam)
lineis ad terminum axis pyramidis, illa communis sectio semper trigonus maior vel minor. due linee longitudinis pyramidis, que sunt EF et GF, concurrunt in puncto F preter verticem
Patet ergo propositum, quoniam et conversa per se et ex premissis patet.” Cf. pp. 111-12 pyramidis, quod est impossible et contra 91am huius. Trigonum quoque curvilineum fieri
for Unguru’s translation. See also ed. of Risner, p. 35. est impossibile, quoniam superficies secans supponitur esse plana, et superficies illius
12 Ed. of Heiberg, Gr 1, p. 12; Taliaferro, op. cit. in Chap. 1, note 19, p. 606. trigoni est curva, ut patet ex diffinitione. Erit ergo linea EFG linea una. Cum itaque illa
13 Ed. of Unguru, pp. 269-70: ‘‘[Propositio] 91. Omnis pyramidis rotunde vel laterate sectio sit linea una dicatur sectio conica vel pyramidalis. Si itaque axis pyramidis, qui
linee longitudinis super axem in vertice tantum se intersecant; producte quoque aliam similem est AD, sit equalis semidyametro basis, que est DB, palam quia pyramis ABC est ortho­
pyramidem principiant cuius linee longitudinis, secundum positionem et situm, priori pyra­ gonia, quoniam angulus BAC trigoni ABC est rectus. Si ergo linea FH, que est communis
midi modo contrario se habent. Quod (Risner, Quoniam Ung.) omnes linee longitudinis sectio superficiei EFG et trigoni ABC, equidistet linee AC, que est latus trigoni et linea
pyramidis cuiuscunque producte se super axem in vertice secent evidens est, quoniam longitudinis pyramidis, palam, per 29am Γ, cum angulus BAC sit rectus, quod etiam angulus
concurrent omnes in illo puncto verticis. Et quoniam omnes sunt equales. per 89am huius, BFH erit rectus, et similiter angulus HFA. Tunc itaque sectio EFG dicetur sectio rectangula,
patet quia citra verticem nulla ipsarum aliam intersecat. Quod etiam producte aliam pyra­ vel parabola, et est illa quam Arabes dicunt mukefi. Si vero linee HF et AC non equidistent,
midem priori similem principiant, patet. Secet enim superficies plana pyramidem secundum sed concurrant, si concursus fiat ad partem puncti A, qui est vertex pyramidis, tunc patet,
axis longitudinem, erit ergo, per precedentem, communis sectio istius superficiei et super­ per 14am huius, quod (Risner; quoniam Ung.) angulus HFA erit obtusus, et tunc sectio
ficiei conice pyramidis trigonum equum duplo trigoni rectanguli pyramidem causantis. Sed EFG dicetur ampligonia, vel yperbole, vel mukefi addita; si vero linee HF et AC concur­
palam, per 36am huius, quod latera cuiuslibet trigoni producta principiant alium trigonum rant versus punctum C, qui non est vertex pyramidis, tunc, per 14am huius, erit angulus
priori simile, cuius latera positionem et situm prioris trigoni lateribus contrariam habent. HFA acutus, et tunc sectio EFG dicetur oxigonia, vel ellipsis (Risner; elipsis Ung.), vel
Et quoniam tot possunt ymaginari plane superficies trans axem pyramidem secantes quot mukefi diminuta. Et secundum hunc modum iste sectiones et earum passiones amplissime
sunt linee longitudinis ymaginabiles in medietate pyramidis, patet quod omnes linee longi­ variantur.” I have changed the punctuation rather more than usual in this proposition,
tudinis producte principiant aliam pyramidem priori similem, lineis longitudinis a dextro and I have not italicized mukefi. Note that Unguru here and in his article, “ A Very Early
prioris prodeuntibus in sinistrum posterioris, et a sinistro prioris in dextrum posterioris, Acquaintance,” p. 122, first read mukefi as mukesi but on reexamining the manuscripts
et econverso [Fig. 3.11]. Patet ergo propositum.” Cf. p. 112 for Unguru’s translation. he has found that mukefi is the correct reading, as it should be. Cf. pp. 115-16 of his
See also ed. of Risner, p. 35. ed. for Unguru’s translation. See also ed. of Risner, p. 37.
72 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 73

erected [perpendicularly] on triangle ABC cut the same [cone] not through the rectilinear or curvilinear. It is not rectilinear because, if so, the lines EF
apex, according to the section EFG, whose highest point is F, and let line EG, and FG would at the same time be straight lines and lie in the surface of
whose middle point is H, be parallel to some diameter of the base of the cone; the cone and thus in effect would be generators; hence, by his definition
also let line FH be drawn from the highest point of the section to the middle of of the generators of a cone, they would pass through the apex A, as well
its base. And since line EG is a straight line parallel to the diameter of the base as through F, which is impossible since it was posited that the section
of the cone, and the point marked F is above in the surface of the cone, plane
does not pass through the apex. Nor can section EFG be a curvilinear
EFG cuts the surface of the cone. And so, if section EFG is a rectilinear
[and curviplanar] triangle, for by the definition of such a triangle its surface
triangle, it is plain that two lines of length [i.e. generators] of the cone, which
are EF and GF, meet in point F outside of the apex of the cone, which is must be curved and yet it was posited in this proposition that the section
impossible and against the 91st [prop.] of this [book]. Further, the triangle is formed by a plane surface that cuts the cone and consequently the
cannot be made curvilinear, because the intersecting surface is assumed to be a section’s surface must be a plane surface. Having shown that the figure
plane, and the surface of that triangle is curved, as is clear from the definition is not triangular and in effect that EF and FG are not straight lines, he
[of such a triangle]. Therefore, line EFG will be one line. And so, since that concludes, not without ambiguity, that [curved] line EFG is a single
section is one line, it may be called a conic or pyramidal section. And so, if the [continuous] line which may be called a conic or pyramidal section.
axis of the cone, which is AD, is equal to the radius of the base, which [radius] After completing this proof, Witelo includes a brief excursus in which
is DB, it is evident that the cone ABC is right-angled, because the angle BAC of he constructs the three sections and identifies each of them with three
triangle ABC is right. Hence, if line FH, which is the common section of equivalent names: (1) sectio rectangula = parabola = mukefi, (2) sectio
surface EFG and of triangle ABC, is parallel to line AC, which is a side of the
ampligonia = yperbole = mukefi addita, and (3) sectio oxigonia = ellip­
triangle and a line of length [i.e. a generator] of the cone, it is evident by 1.29 [of
the Elements of Euclid], since angle BAC is right, that angle BFH will also be
sis = mukefi diminuta. Now the source of the Arabic names is certainly
right, and similarly angle HFA. In such a case then, the section EFG will be the fragment translated by Gerard of Cremona from the Arabic text of
called a right-angled section or parabola, and it is that [section] which the the Conics (see above, Chap. 1, note 21). While this is undoubtedly so,
Arabs call mukefi. If now lines HF and AC are not parallel, but meet in the we should observe that Witelo has committed errors in the Arabic names
direction of point A , which is the apex of the cone, then it is evident, by the 14th for a hyperbola and an ellipse by inserting mukefi before addita and
[prop.] of this [book], that angle HFA will be obtuse; and then section EFG diminuta, 15 It is abundantly clear that Witelo had no idea as to the force
will be called an obtuse-angled [section], or yperbole, or mukefi addita·, but if of mukefi. Though Gerard of Cremona himself did not know the proper
lines HF and AC meet toward point C, which is not the apex of the cone, then, Latin term for mukefi, he clearly revealed that the term was not in the
by the 14th [prop.] of this [book], angle HFA will be acute; and then the section Arabic names for hyperbola and ellipse. The source of the other names
EFG will be called an acute-angled [section], or ellipsis, or mukefi diminuta. for the conic sections given by Witelo is less easily and surely determined.
And according to this method [of generation] these sections and their
properties vary greatly. It is clear, however, that they are all of Greek origin and that we should
accordingly look toward one or more Greek sources. We should particu­
The actual enunciation of this proposition and the first part of the proof, larly look for a Greek source that contains alternate names, and indeed
namely, that the section produced by a plane not through the apex is not there are three possibilities of just such a source being available to Witelo,
a triangular figure (either rectilinear or curvilinear), is completely original although to be sure each of these sources contains the Pre-Apollonian
with Witelo, at least in form if not substance. Of course, the conclusion and Apollonian names, while Witelo gives (as I shall presently show)
is substantially found in that part of the fragment translated by Gerard a distinctly modified form of the Pre-Apollonian names alongside those
of Cremona that was written by the Arabic editor of the Conics: “ When of Apollonius.
a cone is cut by a plane that does not pass through the apex, then the The first source which identifies the Pre-Apollonian and Apollonian
common section is a surface which a curved line contains” (see above, names for sections, and which was almost certainly available to and used
Chap. 1, note 18). This same statement is found in the Proem of the by Witelo, is William of Moerbeke’s Latin translation of Eutocius’ Com­
De duabus lineis (see above, Chap. 2). Furthermore, the proof that the mentary on the Sphere and the Cylinder o f Archimedes.16 In that work
line is a curved line is guaranteed in Proposition 1 of the De duabus lineis, Eutocius speaks of having discovered in an old book the solution of a
where it is proved that the straight line connecting any two points on the
surface of a cone that are not in line with the apex must fall within the 15 See the discussion of these Arabic terms in Chap. 1 above. I first thought that perhaps
some scribe, and not Witelo, had introduced the intrusive mukefi. But Unguru, who kindly
cone, that is, not on the surface of the cone. This proposition does indeed
checked the manuscripts for me, tells me that they all have the false mukefi before both
assure that the section cannot be a triangle. But it is clear that Witelo addita and diminuta. Hence it is clear that Witelo thought they belonged there.
does not actually take this line of proof but rather presents an indirect 16 See note 7 above. See also, above note 6, my discussion of Witelo’s apparent bor­
proof to the effect that if the section were triangular, it must be either rowing from Eutocius' work of the longer title De conicis elementis for Apollonius' treatise.
74 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 75

problem subsidiary to Proposition II.4 of Archimedes’ On the Sphere and [when compared to an exact application to the parameter] by a square [i.e.
the Cylinder, a solution that he believes may have been based on the a quadrilateral rectangle], while in the section of an obtuse-angled cone this
missing solution promised by Archimedes because the theorems of that area is in excess (νπερβάλλον) [of that exact application] by a square,
old book partly conserved the old Doric dialect affected by Archimedes and in the right-angled cone it is neither deficient nor in excess.
and retained “ the names familiar in old usage, the parabola being called Now I expect that the reader who compares Witelo’s Proposition 1.98
a section of a right-angled cone, and the hyperbola a section of an obtuse- with this passage from Pappus’ Collectio will agree with me that it is very
angled cone.” 17 The case of the ellipse is missing in this passage, but its doubtful that Witelo was influenced by the latter. Not only is the order
identity with a section of an acute-angled cone is suggested in a marginal of presenting the three sections different in Pappus’ account (i.e., ellipse,
comment by Moerbeke (see note 17). parabola, hyperbola), but, more important, there is no trace in Witelo’s
A second possible source containing the Pre-Apollonian and Apollonian proposition of Pappus’ explanation of the Apollonian names by means
terms for conic sections is found in a passage of Book VII of Pappus’ of the application of areas. This apparent lack of influence on Witelo of
Collectio, some part of Book VI of which appears to have been read by this passage of Pappus’ work is reinforced by a similar lack of influence
Witelo (see note 19 below). In this passage,18 Pappus first notes that Apol­ on the Polish mathematician of the long section in Book VII that includes
lonius brought to completion the four books on Conics of Euclid and added lemmas useful for the propositions of all eight books of Apollonius’ Conics
four additional books. Pappus also mentions the attention to conics in (see note 18). The probable inference from the overall lack of influence
Aristaeus’ Solid Loci and says that “ [like the predecessors of Apollonius,] of Pappus’ treatment of conic sections on Witelo is that Witelo had not
he (Aristaeus) calls one of the conic sections the section of an acute- read Book VII of the Collectio.
angled cone, the second the section of a right-angled cone and the third A third possible source identifying the Pre-Apollonian and Apollonian
the section of an obtuse-angled cone.” (Note that the bracketed phrase names is Eutocius’ Commentary on the Conics o f Apollonius, which
is in the earliest extant Greek manuscript but is believed by Hultsch to together with the Conics might have been available to Witelo in the form
be an interpolation.) He (Pappus) goes on to say that Apollonius could of a translation made by William of Moerbeke from Vat. gr. 203.19 In this
not see why his predecessors had so limited the names of the conic sec­
tions, since in fact each could arise from cutting any one of the three kinds 19 Archimedes in the Middle Ages, Vol. 2, p. 27, n. 21. I argue there (and on p. 10, n. 35)
of cone, the different sections being determined by the different modes in a preliminary fashion that extant manuscript Vat. gr. 203 might well be identified with
item no. 425 in the Vatican catalogue of 1295 and item 620 of the Vatican catalogue of
of cutting the cone. Consequently, he changed the names: “ he called the 1311. This manuscript contained texts of both the Conics and Eutocius’ commentary thereon
section customarily spoken of as the section of an acute-angled cone an and seems to be the only possible candidate among known manuscripts for an exemplar
‘ellipse,’ that spoken of as the section of a right-angled cone a ‘parabola’ that might have been available to Moerbeke, if indeed the Flemish Dominican did make
and that spoken of as the section of an obtuse-angled cone a ‘hyperbola,’ a translation of these two works. I think the most likely time for Moerbeke to have com­
each [of his names being adopted] from a particular property.” In this pleted such translations was in the year 1270, or early 1271, for no other translation can
be dated to the period between the completion of his translation of Ptolemy’s De analem-
connection he notes further that: mate in early 1270 and the completion of his translation of Simplicius’ Commentary on
a certain area [i.e. a rectangle] applied to a certain straight line [i.e. the the De caelo of Aristotle on 15 June, 1271 (see ibid., p. 6). I do not think that the translations
parameter] in the section of an acute-angled cone becomes deficient (kXkeinov) of Apollonius’ Conics and Eutocius' commentary could have been done before 1269-70,
for Moerbeke was still groping for the proper way to render in Latin the various terms
17 Clagett, Archimedes in the Middle Ages, Vol. 2, 40rR-S: “ In quodam tamen antiquo relating to conics when he made his Archimedean translations in 1269 (ibid., p. 38). It is
libro, neque enim ab inquisitione in multa destitimus, occurrimus theorematibus scriptis possible that Moerbeke also translated, about the same time (1270-71), some part of Pappus’
non modicam quidem obscuritatem habentibus ex corruptionibus et circa descriptiones Collectio and a fragment of Anthemius’ wepi παραδόξων μηχανημάτων from MS Vat.
multis modis peccantibus, quesitorum tamen habens subsistentiam, in parte autem linguam gr. 218, for it appears that Witelo had consulted both of these works (see Clagett, Archimedes
Doricam Archimedi amicam salvans et consuetis antiquo rerum nominibus inscribebatur, in the Middle Ages, Vol. 3, p. 406, n. 56, where I have noted Unguru’s discovery that
parabola quidme rectanguli coni sectione nominata, yperbole (/) autem amblygonii coni Witelo had access to Book VI of Pappus’ work in Propositions 1.22, 1.38 and 1.39 of the
sectione; ut ex ipsis cogitetur ne forte et ipse sit qui in fine promissa scribit.” I added Perspectiva and have also remarked on Witelo's use of Anthemius’ work in Proposition
the exclamation point after yperbole because ordinarily Moerbeke used yperbola as the V.65, ed. of Risner, p. 223). Cf. S. Unguru, “ Pappus in the Thirteenth Century in the
nominative form (see note 6, passages 40vB -D , 40vR -U , 41vB -E , 42rO-P, 42vH -K ). Latin West,” Archive for the History o f Exact Sciences, Vol. 13, No. 4 (1974), pp. 307-
Notice that Witelo also used the more literal transcription yperbole in Proposition 1.98. 324. However, as I have already mentioned in the text above and the preceding footnote,
See also ibid., p. 414, variant for 42rL: “ ellipsis, ut puto, dicitur oxygonii coni sectio (?).” there is no evidence that Witelo read or mastered the material on conic sections in Book VII
1HPappi Alexandrini Collectionis quae supersunt, ed. of F. Hultsch, Vol. 2 (Berlin, 1877; of Pappus' Collectio. Nor is there any evidence in Witelo’s work of his having mastered
repr. Amsterdam, 1965), pp. 672-74. See also pp. 918-1014 for the long section of Book VII the interesting conclusions relating to conics in Anthemius’ fragment. For the latter, see
that includes lemmas for all eight books of Apollonius’ Conics, and certain other lemmas, T. L. Heath, “ The fragment of Anthemius on burning mirrors and the ‘Fragmentum mathe­
which seems, however, to have had no apparent influence on Witelo, if indeed he had seen it. maticum Bobiense',” Bibliotheca mathematica, 3. Folge, Vol. 7 (1906-07), pp. 225-33,
76 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 77

passage, Eutocius summarizes from Geminus’ De scientia mathematica, But if the cone is an obtuse-angled cone, as in the second figure [see Fig. 3.14],
Bk. VI, the contrasting constructions of conic sections by the Ancients with the obtuse angle posited as BAG and the right angle being AEZ, then
and by Apollonius.20 It is first noted that the Ancients defined a cone as angles BAG and AEZ [together] are greater than two right angles. Hence
DEZ does not meet side AG in the direction of Z and G but only in the direction
arising from the rotation of a right triangle. Such cones [obviously, right
of A an d £ , GA having been extended toward D. Therefore, the cutting plane
cones] were distinguished by their vertical angles as “right-angled,”
will produce in the surface of the cone a section which is called a hyperbola,
“ obtuse-angled,” and “ acute-angled.” Each of these right cones has its thus called because the angles AEZ and BAG [together] “ exceed” two right
characteristic sections: angles or because DEZ “ exceeds” (i.e. extends beyond) the apex of the cone
in the right-angled cone it is that which we now call the parabola, in the and meets with GA outside.
obtuse-angled cone [that which we now call] the hyperbola, and in the acute- But if the cone is an acute-angled cone, with the acute angle posited as BAG
angled cone [that which we now call] the ellipse. . . . And that which is [see Fig. 3.15], then angles BAG and AEZ [together] are less than two right
called the section of a right-angled cone they (the Ancients) sought only in angles. Hence EZ and AG extended will meet somewhere, for we can increase
a right-angled cone when it was cut by a plane perpendicular to a side [i.e. the size of the cone. Therefore, a section will be produced in the surface which
generator] of the cone, while they were accustomed to demonstrate that the is called an ellipse, thus called because the two angles “ fall short” of two right
section of an obtuse-angled cone arises [only] in an obtuse-angled cone and angles or because the ellipse “ falls short” of being a circle.
the section of an acute-angled cone [only] in an acute-angled cone, when in
each case the plane perpendicular to a side of the cone is drawn. This is In this passage Eutocius correctly describes the generation of the conic
what these ancient names of those lines indicate. sections according to the ancients, but at the same time he adds false
etymologies of their Apollonian names (for their correct explanation by
The passage continues by saying that afterwards Apollonius of Perga Pappus, see note 18 above). At any rate, Eutocius goes on to illustrate
more universally investigated, in any cone, right or scalene, all the sections the different technique of Apollonius, who supposes a cone that is either
that arise as we vary the position of the cutting plane with respect to the right or scalene and who abandons the condition of having the cutting
cone, and he was so admired for his demonstrations of conic theorems plane perpendicular to a side of the cone— but instead adopts the proce­
that “ they were wont to call him the Great Geometer.” Eutocius follows dure of varying the angle of the cutting plane:
the discussion from Geminus with a geometric illustration of the generation
of the conic sections, first according to the Ancients and then according For again, as in these same figures [Figs. 3.13-3.15], let the cutting plane
to Apollonius:21 be KEL, and let the common section of it and the base of the cone be KZL,
and let the common section of plane KEL and triangle ABG be EZ, which
Let axial triangle ABG be posited and from some point E let DE be drawn latter is called the diameter of the section. Now in all of the sections he
perpendicular to AB, and let a plane drawn through DE perpendicular to supposes KL to be perpendicular to BG, the base of the triangle. Then further,
AB cut the cone. And so angles AED and AEZ are right angles. Now if the if EZ is parallel to AG, the section KEL in the surface of the cone becomes
cone is right-angled and hence angle BAG is a right angle, as in the first a parabola; but if EZ meets side AG beyond the apex of the cone (as at D),
figure [see Fig. 3.13], then angles BAG and AEZ [together] equal two right section KEL becomes a hyperbola; while if EZ meets AG beneath [the apex
angles. Hence DEZ is parallel to AG. And in the surface of the cone a section of the cone], the section becomes an ellipse, which latter they call a “ shield.”
is formed which is called a parabola, thus called because DEZ, which is the Therefore, universally, the diameter of a parabola is parallel to one side of
common section of the cutting plane and the axial triangle, is parallel to side the triangle; the diameter of the hyperbola meets the side of the triangle in
AG of the triangle. the direction of the apex of the cone; and the diameter of the ellipse meets
with the side of the triangle in the direction of the base.
and particularly pp. 228, 230, and G. L. Huxley, Anthemius o f Tralles: A Study in Later Now if Witelo read this passage from Eutocius’ commentary, he must
Greek Geometry (Cambridge, Mass., 1959), pp. 39-43. I doubt Huxley’s conclusion that have immediately seen that the Pre- Apollonian names of the conic sections
Witelo’s Proposition IX.44 depended on Anthemius' fragment since it can be adequately
explained on the basis of Alhazen’s De speculis comburentibus and Witelo’s own proposi­
had to be transformed if they were to be used in his work. For he was
tion concerning reflection. For the question of the relationship between Witelo and William concerned here not only exclusively with a right cone (as were the An­
of Moerbeke, see the important article of A. Paravicini, “ Witelo et la science optique a la cour cients) but with one specific kind of such a cone, the right-angled right
pontificale de Viterbe (1277),” Melanges de I'Ecole Franqaise de Rome: moyen age, temps cone. Assuming such a cone for the generation of all three conic sections,
modernes, Vol. 87 (1975), pp. 425-53. He establishes with surety that the two scholars as he did, he could no longer use the Pre-Apollonian names that depended
were together at Lyon in 1274 and at Viterbo in 1277.
on variations of the vertical angle of the cone. And so he now adopted
20 Apollonii Pergaei quae Graece exstant cum commentariis antiquis, ed. of J. L. Heiberg,
Vol. 2 (Leipzig, 1893), pp. 168, line 12— 170, line 27. the Apollonian procedure of varying the angle of the cutting plane. Thus
21 Ibid., pp. 170, line 28— 176, line 12. instead of applying the adjectives “right-angled,” “ obtuse-angled” and
78 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 79

“ acute-angled” to the cones from which the sections were constructed cylinder) is parallel to the base, then the section is circular. Conversely,
by the Ancients he now applied them to the sections themselves. Hence if the section is a circle, then the cutting plane is parallel to the base.
by a “right-angled section” he understood the section formed by a plane Further, the parallel cutting plane forms a new [right] cone or [right]
cutting a side of the axial triangle at a right angle, by an “ obtuse-angled cylinder. The first part of the enunciation (with respect to a cone) is essen­
section” the section formed when that same plane cut the same side at an tially equivalent to Proposition 2 of the De duabus lineis, edited in Chapter
obtuse angle, and by an “ acute-angled section” the section formed when 2, and to Proposition 1.4 of the Conics of Apollonius. It may seem doubtful
the same plane cut the same side at an acute angle. So in essence Witelo to the reader who compares the two proofs that Witelo took his proposition
abandoned the procedure for generating conics followed by the Ancients from Apollonius, since he does not mention Apollonius, nor does the prop-
while at the same time transferring the distinguishing adjectives used by basis D\ secetque ipsam superficies plana equidistanter basi, et sit communis sectio super­
the Ancients from the cones from which they produced the sections to ficiei illius et superficiei conice pyramidis linea EFG. Dico quod linea EFG est periferia
the sections themselves, i.e. to their inclinations. circuli. Secet enim alia superficies plana pyramidem per verticem et per axem, qui est
Despite the difference of approach evident in Witelo’s and Eutocius’ AD. Communis itaque illius superficiei et pyramidis sectio est trigonum, quod sit ABC,
per 90am huius; secetque superficies EFG axem AD in puncto H et trigonum ABC secet
accounts, a difference dictated by Witelo’s assumption of a single right- superficiem EFG in linea EHF. Erit ergo linea EH equidistans linee BD, per 16am XI1.
angled right cone, I think it probable that Witelo had read Eutocius’ Est ergo, per 29am I' et per 4am VF, proportio linee BA ad EA sicut linee CA ad lineam
discussion, for like Eutocius he still focused (through his citations of AF. Ergo, per 7am huius, erit eversim proportio linee BA ad lineam BE sicut linee CA
1.29 of the Elements and his own Proposition I.14)22 on the sum of the ad lineam CF. Ergo, per 16am V‘, erit permutatim proportio linee BA ad lineam CA sicut
two interior angles that determine if and in what direction the diameter linee BE ad lineam CF. Sed linea BA est equalis ipsi CA, per 89am huius; ergo erit linea
BE equalis linee CF. Ducantur itaque linee DE et DF. Et quoniam, per 89am huius, anguli
of the section and the opposite side of the axial triangle meet. In the
quos continent linee longitudinis pyramidum cum semidyametris basium sunt equales,
Eutocian passage the angle of the cutting plane is always a right angle palam, per 4am l1, quia linea DE est equalis linee DF, et angulus EDB est equalis angulo
and the vertical angles of the cones are successively right, obtuse, and FDC. Quia vero angulus HDB equalis angulo HDC, quoniam ambo sunt recti, et angulus
acute, which produce successively an angular sum which is equal to two EDB equalis angulo FDC, remanet angulus EDH equalis angulo FDH, quoniam sunt residue
right angles, is greater than two right angles, and is less than two right partes rectorum super angulos equales. Palam ergo, per 4am I', quod (correxi ex quoniam)
linea EH est equalis linee HF, Similiterque ductis lineis HG et DG, et completa prout
angles, thus producing successively the conditions necessary for the parab­ in premissis figuratione, declarabitur quod (correxi ex quoniam) lin ea/7/ / est equalis linee
ola, the hyperbola, and the ellipse. In the Witelonian passage the vertical GH. Sunt enim trigona equiangula, ut patet intendenti Ergo, per 9am IIP, punctum H est
angle is always a right angle and the angles of the cutting plane are suc­ centrum circuli. Est ergo EFG linea circumferentia circuli, quod est propositum. Et si
cessively right, obtuse and acute; and thus the angular sums and resulting sectio EFG est circulus, palam quod (correxi ex quoniam) superficies plana secundum
conics are as before. Whatever we say about the particular source or illum circulum secans pyramidem est equidistans basi. Erit enim EAF pyramis cuius axis
AH et centrum basis H. Erit itaque linea longitudinis, que est EA, equalis linee FA, per
sources of this proposition of Witelo’s, it does reveal at least a rudimentary 89am huius. Sed et linea BA equalis est ipsi CA; remanet ergo linea BE equalis ipsi CF.
knowledge of the Apollonian construction of conic sections. Erit quoque linea ED equalis linee FD, per 4am Γ. Et quia trigona EHD et FHD sunt
Little attention need be given to Proposition 1.100 of Witelo’s Per­ equalia inter se latera habentia, ergo, per 8am F, angulus EHD est equalis angulo FHD.
spectiva,23 It asserts that if a plane which cuts a [right] cone (or a [right] Ergo, per diffinitionem linee super superficiem erecte, patet quod DH erecta est super
superficiem EFG. Sed eadem linea HD est erecta super basim pyramidis cuius dyameter
22 See the Campanus version of the Elementa (Basel, 1546), p. 25: “ 29. Si duabus lineis est BC. Ergo, per 14am XF, superficies EFG est equidistans basi date pyramidis, quod
aequidistantibus linea supervenerit, duo anguli coalterni aequales erunt, angulusque extrinse­ est propositum.
cus angulo intrinseco sibi opposito aequalis, itemque duo anguli intrinseci ex alterutra parte Quoniam simpliciter, secundum premissum in pyramidibus modum, in columpnis quoque
constituti duobus rectis angulis aequales.” For Proposition 1.14 of Witelo’s Perspectiva, rotundis potest demonstrari. Et propter equidistantiam linearum longitudinis columpne
see the ed. of Unguru, p. 222: “ [Propositio] 14. Si linea recta super duas rectas ceciderit facilitas accedit demonstrationi. Fuerit enim linee DF, DG, DE equales, ergo et linee HE,
feceritque angulos coalternos inequales, art duos intrinsecos minores duobus rectis, vel HF, HG\ eritque sectio EGF circulus, per 9am IIF. Et conversa simpliciter patet, per 14am
extrinsecum inequalem intrinseco, illas duas lineas ad minorum angulorum partem con­ XF, ut prius, et hoc proponebatur. Per hec itaque patet manifeste quod (corr. ex quoniam)
currere est necesse, ad aliam vero partem impossibile; et si linee concurrunt necesse est omnis plana superficies secans quamcunque pyramidem equidistanter sue basi novam consti­
dictos angulos aliquo propositorum modorum se habere.” See also ed. of Risner, p. 8. tuit pyramidem cuius in pyramide rotunda basis est circulus, et in laterata pyramide super­
23 Ed of Unguru, pp. 274-76 (with bracketed phrase from ed. of Risner, p. 38): “ [Proposi­ ficies similis basi illius secte pyramidis, ut patet per 99am huius. Semper tamen vertex
tio] 100. Omnis superficiei plane secantis pyramidem vel columpnam rotundam trans axem illius pyramidis abscise est idem cum vertice prioris, et axis abscise pars axis ipsius prioris
equidistanter basi et curve superficiei pyramidis, vel columpne, communis sectio est circulus; date, basis quoque equidistat basi. Similiter quoque sit in columpnis rotundis vel lateratis,
et si illa sectio est circulus, superficies secans est equidistans basi; ex quo patet quod superficies enim equidistanter basibus secans quamcunque columpnan novam efficit colump­
omnis plana superficies equidistanter basi secans pyramidem vel columpnam novam pyrami­ nam rotundam vel lateratam; ymmo duas, scilicet abscisam et ipsam residuam, quod non
dem constituit vel columpnam. [4. theor. I Conicorum Apollonii et 5. the. Cylindricorum accidit in pyramidibus. Patet ergo totum quod proponebatur.” Cf. pp. 117-19 of his ed.
Sereni.] Sit pyramis rotunda ABC [Fig. 3.16] cuius vertex A , dyameter basis BC, et centrum for Unguru’s translation.
80 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 81

osition have Apollonius’ generality (Witelo’s cone is simply a right cone), to the base; and let the common section of this plane surface and of the
and the proposition is a commonplace of Alhazen’s Perspectiva, which conic surface be line FGHK. 1 say that it is not possible that this section
was Witelo’s primary source.24 But still Witelo’s modified proposition is a circle.
was no doubt suggested to him by one of his sources; hence it is not im­ For let a circle be drawn around point E [as a center] in the conic surface,
possible that he had read Conics, 1.4, and modified it to fit his right cone. in accordance with the preceding [proposition]. And so this circle will be
As for the converse, it is true of a right cone but is not universally true parallel to the base, by the 100th [prop.] of this [book]; and let it (the circle)
be FGLM. And let the lines of length of the cone [i.e. the generators] AF,
of an oblique cone because the subcontrary sections of the latter are
AG, AL and AM be designated. And so all of these [lines] will be equal,
circles but are not parallel to the base (see the Conics, Proposition 1.5).25 by the 89th [prop.] of this [book]; therefore, because the surface [is] parallel
But since Witelo understood only the right cone in his proposition, he to the base of the cone, it cuts off a new cone, by the 100th [prop.] of this
had no reason to mention the subcontrary sections. [book]; and since section FGHK is not parallel to the base of the cone, it is
We may also treat Witelo’s Proposition 1.103 rapidly.26 It held that if plain that it is not equally distant from the apex of the cone, which is point
a plane cuts a [right] cone or a [right] cylinder [wholly] and the plane is A. And so let point H be more remote from apex A and fall in line AL
not parallel to the base, the section cannot be circular. Its proof follows: extended, and let point K be nearer to apex A and fall in line AM. Conse­
quently, line AH will be greater than line AL, and line AK is less than line
Let there be a cone whose apex is A [see Fig. 3.17], whose diameter is AM. And let lines HE, KE, FE, GE, EL and EM be drawn. And since angle
BC, whose center of the base is D and whose axis is AD; and let a plane ALE is acute, by the 89th [prop.] of this [book], anglt HLE will be obtuse,
surface cut the cone across axis AD in point E, [the plane being] not parallel by 1.13 [of Euclid’s Elements]. Therefore, by 1.19 [of the Elements], side
HE of triangle HEL is greater than side EL. But side EL is equal to side
24 In note 40 of Chap. 1 above, I have listed the various propositions in Alhazen’s Per­ EF, by the definition of a circle. Now line EF comes from a point of the
spectiva that appear to reflect Proposition 1.4 of the Conics as well as other of Apollonius’ axis to a point of the section, since it is the common section of the circle
initial propositions. and of the oblique surface intersecting the cone. And so unequal lines are
23 Ed. of Heiberg, Gr 1, pp. 16-20. drawn from this point E to the periphery of the section. Hence that section
z(i Ed. of Unguru, p. 277 (with bracketed phrase from ed. of Risner, p. 39): “ [Propositio] is not a circle, by the definition of a circle.
103. Omnis superficiei secantis pyramidem vel columpnam rotundam trans axem, non We shall, therefore, call that section “ conic” in cones and “ cylindrical”
equidistanter basibus, et superficiei curve communem sectionem circulum esse est impos­ in cylinders. But that section in cones has earlier (in the 98th [prop.] of this
sibile. [5 theo. I Conicorum Apollonii item 9 theor. Cylindricorum Sereni.] Sit pyramis [book]) been called an acute-angled section or ellipse. And since such a
cuius vertex A [Fig. 3.17], dyameter basis BC, et centrum basis D, et axis AD; secetque
section is an oblong figure, it is plain that it has many diameters, all unequal,
ipsam superficies plana trans axem AD in puncto E, non equidistanter basi, et sit communis
sectio huius superficiei plane et superficiei conice linea FGHK. Dico quod hec sectio non
which pass through the same point of the axis of the intersected body and
est possibile ut sit circulus. Esto enim ut circa punctum E, in pyramidis conica superficie, which bisect this section. The greatest of these [diameters] is that which
ducatur circulus, per premissam. Hic itaque equidistabit basi, per 100am huius, sitque transits the length of the section and the least is that which transits the
FGLM. Et signentur linee longitudinis pyramidis AF, AG, AL, AM. Hee itaque omnes width and is erected orthogonally to the greatest diameter. And so what has
erunt equales, per 89am huius, ideo que superficies equidistans basi pyramidis novam been proposed is evident.
pyramidem abscindit, per 100amhuius, et quoniam sectio FGHK non equidistat basi pyrami­
dis, patet quod non equaliter distat a vertice pyramidis, qui est punctus A . Sit itaque punctus This is to be compared with Proposition 1.9 of the Conics: “ If a cone
H remotior a vertice A et cadat in linea AL producta, et punctus K sit propinquior vertici is cut by a plane meeting both sides of the axial triangle, [a plane that is]
A et cadat in linea AM. Erit itaque linea AH maior quam linea AL, et linea A K minor neither parallel to the base nor situated subcontrariwise, then the section
est quam linea AM; et continuentur linee HE, KE, FE, GE, et linee EL, EM. Et quoniam will not be a circle.’’27 While this Apollonian proposition may well have
angulus ALE est acutus, per 89am huius, erit angulus HLE obtusus, per 13am F. Ergo, per
19am I‘, latus HE trigoni HEL est maius latere EL. Sed latus EL est equale lateri EF, per
been the source of Witelo’s enunciation, it is clear that, if it was, Witelo
circuli diffinitionem. Linea vero EF venit a puncto axis ad punctum sectionis, quia est altered it to fit his exclusive concern with a right cone by omitting the
communis sectio circuli et superficiei oblique pyramidem secantis. Inequales itaque linee unnecessary restriction that the cutting plane be not situated subcontrari­
ab hoc puncto E producuntur ad periferiam sectionis. Non est ergo sectio illa circulus, per wise. His concern with a right cone and his consequent omission of any
circuli diffinitionem. Dicemus ergo illam sectionem in pyramidibus pyramidalem et in reference to subcontrary sections made it necessary for him to present
columpnis columpnarem. Est tamen illa sectio in pyramidibus in 98a huius prius dicta
a proof completely different from Apollonius’, which latter proof was
sectio oxigonia, vel ellipsis (Risner; elipsis Ung.). Et quoniam talis sectio est figure oblonge,
patet quod ipsa habet dyametros plurimas omnes inequales, et per idem punctum axis based on showing that if, contrary to the apodosis, the section were as­
secti corporis transeuntes, ipsam quoque sectionem per equalia dividentes, quarum maxima sumed to be circular, then it must be a subcontrary section and thereby
est que transit longitudinem sectionis, minima vero est que pertransit latitudinem et est contradict one of the conditions of the protasis. Incidentally, the fact that
super maximam dyametrum orthogonaliter erecta. Patet itaque propositum.’’ Cf. p. 120
for Unguru's translation. 27 Ed. of Heiberg, Gr 1, p. 32; Taliaferro, op. cit. in Chap. 1, note 19, p. 613.
82 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 83

Witelo focused on a right cone, thereby making any reference to sub­ between one of the joined [lines] and the given point is equal to the sum of
contrary sections superfluous, makes Risner’s citation of Proposition 1.5 any given line and the remaining part of the line [to be drawn], which [latter
of the Conics as a source for Witelo’s proposition inappropriate (see the part] lies between the given point and the other of the joined lines.
bracketed phrase in the enunciation in note 26). Risner also cited Serenus’ For example, let there be two straight lines that meet and are joined angu­
On the Section of a Cylinder, Prop. 9, as a possible authority for Witelo’s larly in point K, which lines [we] let be FK and TK [see Fig. 3.18]. Between
proposition {ibid.), but I doubt that Serenus’ work influenced Witelo at these [lines] let point M be given, and let the given line be MC. It is proposed
this point.28 If it had, I would expect that Witelo would have noted that to us that a straight line be drawn through point M between lines FK and
TK such that it cuts those lines in points O and L and its part LM = given
the cylindrical section was also an ellipse, which he failed to do.
line MC + remaining part MO [i.e., LM = CO]. But [seeking] to demonstrate
Turning to the second part of the proof where Witelo identifies the this by straight lines or circular lines, we spent long and greatly diverse labor
section in cones as an ellipse, we should note here the lack of precision. without being able to achieve it by means of such lines except by motion
All that Witelo does is to note the plurality of unequal diameters and and mechanical imagination. So, since lines FK and TK are given as indefinite
to identify the major and minor axes of the ellipse. I think it very doubtful in length, let us imagine line LO (fixed [only] in point M) to be moved until
that he dredged this modest information out of the rather complicated we have the desired [relationship of its parts].
proposition of 1.15 of the Conics, which was concerned with conjugate However, Apollonius of Perga has demonstrated this in his book On Conic
diameters.29 Much more likely is it that he was briefly expressing what Elements, Bk. II, Prop. 4, by drawing an obtuse-angled section through the
he had learned of conjugate diameters and axes of an ellipse from the given point and between the two asymptotic lines so that it (the hyperbola)
fragment translated by Gerard of Cremona from the Arabic (see above, cuts neither of these lines. [Here] we [simply] supposit the demonstration
Chap. 1, note 5, passage [6], and note 31). of it, inasmuch as it depends on the many prefatory principles of his book,
and we use it as [if it were] demonstrated.
We must now skip such intervening propositions which have only periph­
eral interest to our investigation and concentrate on Propositions 1.129 and This proposition, which is crucial for the demonstration of the next
1.131 of Witelo’s Perspectiva, the first propositions in which Witelo ac­ proposition, 1.130,31 which latter he borrowed in toto from Alhazen (see
knowledges his dependence on Apollonius. First let us examine 1.129
in toto:30 31 Ed. of Unguru, pp. 300-03: “ [Propositio] 130. Sumpta circuli dyametro et sumpto in
circumferentia puncto inequaliter distante a terminis dyametri, possible est a sumpto puncto
129. From a point given between two straight [lines] that are joined angu­ ad eductam dyametrum lineam ducere, que vel cuius pars interiacens periferia et dyame-
larly, to draw a straight [line] such that the one part of the line that lies trum sit date linee equalis. Disponantur omnia ut in 128a huius [=Alhazen, Perspectiva,
V.32], nisi quod punctus datus in circumferentia circuli, qui sit A [Fig. 3.19a], inequaliter
28 Unguru {ed. cit., p. 194) says that Risner’s reference to Serenus is more “ defensible,” distat a terminis dyametri qui sint G et B; eruntque linee AB et AG inequales, ideo quod
and so it is, but I am still inclined to think that, even if it can be argued that Witelo used punctum A inequaliter est distans a punctis G et B. Protrahatur ergo a puncto G linea
Serenus’ work elsewhere, Witelo here in Proposition 1.103, restricting his consideration equidistans linee AB, ex 31a P, que sit GN, et sumatur linea quecunque, ut pote ZT [Fig.
to a right cone and right cylinder, did not use Serenus’ more general proposition. 3.19b], et fiat supra punctum eius Z angulus equalis angulo AGD, per 23am P, qui sit angulus
29 Ed. of Heiberg, Gr 1, pp. 58-64; Taliaferro, op. cit. pp. 621-24. TZF, ducta linea ZF\ et ducatur a puncto T linea equidistans linee ZF, ut prius, que sit
30 Ed. of Unguru, p. 300 (with bracketed phrase from the ed. of Risner, p. 52): “ [Proposi­ TM, et ex angulo TZF secetur angulus equalis angulo NGD, per 27am huius, qui sit angulus
tio] 129. Inter duas rectas angulariter coniunctas, a dato puncto rectam ducere cuius una TZM, ducta linea ZM que, per 2am huius, necessario concurret cum linea TM, cum sit
partium interiacens unam coniunctarum et datum punctum sit cuicunque date linee et insuper ducta inter equidistantes; sit ergo punctus concursus M. Restat ergo ut angulus MZF sit
relique sue parti datum punctum et alteram coniunctarum interiacenti equalis. [4 theor. equalis angulo AGN. A puncto itaque T ducatur linea equidistans linee ZM, que sit TO.
2 conicorum Apollonii.] Exempli causa sit ut due linee recte in puncto uno angulariter Hec quoque necessario concurrat cum linea FZ, per 2am huius; sit ergo earum concursus in
coniungantur, que sint FK et TK [Fig. 3.18] concurrentes in puncto K, inter quas sit datus puncto K. Sumatur quoque, per 3am huius, linea cuius proportio ad lineam ZT sit sicut
punctus M, et sit data linea MC. Proponitur nobis ut a puncto M ducatur linea recta inter dyameter GB ad lineam QE, lineam datam, et hec sit linea I. Deinde a puncto M dato
lineas TK et FK, secans illas in punctis O et L ita ut eius pars que est LM sit equalis date inter duas lineas FK et KO ducatur, per premissam, linea que sit LCMO, secans lineam
linee MC et insuper relique sue parti que est linea MO. Ad hoc autem per lineas rectas LK in puncto L et lineam KO in puncto O ita ut eius pars CM sit equalis date linee 1
vel circulares demonstrandum, longus labor et multe diversitatis nobis incidit, et non fuit et eius pars LC sit equalis linee MO\ et a puncto T ducatur linea TF equidistans linee
nobis possibile complere per huiusmodi lineas absque motu et ymaginatione mechanica, LO, per 31am P; hec quoque, per 29am P huius, secabitur a linea ZM. Sit ergo punctus
ita ut cum linee FK et TK date sint nobis indefinite, linea LO fixa in puncto M, ymaginetur sectionis Y. Fiat ergo supra punctum A, terminum linee GA (punctum scilicet quod est in
moveri quousque nobis accidat res quesita. Hoc tamen Apollonius Pergeus in libro suo circumferentia circuli), angulus DAG equalis angulo ZFT, per lineam AND. Palam autem
de conicis elementis, libro secundo, propositione quarta, per ductionem sectionis amply- quod hec linea concurret cum producta dyametro GD. Cum enim angulus DAG sit equalis
gonie a dato puncto inter duas lineas asymptotas (Risner; assymptotas. Ung.), nullam angulo ZFT, et angulus AGN equalis angulo FZM, et angulus DGN equalis angulo TZM,
illarum linearum secante demonstravit; cuius nos demonstrationem ut a multis sui libri totusque angulus AGD equalis toti angulo FZT, et cum linee FT et ZT concurrant, ergo et
principiis preambulis dependentem hic supponimus, et ipsa utimur sicut demonstrata.” linee AD et GD concurrent; aut ergo linea AD continget circulum aut secabit ipsum. Sit
Cf. p. 147 for Unguru’s translation. ergo linea AD primo contingens circulum in puncto A . Cum ergo angulus GAN sit equalis
84 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 85

angulo ZFT et angulus AGN sit equalis angulo FZM, palam, per 32am I1, quia angulus above, Chap. 1, for my discussion of Proposition V.33 of Alhazen’s Per­
ANG erit equalis angulo ZYF eritque triangulus AGN equiangulus triangulo ZFY. Est ergo, spectiva), can be summarized as follows. Given two indefinitely long
per 4am VT, proportio linee AN ad lineam AG sicut linee FY ad lineam FZ. Similiter, cum straight lines F K and TK , which are joined in an angle at K , and given
angulus AGD sit equalis angulo FZT et angulus GAD equalis angulo ZFT, erit, per eandem, any point that lies between these lines, to draw a straight line OMCL
triangulus AGD similis triangulo FZT. Ergo, ut prius, que est proportio linee AG ad lineam
through M such that ML = MC + MO. First, Witelo confesses that he
GD eadem est linee FZ ad lineam ZT. Si ergo que est proportio linee AN ad lineam AG
eadem est linee FY ad lineam FZ, et que est proportio linee AG ad lineam GD eadem est tried to solve this neusis by circles and straight lines but without success.
linee FZ ad lineam ZT, erit ergo, per equam proportionalitatem, per 22am V', ut que est He then says it might be solved by “ motion and mechanical imagination.”
proportio linee AN ad lineam GD eadem sit linee FY ad lineam ZT. Quia vero linea TM That is to say, one could try to adjust and move the line, on which CM
est equidistans linee FL et linea FT equidistans linee LM, erit superficies LFTM equi- had been marked off, until the parts beyond it in both directions and ex­
distantibus contenta lateribus. Palam ergo, per 34am 1‘, quod (correxi ex quoniam) linea
tending to the joined lines were equal, making sure that the rule or line
FT est equalis linee L M . Quare erit linea FT equalis linee CO, quoniam linea MO est equalis
ipsi LC, per premissam; linea ergo CM addita utrique adhuc sunt equales; eritque LM moved always was passing through point M. In short, the author’s first
equalis linee CO. Sed linea MO est equalis linee YT, per eandem 34am l·, et linea YM est solution is a mechanical one. But he then goes on to tell us that Apollonius
equalis linee TO. Restat ergo ut linea ET sit equalis linee CM. Sed linea CM, ex premissis, has given the solution in his On Conic Elements, Bk. II, Prop. 4, by
est equalis linee 1. Est autem, ex premissis, et proportio linee /, per 5am huius, ad lineam drawing a hyperbola through M within asymptotes F K and TK (see above,
ZT sicut dyametri BG ad lineam EQ. Erit ergo, per 7am V1, proportio linee FY ad lineam
Chap. 1, note 50). Once this is done, then a line equal to given line MC
TZ sicut dyametri BG ad lineam EQ. Quia vero est proportio linee AN ad lineam GD sicut
linee FY ad lineam ZT, ergo, per equam proportionalitatem, per l l am V', erit proportio can be inserted as a chord within the hyperbola by using MC as a radius
linee A N ad lineam GD sicut linee GB ad lineam EQ. Verum angulus GAN est equalis and M as the seat of the fixed foot of the compass. Then line MC is
angulo GBA, ex 32a IIP. Sed angulus NGD est equalis angulo GBA, per 29am P, quia extended in both directions until it meets the joined lines. So it is evident
linea NG equidistat linee BA. Igitur angulus NGD est equalis angulo AAG; et angulus that the solution of the problem is accomplished by the intersection of
NDG est communis ambobus trigonis NDG et ADG; ergo, per 32am P, erit angulus DNG
a circle and an hyperbola. Now to prove that this construction gives the
equalis angulo DGA. Sunt ergo dicti trianguli equianguli; erit ergo, per 4am VP, proportio
linee AD ad GD sicut linee GD ad ND. Ergo, per 17am VP, erit id quod fit ex ductu linee desired relationship of the parts of the line another proposition of Apol­
AD in DN equale quadrato GD. Sed id quod fit ex ductu linee BD in GD, per 36am IIP, est lonius is required, and indeed it was so mentioned by Alhazen without
equale quadrato D A ; quadratum vero linee DA est equale ei quod fit ex ductu linee AD specification of the proposition number, namely, Proposition II.8 of the
in DN et AD in N A , per 2am IP; et id quod fit ex ductu linee BD in DG est equale quadrato Conics (see above, Chap. 1, note 51). This proposition guarantees that
linee DG et ei quod fit ex ductu BG in DG, per 3am IP. Ablatis ergo equalibus hinc inde,
line CL = line MO. So, if line MC is added to both magnitudes, then
que sunt quadratum GD et rectangulum A D N , restat ut illud quod fit ex ductu linee AD in
A N sit equale ei quod fit ex ductu linee BG in DG; eritque, per 16am VP, proportio linee obviously ML = MC + MO, as Witelo (and Alhazen before him) wished.
AN prime ad lineam GD secundam sicut linee BG tertie ad lineam AD quartam. Ostensum
est autem supra quod est proportio AN ad lineam GD sicut linee BG ad lineam EQ. Erit
quod proportio linee AN ad lineam DG est sicut dyametri BG ad lineam EQ. Igitur, per
ergo, per 9am V', linea EQ equalis linee AD, quod est propositum: quoniam ipsa linea
9am V1, linea DH est equalis linee EQ, quod est propositum. Si vero linea AG sit minor
AD est date linee equalis, interiacet autem periferiam circuli et eductam dyametrum, eo
quam linea BA, secabit linea DA circulum in arcu AB [Fig. 3.19d]. Sit ergo ut secet ipsum
quod est contingens circulum. Quod si linea AD non sit contingens sed secans circulum,
in puncto H, et ducatur linea H G , et linea GN equidistans linee BA. Palam ergo, per
aut igitur linea AG est maior quam linea AB aut econtrario. Sit autem nunc linea AG maior
29am I', quod (correxi ex quoniam) angulus NGD est equalis angulo ABG. Sed angulus
quam linea BA [Fig. 3 .19c]. Palam quia linea a puncto A ad dyametrum BG extra circulum
ABG est equalis angulo AHG, per 27am IIP, quoniam ambo cadunt in arcum GA et sunt
ducta secabit circulum in arcu AG. Sit ergo ut secet ipsum in puncto H, et ducatur linea
super circumferentiam circuli. Ergo angulus NGD est equalis angulo AHG et angulus NDG
HG. Palam itaque, cum quadrangulum ABGH sit inscriptum circulo, quia duo anguli AHG
communis est ambobus trigonis, scilicet NDG et DHG. Est ergo tertius DNG equalis tertio
et ABG, per 22am IIP, sunt equales duobus rectis. Ducatur quoque linea GN equidistans
scilicet D GH, per 32am P. Ergo, per 4am VP, erit proportio linee HD ad lineam DG sicut
linee BA. Erit ergo, per 29am P, angulus NGD equalis angulo GBA. Igitur angulus NGD
linee DG ad lineam DA; ergo, per 17am VP, illud quod fit ex ductu HD in DN est equale
et angulus AHG sunt equales duobus rectis. Sed, per 13am P, angulus NHG cum angulo
quadrato linee GD. Sed illud quod fit ex ductu BD in DG, per 36am IIP, est equale ei
AHG valet duos rectos. Angulus ergo NGD est equalis angulo NHG. Angulus vero NDG est
quod fit ex ductu HD in DA; illud autem quod fit ex ductu HD in DA est, per lam IP, equale
communis ambobus trigonis GDN et HDG. Erit ergo tertius angulus, qui est DNG, equalis
ei quod fit ex ductu linee HD in DN et linee HD in AA; illud vero quod fit ex ductu linee
tertio, qui est D GH , per 32am P. Ergo, per 4am VP, latera equos angulos respicientia sunt
BD in DG, per 3am IP, valet idem quod fit ex ductu linee BG in GD et quadratum GD.
proportionalia. Est igitur proportio linee HD ad lineam DG sicut linee DG ad lineam DN.
Ablatis ergo equalibus hinc inde, erit illud quod fit ex ductu HD in NA equale ei quod fit
Ergo, per 17am VP, illud quod fit ex ductu HD in DN est equale quadrato DG. Sed id quod
ex ductu BG in GD. Erit ergo, ut prius, proportio linee AN ad lineam DG sicut linee BG
fit ex ductu AD in DH est equale ei quod fit ex ductu BD in DG, per 36am IIP. Item id
ad lineam HD. Sed iam ostensum est supra quod est proportio linee AN ad lineam DG
quod fit ex ductu AD in DH est equale ei quod fit ex ductu DH in DN et DH in A N, per
sicut linee BG ad lineam EQ. Igitur linea EQ est equalis linee HD, per 9am V1, quod est
p,m IP. Illud vero quod fit ex ductu BD in DG est equale ei quod fit ex ductu BG in GD
propositum: quoniam a puncto A dato ducta (om. Unguru hic sed habet in sua dissertatione)
et quadrato GD, per 3am IP. Ablatis igitur equalibus ab utrisque, scilicet quadrato GD ex
est linea secans circulum cuius pars a puncto sectionis usque ad concursum cum dyametro
una parte et illo quod fit ex ductu DH in DN ex altera, restat ut illud quod fit ex ductu
producta equalis est date linee. Patet ergo quod proponebatur.” Cf. pp. 147-52 for Unguru's
DH in AN sit equale ei quod fit ex ductu BG in DG. Erit ergo, per 16am VI1, proportio
translation. See also the ed. of Risner, pp. 52-54.
AN primi ad GD secundum sicut EG tertii ad DH quartum. Sed probatum est in precedentibus
86 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 87

There are several points of interest to our study of conic sections in I have given in note 31, but will not analyse because of my earlier treat­
Witelo’s comments to 1.129. (1) Witelo was quite obviously directed to ment of its model, Alhazen’s V.33. Turning now to Witelo’s Proposition
Proposition II.4 of the Conics by the specific declaration of Alhazen that 1.131, I first present it in full:33
we can construct the pyramidal section (i.e., the hyperbola) by that prop­ 131. To draw an hyperbola through a given point, on one side, between
osition of Apollonius (see above, Chap. 1, note 43), as he (Witelo) set two intersecting straight [lines, such that the hyperbola] does not touch these
about to reproduce Proposition V.33 of Alhazen’s Perspectiva. (2) But lines, [and] to describe on the other side of the common point of these lines
in following Alhazen, Witelo diverges from the Latin text of the former’s a hyperbola opposite to the prior one. From this it is plain that when two
work by using Greek terminology, i.e., by naming the conic section in­ opposite [hyperbolic] sections are [given] between two lines, and the least
volved as an “ obtuse-angled section” and the lines that it does not meet line is drawn from one section to the other, the part of that [least] line
as “ asymptotic lines.” I think it probable that he found the relevant which lies between one section and one [asymptotic] line will be equal to
proposition in the text of the Conics (most likely in a translation of it by that part of it [i.e., of the least line] which lies between the other section
Moerbeke), where the terms “ asymptotes” and “ hyperbola” are found. and the other [asymptotic] line.
If so, his only change was to substitute for “ hyperbola” his equivalent What is here proposed has been demonstrated by Apollonius in his book
On Conic Elements. Indeed, obtuse-angled sections or hyperbolas are called
name “ obtuse-angled section” (which, I have already suggested, he prob­
’’opposite” when the gibbosity of the one of them follows the gibbosity of
ably devised for 1.98 after reading Eutocius’ Commentary on the Conics
the other, so that these gibbosities reflect one another [i.e. are oppositely
of Apollonius). But one cautionary remark must be made. It is not impos­ symmetrical] and the diameters of both of them are in one straight line.
sible that he learned the Greek terminology and indeed the basic construc­ For example, let it be that the two lines HL and NZ intersect one another
tion of the hyperbola from Moerbeke’s translation of Eutocius’ Com­ in point X [see Fig. 3.20], and let an obtuse-angled section TP be drawn
mentary on the Sphere and the Cylinder o f Archimedes, which he almost through point T on one side of these [lines], namely, within the angle HXZ,
certainly read (see note 6 above, passage 42vH-K). (3) In his comments or within the angle H X N , and let CU, a section opposite to that, be drawn
to 1.129, Witelo failed to mention the need to cite II.8 of the Conics to on the other side, either within the angle N XL, or within the angle ZXL,
prove that the hyperbola constructed through M and within asymptotes so that the diameters of both sets of these opposite sections are in one line,
FG and TG provided the solution to the neusis presented in 1.129. Perhaps which is TC, drawn from the vertex of one [branch] to the vertex of the
the reason for this failure was that Alhazen, while noting Apollonius’ other, which [line TC] is necessarily the least of all lines drawn between
those two sections. And from these [constructions] Apollonius has shown
proof, did not himself specify the proposition number and consequently
that which is proposed as a corollary, namely, that if line TC cuts line HL
Witelo was unable to find the relevant proposition. Or it could be that he
in point F and line ZN in point Q, then line TQ will be equal to line CF;
found Proposition II.8 but for some reason did not wish to cite it. It
would certainly not have been difficult for him to have found it if we
33 Ed. of Unguru, pp. 303-04: ’’[Propositio] 131. Inter duas rectas se secantes ex una
assume that he had found Proposition II.4 of the Conics, for the copy parte a puncto dato yperbolem illas lineas non contingentem ducere, ex altera parte com­
that he or Moerbeke would probably have seen was Vat. gr. 203, which, munis punctis illarum linearum yperbolem priori oppositam designare; ex quo patet quod,
unlike some other copies, did have proposition numbers.32 Now if he had cum fuerint due sectiones opposite inter duas lineas, et producatur linea minima ab una
found Proposition II.4, it would have required his reading of only the sectione ad aliam, erit pars illius linee interiacens unam sectionem et reliquam lineam equalis
sue parti aliam sectionem et reliquam lineam interiacenti. Quod hic proponitur demon­
next four propositions for him to have found Proposition II.8. Be that
stratum est ab Apollonio in libro suo de conicis elementis. Dicuntur (Risner\ Ducunter
as it may, he not only did not cite Proposition II.8, he omitted Alhazen’s Ung.) autem sectionis ampligonie sive yperbole opposite quando gibbositas unius ipsarum
reference to the Conics for this crucial conclusion. Still I remain persuaded sequitur gibbositatem alterius ita ut ille gibbositates se respiciant, et ambarum dyametri
that Witelo had seen and read the Conics in at least a cursory fashion sint in una linea recta. Verbi gratia, sit ut due linee HL et ZN secent se in puncto X
because of his statement that he would not present the demonstration [Fig. 3.20], et ex una parte illarum, scilicet sub angulo HXZ, vel sub angulo HXN, a dato
puncto, qui sit T, ducatur sectio ampligonia, que sit TP, et altera parte sub angulo NXL,
of Proposition II.4, “ inasmuch as it depends on the many prefatory prin­
vel sub angulo ZXL, ducatur sectio illi opposita, que sit CU, ita quod dyametri quarumlibet
ciples of his book.” These are, I believe, the words of some one who oppositarum ambarum illarum sectionum sint in una linea, que est TC, a vertice unius
has looked at the text and decided that it was too complex to epitomize ad verticem alterius producta, que necessario est minima omnium linearum inter illas duas
it for the mere purpose of utilizing one or two propositions. Now, as I sectiones productarum. Et ex hiis declaravit Apollonius illud quod correlarie proponitur,
have already stated, the purpose of singling out this Apollonian material scilicet quod si linea TC secet lineam HL in puncto F et lineam ZN in puncto Q, quod
linea TQ erit equalis linee CF, et si linea TC pertranseat punctum X, erit linea TX equalis
was to present it as a theorem ancillary to his next theorem, 1.130, which
linee X C . Et nos utimur hic illo ut per Apollonium demonstrato, et propter conformitatem
positionis sectionum respectu linearum se intersecantium. Patet ergo propositum.” Cf. pp.
32 See the apparatus in Heiberg’s edition, where Vat. gr. 203 is abbreviated as v. 152-53. See also the ed. of Risner, p. 54.
88 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 89

and, if line TC passes through point X , line TX will be equal to line XC. osition V.34 of Alhazen’s Perspectiva, which I have presented and ana­
And we use that [proposition] here, inasmuch as it has been demonstrated lysed in Chapter 1. Witelo’s comment to Proposition 1.131 is not a proof.
by Apollonius and because of the conformity of the position of the sections It consists only of a definition of the opposite branches of a hyperbola
with respect to the intersecting lines. Therefore that which has been proposed and a specification of the proposition (and corollary) in terms of particu­
is evident. lar magnitudes. Witelo first notes that this proposition has been demon­
strated by Apollonius in his On Conic Elements. Then follows the definition
This proposition has been singled out by Witelo because it is crucial of opposite branches, a definition taken almost verbatim from Gerard of
for his Proposition 1.133.34 The latter proposition he copied from Prop- Cremona’s translation of that part of the proem of the Conics prepared
by the Arabic editor (see above, Chap. 1, note 23). I have already men­
34 Ed. of Unguru, pp. 305-06: "[Propositio] 133. A puncto dato in circuli circumferentia tioned that this statement concerning opposite branches was but a pale
extra dyametrum possibile est ducere lineam per dyametrum ad circumferentiam ita ut pars reflection of Proposition 1.14 of Apollonius’ Conics because it simply
eius interiacens dyametrum et reliquam partem circumferende sit equalis linee date eidem
circulo inscriptibili premisso modo; sed harum linearum equalium ab eodem puncto dato
declared that the gibbosities (i.e., the convexities or curvatures) of the
in eodem circulo producibiles sunt tantum due. Esto ABG circulus [Fig. 3.21], cuius dya- opposite branches were the same or symmetrical, while Apollonius’ prop­
meter sit BG, et punctus datus in sui circumferentia sit A , et sit HZ linea data minor dyametro osition proved that their parameters (already properly defined in Proposi­
BG, premisso modo possibilis inscribi circulo. Dico quod a puncto A possibile est ducere tion 1.12 of the Conics) were the same. Witelo’s definition of opposite
lineam transeuntem per dyametrum BG, cuius pars interiacens dyametrum BG et circum­
branches, then, cannot be considered as having been taken directly from
ferentiam sit equalis linee date, que HZ. Ducantur enim in circulo linee BA et AG, et super
punctum H linee date HZ fiat angulus equalis angulo AGB, qui sit MHZ, ducta linea M H ,
the Conics and certainly cannot be used as evidence that he had read or
et super idem punctum H fiat angulus equalis angulo ABG, qui sit LHZ, ducta linea HL\ mastered Apollonius’ Proposition 1.14.35 In the specification that follows
et a puncto Z ducatur linea equidistans linee HM, que sit Z N , que quidem secabit lineam the definition of opposite branches, Witelo constructs one branch through
HL. Sit ut secet ipsam in puncto X. Et a puncto Z iterum ducatur alia linea equidistans point T, presumably on the authority of Proposition 1.4 of the Conics,
linee HL, que sit ZT, secans lineam HM in puncto 7. Secabit autem per 2am huius, et a and then he simply lets the other branch be constructed (no point being
puncto T ducatur sectio conica, que sit TP, sicut premissum est in 13l a huius. Hec itaque
sectio non contingit aliquam linearum ZN et HL, inter quas ipsa iacet. Similiter fiat sectio
specified) so that the branches have the same asymptotes and a common
alia conica isti opposita, inter easdem lineas, ex parte alia que sit CU. Et inter illas sec­ diameter without reference to Proposition 1.14 or to any other proposition.
tiones omnium linearum ductarum minima ducta a puncto T ad sectionem CU sit linea He is, I believe, simply implying that the construction is of the type defined
TC. Hec ergo linea TC si fuerit equalis dyametro circuli BG, circulus factus secundum
semidyametrum TC, posito pede circini in puncto T, palam quia sectionem CU continget.
Si vero linea TC fuerit minor dyametro BG, circulus factus modo predicto secundum quan­ HZ, sed sicut per 131am huius, declaratum est, patet quod linea QT est equalis linee FC;
titatem linee BG secabit sectionem CU in duobus punctis, ut patet per premissam. Sit ergo sed linea TQ est equalis linee MZ, per 34amE, cum parallelogramum MTQZ sit equidistantium
nunc primum linea CT equalis dyametro BG. Cum ergo linea TC ducatur ad sectionem laterum, ut patet ex premissis. Est igitur linea MZ equalis linee FC. Sed, per eandem 34am
conicam, que interiacet lineas HL et ZN, necessario secabit linea TC illas ambas lineas; E, linea ZL est equalis linee TF. Est igitur totalis linea ML equalis totali linee TC. Ergo,
quas si in puncto A', qui est punctus communis sectionis illarum linearum, secuerit, erit linea per 7am V', est proportio linee TC ad HZ sicut linee LM ad HZ. Est ergo proportio linee
TX equalis linee XC\ quod si ipsas in aliis punctis secuerit, secet ergo lineam ZN in puncto GB ad lineam DE sicut linee TC ad HZ, et permutatim. Cum ergo linea TC sit equalis
Q et lineam HL in puncto F. Et ducatur a puncto Z, per 3 lam E, linea equidistans linee linee BG, erit linea ED equalis ipsi HZ date linee, quod est propositum. Si autem linea
TC que, per 2am huius, secabit lineas HM et HL, sicut etiam sua equidistans TC. Secet TC sit minor dyametro BG, producatur ultra sectionem donec ipsa sit equalis dyametro BG,
ergo eas in punctis M et L, et sit ipsa linea MZL. Super dyametri ergo GB terminum et secundum quantitatem eius fiat circulus. Palam, per premissam, quod ille secabit sectionem
G, per 23am E, fiat angulus equalis angulo HLM, qui sit angulus BGD, et ducantur due in punctis duobus, que sint C et U, a quibus linee ducte ad punctum T erunt equales
linee AD, DB. Palam ergo, cum angulus GAB sit rectus, per 31am IIE, quod alii duo anguli linee BG, per diffinitionem circuli; et tunc a puncto Z ducatur linea equidistans alteri
trianguli GAB, scilicet AGB et ABG, valent rectum, per 32am E. Angulus ergo LH M , qui illarum, et item alia equidistans alteri; et tunc erit ducere a puncto A , per modum predictum,
equalis est illis duobus angulis, est rectus; est ergo equalis angulo GDB. Angulus vero HLM duas lineas ED equales linee date; et erit idem penitus probandi modus, qui supra. Patet
est equalis angulo DGB\ igitur, per 32am E, angulus tertius unius trigonorum GBD et HLM ergo propositum.” Cf. pp. 153-55. See also ed. of Risner, pp. 55-56.
erit equalis angulo tertio alterius, scilicet angulus HML angulo GBD. Erit ergo, per 4am 35 Unguru (ed. cit., p. 204) suggests, without proof, that Witelo might have been in­
VE, proportio linee GB ad BD sicut linee LM ad M H . Sit autem punctus in quo linea AD spired by Proposition 1.14, or perhaps 1.60, of the Conics but he did not realize that Witelo
secat dyametrum BG punctus E. Quia ergo, per 27am IIE, angulus ADB est equalis angulo had employed the definition of opposite branches presented by the Arabic editor and trans­
BGA, quia cadunt in eundem arcum, qui AB, et angulus BGA equalis angulo MHZ, ex lated by Gerard of Cremona. Though, as Unguru suggests, there is some similarity between
premissis, erit ergo angulus ADB equalis angulo MHZ, et patuit prius quod angulus DBG the diagram given by Witelo and that given in 1.60 of the Conics, it is probable that Witelo’s
est equalis angulo HMZ\ erit ergo tertius angulus trianguli DEB, per 32am E, equalis tertio diagram in fact was constructed on the basis of the diagram given in V.34 of Alhazen’s
angulo trigoni MZH, scilicet angulus DEB angulo MZH. Quia ergo trigona DEB et MZH Perspectiva, to which Witelo merely added a further set of opposite branches, thus giving
sunt equiangula, erit, per 4amVE, proportio linee BD ad DE sicut linee MH ad HZ. Ostensum it its appearance of similarity with the diagram for 1.60. But the basic objective of Witelo’s
est autem superius quod est proportio linee GB ad BD sicut linee LM ad MH. Ergo, per treatment in 1.131 of the Perspectiva is quite foreign to that of 1.60 of the Conics (see the
22am V', erit, per equam proportionalitatem, proportio linee BG ad DE sicut linee LM ad ed. of Heiberg, Gr 1, pp. 186-90).
90 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 91

by his definition of opposite branches which he took from Gerard’s trans­ Having already analysed Proposition V.34 of Alhazen’s Perspectiva in
lation of the Arabic text. It ought to be recalled at this point that Alhazen Chapter 1 and presented Witelo’s copy of it in note 34 of this chapter,
in Proposition V.34 of his Perspectiva, the source that led Witelo to I now pass from Book I of Witelo’s Perspectiva to Book IX, where again
mention Apollonius in his comment to his own Proposition 1.131, merely he takes up a conic section, this time, the parabola. The pertinent prop­
held that the first branch could be constructed by the authority of Apol­ ositions are IX.39-IX.44. These propositions relate to the theory and
lonius without specifying Proposition 1.14 in connection with the opposite construction of a paraboloidal mirror and their principal source is Al­
branch. So it appears that Witelo was left up in the air as to the proper hazen’s De speculis comburentibus, which I have discussed at some length
authority to cite. He solved the puzzle in a not too satisfactory way by in Chapter 1. Our paramount concern is with Propositions IX.39 and
turning to the fragment of the Arabic text translated by Gerard. IX.40, in which Witelo acknowledges dependence on Apollonius. Let
Following the construction of the opposite branches of the hyperbola, us first examine IX.39:38
Witelo indicates that Apollonius has proved the corollary, namely, that
TQ = CF, which fact was called to his attention by Alhazen in the latter’s 39. If a straight line is tangent to a parabola, and from the point of tangency
Proposition V.34; and like Alhazen he does not specify the appropriate a straight line is drawn perpendicular to the diameter of the section, [which
proposition of Apollonius, namely 11.16 of the Conics. Hence, Witelo’s diameter] has been extended until it meets the tangent, the part of the diam­
mention of Apollonius here constitutes no evidence that he (Witelo) had eter lying between the perpendicular and the periphery of the section will
read or mastered II. 16 of the Conics. Thus it seems obvious that the exam­ be equal to the part [of it] lying between the section and the tangent.
Let there be a parabola (whose name we have explained earlier in the
ination of the Conics which Witelo might have made in connection with
comment to Proposition 98 of Book I), which [parabola we] let be LAG [see
his Propositions 1.129 and 1.131 was at best superficial and passing, and Fig. 3.22], whose latus rectum [we] let be LG and [whose] diameter [we
confined to adopting the Greek terminology of the Conics. Finally, we let be] AD. And let straight line HBK be tangent to this section in point B\
should note that before presenting Proposition 1.133 (see note 34 above), and let the diameter DA, extended beyond the section, meet the tangent
Witelo singled out in Proposition 1.132 one further geometric step needed line HBK in point H . And from the point of tangency B let a line be drawn
for 1.133, a step that involved the construction of the least line between
the branches.36 In this proposition he did not turn to the Conics, as he not have been available to Witelo (since only the first four books were extant in MS Vat.
no doubt should have, but cited only Proposition III. 10 of the Elements gr. 203, or in fact in any of the known Greek manuscripts). Unguru suggests (ed. cit., p. 204,
of Euclid: “ If a circle cuts a circle, it is necessary that it cuts [it] in only note 2) that Witelo could have proved it by using other propositions of the Conics, like 1.17
and 11.17 (ed. of Heiberg, Vol. 1, pp. 68, 220-22). But all of this is beside the point, for
two places.’’37
in fact Witelo neither proves that TC is the least line in the first part of the proposition
nor that a circle intersects the hyperbola in only two points in the second part of the prop­
36 Ed. of Unguru, pp. 304-05: “ [Propositio] 132. In vertice alterius conicarum sectionum osition.
posito pede circini immobili, secundum quantitatem linee brevissime inter illas sectiones 38 Ed. of Risner, p. 398: “ [Propositio] 39. Si sectionem parabolam linea recta contingat,
ducte descriptus circulus sectionem reliquam continget; secundum vero maiorem, in duobus et a puncto contactus ducatur recta perpendiculariter super diametrum sectionis productam
tantum punctis reliquam secabit. Quod hic proponitur facile est et sola indiget declaratione, ad concursum cum contingente: erit pars diametri interiacens perpendicularem et peripheriam
sint enim, ut in precedenti proponitur, due sectiones conice opposite ad invicem, que sint sectionis aequalis parti interiacenti sectionem et contingentem. Sit sectio parabola, cuius
TP et CU [vide iterum Fig. 3.20], inter quas linea minima, vertices scilicet ambarum sec­ nomen prius libro primo in commento propositionis 98 exposuimus, quae sit LAG [Fig.
tionum continuans, sit linea TC; et posito in altero punctorum T vel C, pede circini, utpote in 3.22], cuius latus rectum sit LG, et diameter AD, contingatque hanc sectionem in puncto
puncto T, describatur circulus secundum quantitatem semidyametri TC. Hic ergo circulus, B linea recta quae sit HBK, concurratque diameter, quae sit DA, producta extra sectionem
quia sectionem CU non attingit nisi in puncto C, et omnes alie linee ducibiles inter ipsas cum linea contingente, quae est HBK, in puncto H, et a puncto contingentiae, quod est B,
sectiones sunt maiores quam linea TC, sunt ergo maiores semidyametro circuli. Secabuntur ducatur, per 12 p 1 linea perpendicularis super diametrum AD, secans ipsam in puncto Z,
ergo omnes per circulum nec attinget circulus alicubi sectionem nisi in puncto C. Patet ergo et sitBZ. Dico quod linea ZA, pars diametri interiacens punctum sectionis perpendicularis
primum propositum. Quod si linea TG, semidyameter circuli, sit maior quam minima BZ et peripheriam sectionis, quae est LAG, est aequalis lineae AH, parti eductae diametri
linearum inter oppositas sectiones productarum, ut est TC, patet quod (correxi ex quoniam) quae interiacet punctum H, quod est punctum concursus diametri cum linea contingente,
illa minima linea intra superficiem sectionis producetur ad periferiam circuli, ut in punctum quae est HBK, et punctum A, quod est terminus diametri cadens in ipsam peripheriam
M. Aliqua ergo superficies communis erit circulo et sectioni. Circulus ergo et sectio se sectionis. Et hoc universale est, etiam si linea recta sectionem contingat in puncto G. Hoc
secabunt. Hec itaque sectio non erit nisi in duobus tantum punctis G et K, quod per modum autem demonstratum est ab Apollonio Pergaeo in libro de Conicis elementis, et hic utemur
10amIII' convinci potest. Patet ergo propositum.” Cf. p. 153. See alsoed. of Risner, pp. 54-55. ipso ut demonstrato. ” I have followed the practice of Unguru for the first book in capitalizing
37 See the version of Campanus' Elementa (Basel, 1546), p. 62: “ Si circulus circulum the letters standing for magnitudes and points, though Risner and five manuscripts give them
secet, in duobus tantum locis secare necesse est." It is obvious, however, that the case as minuscules. I have also altered the punctuation slightly. I have checked Risner’s text
under discussion involves rather the intersection of a circle and a hyperbola. Cf. the comment of this and the succeeding propositions against a few manuscripts and it agrees quite closely
of Unguru (ed. cit., p. 204, note 5). That TC is the least line could be proved by reference with them. Note Risner’s form of expressing Witelo’s citations of Euclid: e.g. 12 p 1 = 12amI1
to V.34 of the Conic s (see above, Chap. 1, note 53). But the fifth book of the Conics would = 1.12 of the Elements.
 MOERBEKE AND WITELO 93
92 ARCHIMEDES IN THE MIDDLE AGES

at Proposition 1.35 of the Conics and that he decided not to reproduce

perpendicular to diameter AD, by 1.12 [of the Elements of Euclid], cutting
it in point Z, and let it [the perpendicular] be BZ. I say that line Z A , the the proof, nor even to specify the proposition number.
part of the diameter lying between the point of intersection of the perpendicu­ One other point worth noting concerning Witelo’s Proposition IX.39
lar BZ [with DA] and the periphery of the section LAG, is equal to line is that, in the figure for this and the succeeding two propositions, the
AH, the extended part of the diameter which lies between point H (the point latus rectum is drawn as if it were the base line of a parabolic segment
of juncture of the diameter and tangent HBK) and point A , which [point] without any further specification as to its length and its determining char­
is the extremity of the diameter falling on the periphery of the section. And acter as the parameter of the parabola. And certainly in none of the figures
this is universally [true], even if the straight line is tangent to the section is LG drawn quadruple AZ, as the propositions demand. Such is the case
in point G. Now, this has been demonstrated by Apollonius of Perga in the not only in the Risner edition but also in the manuscripts I have examined.
book On Conic Elements, and [so] we use it here as [if] demonstrated. I shall return to the confusion of the diagrams when I discuss Proposi­
The demonstration of Apollonius referred to by Witelo is, of course, tion IX.41 below. It is, therefore, no wonder that the author of the later
Proposition 1.35 of the Conics (see above, Chap. 1, note 36). While Al- Speculi almukefi compositio was initially confused when he read Witelo,
hazen made it one of the two cornerstones of his proof of the focal con­ and at first concluded that the latus rectum was any chord cutting off
vergence of lines reflected from the parabola in Proposition 1 of his De a parabolic segment (see the next chapter, Speculi almukefi compositio,
speculis comburentibus (see ibid., note 33), it is of particular interest that proem and the fourth definition). In order to correct this false impression,
Alhazen did not specify that it was demonstrated by Apollonius. Hence Witelo could have done what the author of the later tract did, namely,
the question arises as to how Witelo knew that it was a proposition demon­ to state that he was considering the parabola as generated from a right-
strated by Apollonius of Perga. The simplest and most likely answer is angled right cone and that the latus rectum LG was not any chord but only
that he saw Proposition 1.35 in the text of the Conics. Indeed one might the chord that passed through the midpoint of the segment of the axis
consider this conclusive proof that Witelo had read Apollonius’ text, for of the parabola that extended from the vertex of the parabola to the axis
no other author before him had specifically attributed this proposition of the cone (ibid., Conclusion 4). Or, even better, if he had really read
to Apollonius. I would be more strongly affirmative in expressing this the Conics carefully, he ought to have seen that it was preferable to present
conclusion were it not for the fact that Archimedes introduced the proposi­ his next proposition (IX.40) before IX.39 and in the former proposition
tion without proof as the second proposition of his On the Quadrature o f he ought to have followed Apollonius by defining the latus rectum precisely.
the Parabola and Witelo could easily have read this in Moerbeke’s trans­ But let us see what he did do in his Proposition IX.40, which was the
lation.39 But needless to say, it was not there attributed to Apollonius last proposition in which Witelo acknowledged Apollonius’ authority:41
(who post-dated Archimedes); it was merely said by Archimedes to be 40. Every square of a line drawn from some point on a parabolic section
one of the propositions that were demonstrated “in conic elements,” the perpendicular to the diameter of the section is equal to the rectangle contained
probable reference being to works by Aristaeus and Euclid. It is possible, by (1) the part of the diameter lying between that perpendicular and the
however, that Witelo was unsure of the relative dates of Archimedes and periphery of the section and (2) the latus rectum of the section.
Apollonius and consequently mistook the anonymous reference by Archi­ For example, as in the preceding [proposition, Fig. 3.22], let there be a
medes to be to Apollonius. In such a case, he would then have found an
authority that attributed this proposition to Apollonius. But I doubt this whose newer terminology is everywhere used by Eutocius when he is citing Apollonius
(see the passages in note 6).
argument, for Witelo, as we have said earlier, almost certainly had read 41 Ed. of Risner, p. 398: "[Propositio] 40. Omne quadratum lineae perpendicularis ductae
Eutocius’ Commentary on the Sphere and the Cylinder o f Archimedes ab aliquo puncto sectionis parabolae super diametrum sectionis est aequale rectangulo con­
in the Moerbeke translation and he could have probably deduced there­ tento sub parte diametri interiacente illam perpendicularem et peripheriam sectionis et sub
from that Archimedes’ treatment of conic sections represented an earlier latere recto ipsius sectionis. Verbi gratia, sit, ut in praemissa, sectio parabola, quae sit
stage than Apollonius’.40 So, I suppose that Witelo had at least glanced LAG [vide iterum Fig. 3.22], cuius latus rectum sit LG, et eius diameter sit AD, et a puncto
aliquo sectionis, quod sitR, ducatur super diametrum sectionis, quae est AD, perpendicularis
BZ. Dico quod quadratum lineae perpendicularis, quae BZ, est aequale ei rectangulo quod
39 Clagett, Archimedes in the Middle Ages, Vol. 2, 20vM -N: “ [2.] Si sit rectanguli
fit ex ductu lineae ZA, quae est pars diametri AD interiacens ipsam perpendicularem BZ
coni portio que ABG, sit autem que quidem BD apud dyametrum vel ipsa diameter, que
et peripheriam sectionis, in lineam LG, quae est latus rectum ipsius sectionis. Est ergo,
autem ADG apud eam que secundum B contingentem sectionem coni, que autem EG con­
per 17 p 6, proportio lineae LG ad lineam ZB sicut ipsius ZB ad lineam ZA. Hoc autem
tingens portionem coni apud G, erunt que BD, BE equales. [Fig. 3.23], . . . Demonstrata
similiter demonstratum est ab Apollonio Pergaeo in libro de Conicis elementis, et nos ipso
sunt autem hec [i.e., theoremata 1-3] in elementis conicis.”
utemur ut demonstrato. Haec vero duo theoremata cum aliis Apollonii theorematibus in
40 Witelo might possibly have deduced that Archimedes was earlier than Apollonius
principio libri non connumeravimus, quia solum illis indigemus ad theorema subsequens
from the whole passage, of which I have quoted a part in note 17 above, for it implies
explicandum et in nullo aliorum theorematum totius huius libri.”
that since he used the earlier terminology of conics he ought to have lived before Apollonius,
94 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 95

parabolic section LAG , whose latus rectum [we] let be LG and whose diameter sectionis, omnis linea aequidistanter diametro incidens alicui puncto sectionis, et linea ab
[we] let be AD ; and from some point of the section, [point] B, let perpendicular eodem puncto sectionis ad punctum abscissionis diametri producta, cum linea contingente
BZ be drawn to the diameter of the section, AD. I say that the square of sectionem super illud punctum, continent angulos aequales. Sit, ut superius, sectio parab­
perpendicular line BZ is equal to the rectangle produced from the multiplica­ ola, quae LABG [Fig. 3.24], cuius diameter sit AD, et eius latus rectum sit LG; ab ex­
tremitate quoque diametri AD ex parte peripheriae sectionis, hoc est, a parte puncti A,
tion of line ZA (which is the part of the diameter AD lying between this
resecetur, per 3 p 1, linea AE aequalis quartae parti lateris recti ipsius sectionis, quod est
perpendicular BZ and the periphery of the section) into line LG (which is LG. Incidatque linea TB puncto sectionis, quod est B, aequidistanter diametro AD, et
the latus rectum of the section). Therefore, by VI. 17 [of the Elements of continuetur linea a puncto B ad punctum E, quod separat a diametro AD lineam AE aequalem
Euclid], line LG / line ZB = line ZB / line ZA. But this has been similarly quartae parti lineae LG, et ducatur a puncto B linea contingens sectionem, quae sit HBK. Dico
demonstrated by Apollonius of Perga in the book On Conic Elements, and quod duae lineae TB et BE cum linea sectionem contingente, quae est HBK, in puncto B
[so] we [here] use it as [if] demonstrated. Now we did not enumerate these continent angulos aequales, ita quod angulus TBK est aequalis angulo EBH. Angulus enim
two theorems [IX.39 and IX.40] with the other theorems of Apollonius in BEH non potest evadere unam trium conditionum. Aut enim erit acutus aut rectus aut ob­
the beginning of the work because we need them only for explicating the tusus. Sit primo acutus. Et a puncto B ducatur, per 12 p 1, super diametrum AD perpen­
subsequent theorem [IX.41] and not for any other theorems in the whole work. dicularis BZ, cadetque, per 32 p 1, punctum Z inter duo puncta A et E, et producatur
diameter AD ultra punctum A donec, per 2 th. 1 huius, concurrat cum linea contingente
The enunciation, as given by Witelo, is essentially equivalent to the sectionem, quae est K B H , sitque concursus in puncto H. Eritque angulus AHB acutus.
apodosis of Proposition 1.11 of the Conics of Apollonius (see above, Cadet ergo perpendicularis BZ inter puncta H et E, et erit, per 39 huius, linea AZ aequalis
lineae AH. Quia itaque linea AE est divisa in puncto Z, et ei est aequalis uni parti dividentium
Chap. 1, note 22), except that (and this is a crucial exception) the latus
adiecta, quae est A H , erit ergo, per 8 p 2, quadratum lineae EH aequale ei quod fit ex
rectum is not defined in Witelo’s enunciation. I would suppose that the ductu lineae EA in lineam HA, vel in lineam AZ, quater et quadrato lineae ZE. Sed linea
enunciation was in fact devised by Witelo after studying the citation of EA est quarta pars lineae LG ex hypothesi. Ergo, per 1 p 2, vel per 1 p 6, illud quod
this proposition of Apollonius in Proposition 1 of Alhazen’s De speculis fit ex ductu lineae AZ in lineam AE quater est aequale ei quod fit ex ductu lineae AZ
comburentibus {ibid., note 33). Thus Alhazen gave Witelo his first inkling in lineam LG semel. Illud ergo quod fit ex ductu lineae AZ in lineam LG cum quadrato
lineae ZE est aequale quadrato lineae EH. Sed, per praemissam, patet quod illud quod
that the proposition was one demonstrated by Apollonius. The form of
fit ex ductu lineae AZ in lineam LG est aequale quadrato lineae BZ, quoniam linea BZ
Witelo’s enunciation and the form of Alhazen’s citation are similar in their est perpendicularis super diametrum AD, duo vero quadrata BZ et ZE sunt, per 47 p 1,
omission of any definition of the latus rectum. Indeed the only connection aequalia quadrato lineae BE. Quadrata ergo linearum EH et EB sunt aequalia. Ergo linea
that Witelo’s proposition shows with Apollonius’ Proposition 1.11 is the EB est aequalis lineae EH. Ergo, per 5 p 1, in trigono EBH angulus EHB est aequalis
fact Witelo used the Greek term parabola. His omission of the definition angulo EBH. Sed linea TB et DA sunt aequidistantes. Ergo, per 29 p 1, angulus TBK
extrinsecus est aequalis DHB intrinseco; angulus ergo EBH est aequalis angulo TBK. Eodem
of the latus rectum makes it almost certain that he made no effort to study
quoque modo demonstrandum de qualibet linea aequidistante diametro AD et BE linea copu­
1.11 of the Conics. The only other point of interest in Witelo’s brief ac­ lata ad punctum E, quando illa linea super punctum £ cum diametro AD angulum continet
count of IX.40 is his restatement of the proposition with the perpendicular acutum. Patet ergo propositum secundum hunc modum. Quod si angulus BEH fuerit rectus
(ordinate ZB) expressed as the mean proportional between the latus rectum [Fig. 3.25], adhuc patet propositum, quod angulus TBK est aequalis angulo EBH. Quoniam
LG and the abscissa Z A . As we shall see in the next chapter, this is the enim angulus BEH est rectus, patet quod linea BE est perpendicularis super diametrum AD.
Ergo linea EA, per 39 huius, est aequalis linee AH. Sed linea EA ex hypothesi est quarta
form of the theorem embraced in the second part of Conclusion 4 of the
pars lineae LG. Ergo linea HE, quae est dupla lineae AE, est medietas lineae LG. Ergo,
Speculi almukefi compositio, written after the Perspectiva. But there the per 4 p 2, quadratum lineae EH est quarta pars quadrati lineae LG. Id quoque quod fit
whole proposition is sounder because of the explicit definition of the latus ex ductu lineae EA in lineam LG est aequale quartae parti quadrati lineae LG, per 1 p 6,
rectum given in the first part of the conclusion. quoniam linea EA est ex hypothesi quarta pars lineae LG. Illud ergo quod fit ex ductu
With Propositions IX.39 and IX.40 stated, Witelo follows the lead of lineae EA in lineam LG est aequale quadrato lineae EH. Sed id quod fit ex ductu lineae
EA in lineam LG est aequale quadrato lineae EB, per praemissam, quoniam linea EB est
Proposition 1 of Alhazen’s De speculis comburentibus and proves in his
perpendicularis super diametrum AD. Quadratum ergo lineae EH est aequale quadrato
Proposition IX.41 that any line parallel to the axis of a parabola constructed lineae EB. Ergo linea EH est aequalis lineae BE. Ergo, ut prius per 5 p 1, anguli EBH
to reflect from the parabola to a point on the axis distant by !4 the latus et EHB sunt aequales. Et quoniam linea TB aequidistat lineae AD, patet, per 29 p 1, quod
rectum from the parabola’s vertex produces equal angles with the tangent (correxi ex quoniam) angulus TBK est aequalis angulo EBH. Et simili ter demonstrandum
to the parabola at the point of reflection, which is tantamount to proving de omni linea incidente ipsi sectioni, cum angulus BEH est rectus, et illud est quod pro­
ponebatur. Si vero angulus BEH sit obtusus [Fig. 3.26], dico quod adhuc angulus TBK
that the specified point on the axis is the focus of the parabola. Since
est aequalis angulo EBH. Ducatur enim linea perpendicularis, quae sit BZ, a puncto B ipsius
I have analysed that proposition in Chapter 1, I need here give only the sectionis, cui incidit linea aequidistans diametro AD, quae est TB, illa quoque perpendicu­
text of Witelo’s proposition, without translation or analysis.42 But I should laris super diametrum AD sit BZ; cadetque haec perpendicularis BZ inter puncta diametri
quae sunt D et E, alias enim duo anguli unius trigoni BEZ fierent maiores duobus rectis,
42 Ed. of Risner, pp. 399-400: “ [Propositio] 41. Si in sectione parabolae ab extremitate quoniam uno existente recto, qui BZE, angulus BEZ esset obtusus, quod est impossibile.
diametri ex parte peripheriae sectionis resecetur aequale quartae parti lateris recti ipsius Cadit ergo punctum Z inter puncta E et D. Linea ergo AZ est maior quam linea AE. Et
96 ARCHIMEDES IN THE MIDDLE AGES MOERBEKE AND WITELO 97

point out to the reader that Witelo, in his proofs of all three cases, aban­ investigation of the mathematical considerations of conic sections and I
doned Alhazen’s procedure of giving both a proof by analysis and a proof note them here only that the interested reader may find them and compare
by synthesis for each case. The Polish optician settled for synthetic proofs. them with the similar propositions given in the text of the Speculi al­
Furthermore, the positioning of LG as the latus rectum in the figures for mukefi com positio edited in the next chapter and with the remarks given
Proposition IX.41 (Figs. 3.24-3.26) is erroneous. Such is the case for by Jean Fusoris some decades later.
the figures given by Risner, which represent quite faithfully the figures What does our preceding analysis of Witelo’s treatment of conic sec-
given in the manuscripts of the Perspectiva that I have examined, except
that in the latter the parabolas are formed by arcs of circles. In every sectionis quam facit accidere, tunc omnis linea producta equidistans eius axi et perveniens
case the line of order that constitutes the latus rectum ought to pass through ad superficiem concavam et convertitur ad illud punctum continet cum linea contingente
the focal point E. It is no wonder that the author of the Speculi almukefi superficiem concavam que est differentia communis inter superficiem linee converse et inter
superficiem contingentem superficiem concavam duos angulos equales. . . . 4. Cum op­
com positio was so confused when he began his search for the proofs of
ponitur corpori solis omnis superficies specularis concava concavitate corporis mukefi ita ut
Apollonius’ propositions (see below, Chap. 4) or that Jean Fusoris was sit axis eius in directo corporis ipsius, tunc egrediuntur ex corpore solis ad totam planitiem
highly critical of the figures of this proposition (see below, chap. 5). I eius radii qui omnes convertuntur ad punctum super eius axem cuius elongatio a capite
should also note that here I shall merely locate, without presenting in superficiei per quantitatem quarte lateris erecti sectionis quam efficit superficies. . . . 5.
full, Propositions IX.42-IX.44 of Witelo’s P erspectiva , which are similar Et quia iam ostensum est quod radii que egrediuntur ex corpore solis ad omnem super­
ficiem speculi concavi concavitate corporis mukefi qui sunt equidistantes axi convertuntur
(but by no means identical) to Propositions 2 -5 of Alhazen’s De speculis
omnes ad punctum unum, tunc ostendam nunc qualiter assumamus speculum quod fit se­
com burentibus.43 None of these propositions holds any interest for my cundum hanc figuram. Assumemus ergo laminam calibis boni secundam quam mensuram
voluerimus. Tunc fit sicut laminaABGD. Et protrahamus in ea sectionem ex sectione mukefi,
quoniam linea HBK contingit sectionem, et linea BZ est perpendicularis super diametrum quecunque sectio sit, et sit sectio AEG, et abscindamus laminam super lineam AEG. Qualiter
AD, erit, per 39 huius, linea AZ aequalis lineae AH. Ergo linea HA est maior quam linea vero inveniamus sectionem mukefi et alias sectiones per viam instrumenti multitudo geo­
AE\ fiat, per 3 p 1, linea AM aequalis lineae AE. Remanet ergo linea HM aequalis lineae metrarum variavit, quamvis non invenerunt eam secundum veritatem sui, et ostendam in
ZE. Linea ergo EM addita utrobique, erit linea ZM aequalis lineae HE. Quadratum ergo tractatu, in quo narrabo inventionem omnium sectionum per viam instrumenti, qualiter
lineae ZM est aequale quadrato lineae EH. Quia itaque linea ZA est divisa in puncto E , et ei protrahamus quamcunque sectionem volumus, secundum veritatem instrumenti, qua veri­
est adiecta aequalis uni dividentium, quae est MA, aequalis ipsiAE: patet, per 8 p 2, quod orem non sit possibile invenire in materia, sicut est inventio circuli per circulum, quamvis
illud quod fit ex ductu lineae ZA in lineam AM, vel in eius aequalem lineam AE, quater sit illud valde laboriosum, et super quamcunque superficiem planam voluerimus, et sit
cum quadrato lineae ZE est aequale quadrato lineae ZM, vel lineae EH, quae sunt aequales. angulus ordinis eius, quemcunque angulum voluerimus, et latus eius erectum, quamcunque
Sed illud quod fit ex ductu lineae ZA in lineam AE quater, ut patet ex praemissis, est lineam volumus, et quamcunque sectionem voluerimus ex sectione, scilicet si voluerimus
aequale ei quod fit ex ductu lineae AZ in lineam LG, per 1 p 2 vel per 1 p 6, quoniam linea ab eo quod sequitur capud eius, aut voluerimus ex medio ipsius, et erit elongatio eius
AE est aequalis quartae parti lineae LG ex hypothesi. Illud ergo quod fit ex ductu lineae ex capite ipsius, quamcunque elongationem voluerimus, apparebit ergo per illud qualiter
AZ in lineam LG cum quadrato lineae ZE est aequale quadrato lineae EH. Sed illud quod inveniamus in lamina sectionem mukefi, et nisi abhorrerem prolongare librum et permiscere
fit ex ductu lineae ZA in lineam LG est aequale quadrato lineae BZ, per praecedentem, cum eo quod non est eius, dicerem in hoc loco. . . .” (I have included the beginning of
quoniam linea BZ est perpendicularis super diametrum AD, quadratum vero lineae BE, the proof of Proposition 5 as well as the enunciation to support a later statement in Chapter
per 47 p 1, est aequale quadratis ambarum linearum BZ et EZ. Patet ergo quod quadratum 4 below, note 9.) I have changed the punctuation extensively. For the similar Propositions
lineae BE est aequale quadrato lineae EH. Ergo linea EB est aequalis lineae EH. Ergo, IX.42-IX.44 of Witelo's Perspectiva, see the ed. of Risner, pp. 400-403. Again I give
per 5 p I, anguli EBH et AHB sunt aequales. Sed, ut prius, lineae TB et DH sunt aequidis- here only the enunciations: "42. In omni superficie concava concavitatis sectionis parabolae,
tantes. Angulus ergo TBK, per 29 p 1, est aequalis angulo DHB, ergo et angulus EBH. Et simi­ si ab extremitate axis contingentis sectionem abscindatur pars aequalis quartae lateris recti
liter demonstrandum in omni linea incidente sectioni aequidistanter diametro AD, cum angu­ ipsius parabolae: omnis linea aequidistanter axi incidens illi superficiei et linea a puncto
lus BEH est obtusus. Patet itaque generaliter propositum. Nam omnis linea incidens peri- incidentiae ad punctum signatum in axe producta cum linea in illo puncto superficiem con­
pheriae sectionis aequidistanter diametro et alia linea quae ab illo eodem puncto ducitur ad tingente angulos aequales. . . .4 3 . Speculo concavo concavitatis sectionis parabolae soli
punctum abscindens a diametro ex parte peripheriae sectionis partem aequalem quartae parti opposito ita ut axis ipsius sit in directo corporis solaris, omnes radii incidentes speculo
lateris recti ipsius sectionis cum linea sectione in illo puncto contingente continent angulos aequidistanter axi reflectuntur ad punctum unum axis distantem a superficie speculi secun­
aequales. Et hoc proponebatur."
dum quartam lateris recti ipsius sectionis parabolae speculi superficiem caussantis (/). Ex
4:) For the text of Propositions 2 -5 of Alhazen's De speculis comburentibus, see J. L. quo patet quod a superficie talium speculorum ignem est possibile accendi . . . 44. Specu­
Heiberg and E. Wiedemann, “ lbn al Haitams Schrift fiber parabolische Hohlspiegel," Bib­ lum secundum formam sectionis parabolae, vel lineae eccentralis, vel intersectionis pyra­
liotheca mathematica, 3. Folge, Vol. 10, pp. 224-31, of which I give here only the enuncia­ midalis, vel cuiuscunque alterius regularis vel irregularis datae lineae artificialiter con­
tions: "2. Omnis sectionis mukefi cuius axis figitur et revolvitur sectio donec redeat ad stituere." I have altered the punctuation somewhat in these enunciations. Notice that Propo­
locum a quo incepit per motum, et provenit corpus rotundum, et evenit in corpore con­ sition 2 of Alhazen's De speculis comburentibus is essentially the same as Archimedes',
tento ab ea, quodcunque corpus fuerit, superficies concava, et omnis superficiei plane que De conoidalibus, Prop. 1la, which is given by Archimedes without proof. So far as I know,
egreditur super eius axem et secat illam superficiem differentia communis est sectio mukefi no medieval author mentioned this similarity, although the Archimedean proposition was
equalis sectioni prime, et axis eius est ille axis. . . . 3. In omni superficie concava con­ available in William of Moerbeke's translation (see Clagett, Archimedes in the Middle Ages,
cavitatis corporis mukefi si de extremitate axis eius separetur simile quarte lateris erecti Vol. 2, 47rV-W).
98 ARCHIMEDES IN THE MIDDLE AGES

tions add up to? (1) Witelo was certainly familiar with and used the ma­ CHAPTER 4
terial on conic sections that had been made available by translations
from the Arabic of the introductory material from Apollonius’ Conics
and of the De speculis com burentibus and Perspectiva of Alhazen. Indeed,
in terms of substantial knowledge of conic sections he scarcely went
beyond this material translated from the Arabic. (2) On the other hand,
he did use in rudimentary fashion the Greek terminology associated with
conic sections, obtaining it from one or more of the following texts: Euto-
The Speculi almukefi compositio
cius’ C om m entary on the Sphere and the Cylinder o f A rch im e d es , Apol­
lonius’ Conics and Eutocius’ commentary on that work. (3) But if he saw
these three texts (it is almost certain that he saw the first of them, while Within a generation or so after the preparation of Witelo’s Perspectiva
it is less certain but still quite probable that he saw the second, and it is an unknown monk produced a new work on the parabolic mirror that
at least possible that he saw the third), he read them in the most superficial was one of the most original tracts on conic sections composed by a
fashion, without mastering any of the proofs that he acknowledges as medieval Latin author. For the sake of convenience and to distinguish
Apollonius’. Only a profoundly confused student of mathematics could it from Alhazen’s De speculis com burentibus I have adopted the title of
have ignored so strikingly an important aspect of the parabola as the this tract that appears in two manuscripts (O and F in the sigla below):
nature and definition of the latus rectum. Speculi alm ukefi com positio, though the original title was perhaps De
So, in brief, we affirm that Witelo collected much of the scattered ma­ speculis co m burentibus , as given in the oldest manuscript (C, see the text
terial on conic sections available at his time, covering it with a Greek below, variant readings for the title; but this title given on folios 233v
veneer, and that he set the stage for the first Latin mathematician to treat and 234r may both times be in a later hand). Though the tract cannot be
conic sections in an unambiguous way. This was the unknown author dated precisely, it is evident that it was composed after Witelo’s work,
of the Speculi alm ukefi co m positio , who is the subject of the next chapter. which it cites (see the Proem of the tract below). On the other hand, the
part of the above-noted oldest manuscript containing this tract (C) cannot
be dated much later than the middle of the fourteenth century, since it
was apparently part of a codex owned by the well-known Oxford scholar,
Simon Bredon,1 who was a fellow of Merton College as early as 1330,
remaining there until about 1348, and dying after a long church-oriented
career in 1372.2
The name of the author is unknown, or at least uncertain. Two manu­
scripts (O and F ) attribute the work to Roger Bacon. But such an attribu­
tion is very doubtful. First, because the only early manuscript (C) has
no such attribution. Second, as I have suggested in an earlier place, be­
cause Bacon’s works reveal no interest in formal mathematical proofs,
a fact also remarked on by the sixteenth-century editor of a shortened
version of this tract, A. H. Gogava.3 The ascription of the tract to Bacon
could easily be explained by the fact that in the earlier of the two manu­
scripts (O) our tract follows immediately after Bacon’s De com positione
m edicinarum .4 1 suggest, then, that perhaps the scribe of this manuscript,

1 See the article of A. G. Watson cited below under the Sigla, MS C.

2 A. B. Emden, A Biographical Register of the University of Oxford, Vol. 1 (Oxford,
1957), pp. 257-58.
3 M. Clagett, “ The Life and Works of Giovanni Fontana,” Annali dell’Istituto e Museo
di Storia della Scienza di Firenze, Anno I (1976), p. 19, n. 49.
4 Ibid., p. 19. In MS O (see the Sigla below), 47r: “ Explicit tractatus de composicione
medicinarum per fratrem rugerium bacon editus. Speculi almukefi (? abnukefi?) composicio
secundum Rugerium bacon ordinis minorum.”

99
100 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UK EFI COMPOSITIO 101

faced with an exemplar that assigned Bacon’s name to De compositione including the fragment of Apollonius’ Conics translated by Gerard of
medicinarum but left unassigned the succeeding Speculi almukefi com­ Cremona, Alhazen’s De speculis comburentibus, the anonymous tract on
positio, decided that Bacon was also the author of this tract. But even the hyperbola De duabus lineis and Witelo’s Perspectiva. In addition to
if the treatise was not composed by Bacon, it does appear to have been mentioning the links that connect the Speculi almukefi compositio with
written by a monk, since the author speaks in the Proem of diligently past treatises, I should note that the tract exercised some significant in­
seeking information about a possible copy of Apollonius’ Conics from fluence on later treatments of the geometry of parabolic mirrors. That
“ brothers from distant parts.’’ Finally, regarding this author, there is some influence was exerted mainly by the somewhat shortened version of the
slim evidence that he was interested in astronomy as well as geometry, tract apparently prepared by Regiomontanus in the middle of the fifteenth
since he introduces into the proof of Conclusion 11 astronomical terminol­ century and published therefrom by A. H. Gogavain 1543.1 give a separate
ogy that is quite unnecessary to the proof but might well reflect an au­ text of this shortened version in the next chapter.
thor customarily used to orienting figures in an astronomical way. Also Now let us look at the Speculi almukefi compositio more closely. The
the quasi-trigonometric character of the author’s original proof of Con­ Proem is remarkably interesting. The author tells us there that he was dis­
clusion 11 (he uses sines and versines of arcs) may well bespeak an turbed by the fact that perspectivists (only Witelo is named) cited the
astronomical author accustomed to applying trigonometric procedures. aforementioned propositions of Apollonius (1.11 and 1.35) without ever
I have alluded to the originality of the Speculi almukefi compositio proving them. Hence he tried mightily to locate a copy of Apollonius’
above. My detailed analysis of the tract will support this statement. But treatise, contacting his brothers and others that he might by their media­
for the present I can merely note that the author of our tract, after the tion come upon this book or upon some revealing discussion on the matter
initial confusion to which he refers in the Proem, set out to prove the two of these proofs. Finding neither, he decided to prove them himself, but
propositions of Apollonius’ Conics (1.11 and 1.35) that had been used by without initial success. The apparent stumbling block (as we learn later in
Alhazen in hisDc speculis comburentibus and by perspectivists (including his discussion of the fourth definition) was that he misunderstood the cor­
Witelo) without proof.5 In constructing these proofs, our author became rect nature of Apollonius’ latus rectum, a misunderstanding which un­
the first to do so in the Latin language, unless, of course, Apollonius’ doubtedly originated in the diagrams of Proposition IX.41 of Witelo’s
own proofs had previously appeared in a Latin translation of the Conics. Perspectiva (see my discussion of that proposition in Chapter 3 above).
But even if a translation of the Conics had been made by Moerbeke He apparently first thought that a latus rectum of a parabola was any
or somebody else, it is obvious that our author had not seen such a straight line or chord that acted as the rectilinear base of a parabolic
translation and that his proofs were entirely independent. In the course segment. With such a false assumption, it is no wonder that he initially
of struggling with these proofs the author of the Speculi almukefi com­ concluded that the Apollonian propositions were false. However, he con­
positio was able to show that the latus rectum of a parabola formed from sidered the matter further and decided that Apollonius and those citing
a right-angled right cone was in fact equal to a line which is double the him would hardly have gone so awry as to accept these conclusions
axis of the parabola drawn as far as the axis of the cone, a parameter if they were false. He was particularly impressed by the fact that Campanus
often used by Archimedes.53 One further important original element of the had cited Apollonius in his version of the Elements and called him there
Speculi almukefi compositio is found in the proof of Conclusion 11, in the “Great Apollonius.’’ (See above, Chapter 3, note 3.) So our author
which the author abandons the earlier proof found in the De duabus lineis persevered and found a way to interpret latus rectum that led him to sound
and based on the axial property of a hyperbola for a proof that uses proofs of the propositions. Not wishing to have labored in vain, he de­
right-sines and versines to prove the asymptotic property of a hyperbola, cided by means of these propositions and others to compose a tract that
as we shall see. would “ teach how to make a burning mirror that would burn at any
Though I have just signaled out some original elements in the treatise, distance.”
I should remind the reader that the author displays a thorough knowledge The author initially presents four definitions. The first one was suggested
of the various tracts we have already examined in the first three chapters, to him by the Proem of the anonymous tract De duabus lineis that I
have edited in Chapter 2, where it was also the first definition mentioned.
5 The reader should observe that the most popular of the perspectivists, John Pecham,
Here, as there, it is attributed to Euclid, Elements, Book XI, and defines
has only a short, non-mathematical reference to the paraboloidal mirror and to the De
speculis comburentibus [of Alhazen]. See D. C. Lindberg, John Pccham and the Science a [right] cone by the rotation of a right triangle about one of the sides
o f Optics (Madison, 1970), p. 208. containing the right angle. But our author, apparently emulating Witelo
5a T. L. Heath, The Works o f Archimedes (Cambridge, 1897; Dover repr., New York, in 1.89 of the Perspectiva (see above, Chap. 3, note 8), added the further
1950), pp. clxvii-clxviii.
distinctions included by Euclid that a right-angled cone is formed if the two
102 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UK EFI COMPOSITIO 103

sides including the right angle are equal, an acute-angled cone if the MSS OFV give a passage that identifies a sectio rectangula with a parabola
fixed side is greater than the moving side and an obtuse-angled cone if and a mukefi, a sectio ambligonia with a hyperbola and a so-called mukefi
the moving side is greater than the fixed side.6 addita, and finally a sectio oxigonia with an ellipse and a mukefi diminuta
The second definition is that of the section contained by a curved line (see the text below, variant reading to line 37 of the Descriptio prima etc.).
described in the surface of a cone by a plane cutting the cone but not The source of this intrusion is clearly Propositio 1.98 of the Perspectiva
through its apex. This is essentially the definition given in the fragment of Alhazen (see above, Chap. 3, note 15). I have already commented on
translated by Gerard of Cremona from the Arabic text of Apollonius (see the erroneous appearance of mukefi in the designations of hyperbola and
above, Chapter 1, note 18, def. [1]), a definition not specifically given by ellipse in the Witelonian passage. I shall say here only that I have rejected
Apollonius but by the Arabic editor of the text of the Conics. It seems, this addition as a genuine part of the original Speculi almukefi compositio,
however, that the author (or some early scribe of the text of the Speculi not only because it does not appear in the oldest manuscript (C) but
almukefi compositio) had not correctly read Gerard’s transliteration of also because the false appearance of mukefi in the Arabic-based terms
the Arabic term for such a curved line: linea munani, for he appears to for hyperbola and ellipse would be immediately at variance with the correct
have settled for linea muniani, though later scribes continued to spell the forms of sectio addita and sectio diminuta given by the author earlier in his
Arabic term in a wild variety of ways.7 While our author almost certainly statement concerning their construction.
took this definition from Gerard’s fragment, it is of interest that this defini­ The fourth and final definition concerns the author’s definition of the
tion was also the second definition given in the Proem of the De duabus latus rectum of a parabola and the discussion is entirely original with him.
lineis, where, however, the line is called a. linea obliqua or obliqua linea He first points out that any line of order of the parabola could be called a
(see above, Chap. 2, text of the Proem). At any rate, when our author “ right base’’ or latus rectum and, as I said earlier, this erroneous view of
speaks of it as a definition given by Apollonius he is following the er­ the latus rectum was in all likelihood the source of the author’s original
roneous assertion found in the De duabus lineis and implied by Gerard difficulty in attempting to prove the Apollonian propositions. But the au­
when he speaks of his whole fragment as being “ in the beginning of the thor here goes on to say that this initial definition is not Apollonius’.
Liber de pyramidibus of Apollonius’’ (see above, Chap. 1, note 3). For as I Rather, the latus rectum is specified as that line of order that passes
remarked earlier this definition was one given by the Arabic editor rather through the midpoint of the segment of the parabola’s axis that extends
than by Apollonius. from the vertex of the parabola to the axis of the cone. Now this is true
The third definition (called the First Presupposition in MS O) is con­ only for a parabola formed from a right-angled right cone, and so we see
cerned with the formation of an axial triangle in a cone and with the the significance of his use of that kind of a cone in the third definition,
construction of the three conic sections: parabola, hyperbola and ellipse. a point that the author does not make but no doubt understood or he
It is directly dependent on the oft-mentioned fragment translated by Gerard would not have limited his parabola to one formed from a right-angled
of Cremona from the Arabic version of the Conics (see above, Chap. right cone. After presenting his version of the definition of a Iatus rectum,
1, notes 18 and 21). However, our author makes one change of great he then remarks that any segment of a parabola cut off between its vertex
significance for his later treatment of the latus rectum. Though following and any line of order is itself a parabola that has the same latus rectum as
the Apollonian constructional procedures for describing the conic sec­ the original parabola. He states, but reserves for later proof in Conclusion 4,
tions, he specifies that the cone to be cut is a right-angled right cone. that the latus rectum of the parabola (so described) “ will necessarily be
As we shall see, the purpose of this addition is to assure that the latus double the axis of the parabola, i.e., the axis extended from the vertex
rectum of a parabola constructed from such a cone is indeed double of the section until it meets the axis of the cone.’’ As I have said earlier,
“the line to the axis of the cone.’’ To these instructions that define the this is the parameter often used by Archimedes in treating parabolas
three conic sections, our author adds further definitions that are termino- formed from right-angled cones. The author closes his discussion of the
logically his own, namely those of the axis or “ arrow’’ of a parabola, fourth definition by revealing clearly that his initial misunderstanding of
its vertex and its “lines of order’’ (i.e. its double ordinates). However, latus rectum lay at the root of his early difficulties with the proof of the
they no doubt depend in a general way on the definitions given in the Apollonian propositions (“ I was never able to come to this definition of
beginning of Gerard of Cremona’s fragment (see above, Chap. 1, note 5, the latus rectum of a parabola from anything I found cited from Apollonius
def. [4]). At the end of all of this matter included under the third definition, but only by the circumstance that I demonstrated with great zeal and
labor that the conclusions of Apollonius were false for any other line of
order and that they were true for this one alone.’’).
KSee the Campanus version of Euclid in Euclidis Megarensis . . . Elementorum geo­
metricorum libri -xv (Basel, 1546), p. 346 and note 8, Chapter 3 above for Witelo. Following the four definitions, the author of the Speculi almukefi com­
7 See my text below, variant readings, Descriptio prima etc., lines 11, 18, 25, 30, 31. positio presents his propositions, or “ conclusions’’ as he calls them. The
104 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 105

first conclusion (already anticipated in the third definition) demonstrates tion from Proposition IX.40 of Witelo’s Perspectiva, where in the com­
that a section through the apex of a cone and the center of its base is a ment Witelo adds this form to the Apollonian form (see above, Chap. 3,
triangle. It is similar to Proposition 1.3 of Apollonius’ Conics (see above, note 41). The second part of this conclusion comprises the first Apollonian
Chap. 1, note 20). But our author’s limitation of it to an axial triangle proposition to be proved by our author. These two parts of Conclusion
leads me to suppose that he fashioned the conclusion from Proposition 3 4 are easily proved by using the basic theorem that a perpendicular from
of the De duabus lineis (see Chap. 2) and/or from the remarks found in the circumference of a circle to its diameter is the mean proportional
Propositions 1.89 and 1.90 of Witelo’s Perspectiva (see above, Chap. 3, between the segments of the diameter produced by the perpendicular and
notes 8 and 11). One element of novelty is the appeal here made to Proposi­ by using simple propositions concerning parallel lines and similar triangles.
tion 1 of the Liber de curvis superficiebus Archimenidis in support of There is no need to repeat the proofs readily available in the text and
the statement that the two sides of the axial triangle are called “ hypot­ translation below. From the first part of Conclusion 4 the author deduces
enuses” of the cone.8 I have already referred to the popularity of this that the latus rectum is four times its axial intercept (when the latus
treatise in Chapter 2 above (see note 12). rectum is presented as the line of order through the focus), i.e. pg = 4
The second conclusion, that a straight line connecting any two points on fh in Fig. 4.7. This conclusion was used extensively by Regiomontanus
the surface of a cone that are not in line with the apex necessarily pro­ in his added notes to the Speculi almukefi compositio (see Text D in
ceeds inside the cone was taken from Proposition 1 of the De duabus Chapter Five below). Finally, we should note that in the course of the
lineis (see Chap. 2). In fact, it is evident from the appearance of the word proofs of Conclusion 4 it is demonstrated that the squares of the ordinates
vadet in the enunciation instead of incidit and from the details of the proof are to each other as the axial intercepts of the ordinates. This is equivalent
that our author had consulted Version B of the De duabus lineis. The to Proposition 1.20 of the Conics of Apollonius (cf. ed. of Heiberg, Gr
third conclusion, that a plane cutting a right cone— the plane being paral­ 1, p. 72), where it follows from the proof of the basic equation of the
lel to the base of the cone— produces in the curved surface of the cone parabola instead of being a part of its proof as it is in the Speculi almukefi
the circumference of a circle whose center is in the axis of the cone, compositio. In any case, the author of the medieval work had no knowl­
was also taken from the De duabus lineis, and the close similarity of word­ edge of Proposition 1.20 of the Conics.
ing indicates that the source was Version B of that work (see Proposition Before proving the second of the two Apollonian propositions that
2 of the text of Version B in Chap. 2 above). stimulated the composition of this treatise, the author advances and proves
The fourth conclusion is the promised conclusion concerning the latus two preliminary propositions that he later admits are not necessary for
rectum of a parabola. It has two parts: (1) that the latus rectum is double the proof of the second Apollonian proposition. These preliminary proposi­
[the segment of] the axis of a parabola that extends from the vertex of tions are Conclusions 5 and 6 of the Speculi almukefi compositio. In the
the parabola to the axis of the cone, and (2) that the latus rectum is to a fifth conclusion it is easily and quickly proved that in the case of a circle
perpendicular drawn to the axis of the parabola from the parabolic curve drawn in the plane of a parabola through the vertex of the parabola and
as that perpendicular is to the segment of the axis intercepted between the extremities on the parabolic curve of any line of order between the
the parabola’s vertex and the perpendicular. The first part is the Archi­ vertex and its latus rectum (conceived as in the third definition), the
medean parameter for a parabola formed from a right-angled right cone center of the circle lies on the axis of the parabola and the distance on
already referred to and is completely distinct from the specification of the the axis between the said line of order and the circumference of the circle
parameter given in the enunciation of Proposition 1.11 of Apollonius’ in the direction of the base of the cone will necessarily be equal to the
Conics. Though the second part is essentially equivalent to the apodosis of latus rectum of the parabola. This conclusion and its converse were very
the enunciation of Apollonius’ Proposition 1.11, here it is presented as a important in the later history of the properties of a parabola and particularly
proportion with the perpendicular as a middle term, while there it was of the drawing of a parabola in a plane when the vertex, the axis and the
given as an equation of the square of the perpendicular with a rectangle latus rectum are given (see Johannes Fusoris’ Libellus, Chap. 5 below,
contained by the latus rectum and the abscissa (see above, Chap. 1, n. 22). Text A, Chap. 4, property 1; Regiomontanus’ notes to the Speculi al­
The author no doubt borrowed the proportional mean form of this enuncia­ mukefi compositio, Chap. 5 below, Text D, passage [1]; Johann Werner’s
Libellus, Chap. 6 below, Text, Element XI; and O. Fine’s De speculo
8 M. Clagett, Archimedes in the Middle Ages, Vol. 1 (Madison, Wise., 1964), p. 450. ustorio, Chap. 7 below, Text A, Prop. IIII, corr. 3). In the sixth con­
“ I. CUIUSLIBET ROTUNDE PIRAMIDIS CURVA SUPERFICIES EST EQUALIS TRI­
ANGULO ORTHOGONIO, CUIUS UNUM LATERUM RECTUM ANGULUM CON­
clusion the author shows that if we take a point on the parabolic curve
TINENTIUM EQUATUR YPOTHENUSE PIRAMIDIS, RELIQUUM CIRCUMFER- and draw a perpendicular to the parabola’s axis (line de in Fig. 4.8), and
ENTIE BASIS." we further draw a line tangent to the parabola at that point (tangent dg)
106 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 107

letting that tangent be extended in the direction of the vertex until it author gives no real proof. He merely says that it follows from the pre­
meets in g the axis of the parabola extended, and we construct a line ceding conclusion and from the well-known principle of optics that the
perpendicular to dg at d so that it meets the axis in point h , then he / ed angle of incidence and the angle of reflection are necessarily equal. It is
= ed I eg. From Conclusion 4, latus rectum le d = ed / ea, or de2 = latus evident that this proposition embraces the essential aspects of Propositions
rectum ca. Therefore, he-eg = latus rectum ca, or by alternation, latus 2 -4 of Alhazen’s De speculis comburentibus and Propositions IX.42-
rectum / he = eg / ea, he being the axial distance between the inter­ IX.43 of Witelo’s Perspectiva (see above, Chap. 3, note 43), though to be
sections of the axis with lines de and dh, and eg being the axial distance sure it differs from them by the author’s novel treatment of latus rectum.
between the perpendicular ed and the point of juncture of the axis and The only other thing concerned with the burning mirror that interests
the tangent, and ea being the axial segment between the vertex and the the author of the Speculi almukefi compositio is the construction of the
perpendicular de. This is what the enunciation has proposed. instrument by which the paraboloidal mirror may be hollowed out. This is
Now in the seventh conclusion (again see Fig. 4.8) the author easily Conclusion 10 and differs considerably from Proposition 5 of Alhazen’s
proves that the axial intercept between the perpendicular de and the vertex De speculis comburentibus and Proposition IX.44 of Witelo’s Perspectiva
a is equal to the axial intercept between the same vertex and the point (see above, Chap. 3, note 43). In fact, none of these accounts has great
of juncture of the axis extended and the tangent dg, which was the second interest for our history of the mathematical knowledge of conic sections
Apollonian proposition ( Conics, 1.35) whose proof the author desired. in the Middle Ages, though it ought to be said that the author of the
He proves this proposition by proving its converse, namely, that if these Speculi almukefi compositio describes much more completely than the
two axial segments are equal, line dg will be tangent to the section in authors of the other accounts the generation of the parabolic cutting edge
point d. The converse is shown by first assuming that if dg is not tangent of his hollowing-out instrument so that the reader has a very clear idea of
at d it will cut the parabola between d and a or between d and the base. how the mathematical properties of the parabola are transformed into a
But in either case a contradiction ensues (by using the property of the parab­ proper cutting edge.9 There is, however, an error in our author’s pro­
ola established in the second part of Conclusion 4), namely that a given cedure. Essentially his method is to form a segment of a right-angled cone
line is simultaneously equal and unequal to the same line. Since this is cut down so that the upper surface is a parabola. The error consists in
not possible, line dg must be tangent at d. And so, if the converse is true, his statement that the radius of the base of the segment (which base
the proposition must be true, that is, lines ae and ag are equal. With circle does not contain the intersection of the axes of the section and the
this proof completed, the author has now proved both of the Apollonian cone as its center) is double the focal distance from the vertex of the sec­
propositions (Proposition 1.11 in Conclusion 4 and Proposition 1.35 in tion. In the example he gives the radius of the base circle of the
Conclusion 7). It will be immediately evident to the reader of the pre­ segment is 40 feet. Now, as Regiomantanus later shows, the focal distance
ceding chapter that these two conclusions are equivalent to Witelo’s Propo­ for such a base circle is, in fact, a little more than 27 feet (see below,
sitions IX.39 and IX.40, where, however, the propositions are given but Chap. 5, Text C, bracketed section at the end of Conclusion 10). Fusoris,
not proved (see above, Chap. 3, notes 38 and 41). in summarizing the procedure of our author, makes a somewhat similar
Passing on to Conclusion 8 of the Speculi almukefi compositio, we error in suggesting that the diameter of the base of the cone (which base
first note that it is essentially the same as Proposition 1 of Alhazen’s does contain the intersection of the axes of the section and the cone as its
De speculis comburentibus (see Chap. 1, note 33), except that, like Witelo’s center) is four times the focal distance (ibid., text A, cap. 3, and see my
Proposition IX.41 (see above, Chap. 3, note 42), our author has given the discussion of Fusori’s error in the next chapter). Completely novel for
proofs of the three cases by synthesis alone. The proofs demand little treatises on burning mirrors is the section on the conditions of a good steel
comment on my part. The equality of the angles of incidence and reflection that is appended to Conclusion 10 in even the earliest manuscript of the
are proved for all possible cases, that is, when the reflected ray meets the tract.10 It is thus clearly an appendix prepared by the author of the tract
axis in an acute angle, a right angle, or an obtuse angle. himself. This appendix ought to be of considerable interest to the student
From this conclusion, our author immediately jumps to the general
conclusion (his Conclusion 9) that all solar rays parallel to the axis that fall 9 Needless to say, the Latin readers would have been very much better off in the matter
on a mirror hollowed out with parabolic concavity are reflected to one of drawing a parabola had the tract on the protraction of ail such sections which Alhazen
and the same point, that is to the point that is both the midpoint [of the promised to write been available in Latin (see above, Chap. 3, note 43, Prop. 5 of Alhazen’s
De speculis comburentibus).
segment] of the axis [of the parabola whose rotation produces the surface 10 Concerning the passage on steel drawn by the author from Albertus Magnus’ Mineralium
of the mirror, the segment, that is, that extends from the vertex of the libri V (see below, Prop. 10, Appendix on steel, lines 69-79); consult Albert’s Opera omnia,
parabola to the axis of the cone from which the parabola is produced] ed. of A. Borgnet, Vol. 5 (Paris, 1890), p. 50 and Albertus Magnus, Book of Minerals,
and the midpoint of the latus rectum [as conceived by the author]. The transl. by D. Wyckof (Oxford, 1967), p. 133.
 SPECULI ALM UKEFI COMPOSITIO 109
108 ARCHIMEDES IN THE MIDDLE AGES
cause the hyperbola is designated in the text as sectio mukefi addita
of medieval technology, for it seems to display a first-hand knowledge of (see variant reading for line 85) rather than simply as sectio addita, which
iron and steel working practices among the smiths of the author’s time. latter is the correct designation our author had given in the third definition.
Having achieved his main purpose of writing a treatise on the parabolic It might be argued that an early scribe falsely introduced only the word
burning mirror, the author was apparently reluctant to leave the wider mukefi rather than the whole note, and this seems very likely. Unfortu­
subject of conic sections without adding a new proof he had devised for nately once again MS C cannot be read at this point. Despite the dif­
Proposition 5 of the De duabus lineis. This is his eleventh conclusion. ficulties connected with both of these statements, the fact that the author
In it he gives three cases of lines that continually approach more closely clearly understood the nature and construction of a hyperbola as evidenced
together without ever meeting. (1) The first is the case of a hyperbolic in the third definition persuades me that whether he or someone else added
curve and a straight line that are not in the same plane. (2) The second the two statements under consideration, the original author could scarcely
is the case found in Proposition 5 of the De duabus lineis, that is, the have failed to recognize that it was a hyperbola that he was employing
case of a hyperbolic curve and straight line in the same plane. (3) The third in this proof.
is the case of two hyperbolic curves in the same plane. The proofs of the first two cases enunciated in Conclusion 11 are similar.
We should realize initially that our author produces the hyperbola in a In the first case (see Fig. 4.13a) we take any two points g and h on the
manner that differs from but is equivalent to the Apollonian construction hyperbolic curve eghf and we pass through each point a plane parallel
found in the De duabus lineis. Starting with the usual right circular cone, to the base of the cone, thus producing two circular sections. These
he first notes an axial triangle abc (see Fig. 4 .13a). Then he cuts the cone circles cut side ac of axial triangle abc in points k and /. Then he proves
by a plane parallel to the plane of the axial triangle, which produces \hathl < gk. The proof is clear from Fig. 4.13b, afigure which I have added
hyperbola eghf. It will be seen immediately that this in fact is equivalent to the text. Since the right-sines ht and gs of the two unequal arcs hi
to the Apollonian procedure, for, if we suppose another axial triangle at andgk are equal, their versines are unequal, i.e. tl < sk. Thus it is obvious
right angles to triangle abc, then it will also be at right angles to the plane that hi < gk. This is true for any two points on the curve no matter how
of the section eghf and the latter plane if extended will somewhere be far the curve and the straight line are extended. Hence the lines con­
intersected by a side of the axial triangle to which the plane is perpen­ tinually approach each other more closely. They can never meet because
dicular. The point of intersection will be in that side of the axial triangle the lines lie in two distinct planes that are parallel to each other and thus
extended beyond the apex of the cone. Such an intersection of plane the lines must at least have between them the distance between the planes.
and side is the constructional condition for producing a hyperbola in (But in fact he has already proved that they can never come that close
Proposition 1.12 of Apollonius’ Conics, in Gerard’s fragment (see above, to each other.) This is a curious case because of the fact that the lines
Chap. 1, note 21) and in Proposition 5 of the De duabus lineis. It is are not in the same plane (though they are in the same surface of the
rather interesting that our author does not specify curved line eghf as hy­ cone) and our author recognizes that this might be an objection. So he
perbolic in the actual proof of his Conclusion 11, though there is little doubt sets out in the second case to prove the enunciation for the curve eghf
that he knew the curve to be hyperbolic. Two statements appear to con­ and straight line mn that are in the same plane.
firm this, though there are problems connected with both. At the end of First the author constructs a plane tangent to line ac (a generator of
the appendix to the tenth conclusion, the text has as follows: “ But though the cone and side of the axial triangle abc). Since the plane of axial tri­
we have brought our desired proposal to its desired end, because a certain angle abc is parallel to the plane of section eghf, the tangent plane will
conclusion that seems on the face of it marvelous follows from the [con­ cut the sectional plane in line mn parallel to line ac. Then once more we
struction of a] hyperbola, therefore we have thought it worthy to add the take any two points o and p on the hyperbolic curve eghf and through
said conclusion here.’’ The difficulty connected with this statement is that, these points circles parallel to the base are constructed as before. These
of the manuscripts that can be read, namely OFRV, only R designates
circles cut line ac at points q and r and their planes cut line mn at points
the section in question as a sectio addita (or as ed calls it, a hyperbole v and w. Therefore, as before, the versines are unequal, that is, yr < xq
[/])· The other manuscripts have some form of sectio mukefi (see the (see my added Fig. 4.13c). But since ov = xq andpw = yr, it is clear that
variant reading for the appendix to Conclusion 10, line 107). Unfortu­ pw < ov. We can take the proof one step further than the author by saying
nately the oldest manuscript C cannot be read at this point. The second
that, since pw < ov, the perpendicular distances to line mn from points
statement occurs after the completion of the proof of Conclusion 11:
o and p are similarly related, that is, pz,2 < oz\ (see added Fig. 4.13d).
“ Note that for the proof of this conclusion a hyperbola has been as­
Hence it is obvious that eghf and mn continually approach each other
sumed.” Nothing could be plainer. However, there is perhaps some doubt more closely. That they will never meet is obvious. For, since the plane
that this statement was written by the author of the original treatise be­
110 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM U K EFI COMPOSITIO 111

tangent to line ac will touch the surface of the cone only in line ac and since likely to have been in the original text (and in C if it could be read);
mn is in that tangent plane and is parallel to line ac, it is evident that line and cases where they agree with C against OF. Hence I have had to pay
mn will never touch the surface of the cone and thus will never touch close attention to all the copies in those places where OF seems clearly
the curve eghf that is in the surface of the cone. to be wrong. However, I have not included all the variant readings of
The third case is also easily proved. We take another cone similar and V, many of which are completely absurd or trivial (e.g., conclusio often
equal to the first one and let it lie so that ac is an element common to the becomes coneavatio, proportio becomes propositio, corelarium becomes
surface of each cone. Then we extend the plane that was parallel to axial correspondens, igitur is almost always changed to ergo, etc.). Though I
triangle abc and produced section eghf in the first cone so that it cuts the have given almost all of the readings from MS O and F, the modernized
second cone in the same manner and thus produces in the surface of the humanistic spellings given in F have generally only been noted the first
second cone a hyperbolic curve precisely like that of eghf. Then, ac­ time they occur, for it is clear that this scribe’s orthography is not that of
cording to the second case above, each of these hyperbolic curves will the original text. While MSS O and F are very close to each other, it is
continually approach closer to line mn without ever meeting it. Hence, demonstrable that F (which I believe to be the later of the two manu­
since the curves will never meet the intermediary line mn, they will never scripts) is not simply a copy of O since on a number of occasions it con­
meet each other though always approaching each other more closely. tains lines that are omitted from O but are necessary for the text and
And so the author’s tract is completed. Then, after making the above- are included in C and V.
discussed remark that the proof of this conclusion assumes a hyperbola, Further word about the orthography of my edition is in order. Apol­
he merely adds that there are more marvelous conclusions which can be lonius’ name is spelled in a great variety of ways, as will be evident
deduced and reduced “ to the conclusions which the great Apollonius to the reader who consults the variant readings that accompany the text
inserted in his book On Conic Figures.” But not having seen Apollonius’ below. But since the proper spelling Apollonius is almost always found
work despite searching for it with great diligence, he will be satisfied with in MS C, I have everywhere adopted that spelling (my only alteration
the conclusions he has presented. being always to capitalize the initial letter). Similarly, I have adopted the
Now I must say a few words about the text that I have edited below. spelling Vitello from MS V (it is not clear in MS C, but I believe it is
There are six copies, five manuscripts and the edition of Gogava that I also Vitello there) instead of Vitulo as found in MSS OF. I have written
have already mentioned. The sigla I have adopted are listed below imme­ piramis instead of pyramis since it appears always to be so written in MS C
diately preceding the text. The six copies represent two versions. The and most of the time it is this way in MSS OFR. I have also adopted the
first version, which must be close to the original text, is found in MSS spelling equidistans instead of equedistans, though both spellings are found
COF-, the second was apparently prepared by Regiomontanus and is repre­ in the manuscripts and equedistans is more prevalent in MS C. I reject
sented by MS R (in Regiomontanus’ hand) and Gogava’s edition (ed). sicud and capud in favor of sicut and caput even though the former
MS V generally follows the earlier version of COF but on occasion shows spellings are found in MS C, primarily because the latter are compatible
influence of the second version. The text I have edited below in this chapter to normal spelling procedures. Similarly I always write -ti- before vowels
is the first version. It is accompanied by an English translation. The second instead of -ci-, though the latter is often found in most of the manu­
version is given in the next chapter without translation. I hasten to add scripts. (MS V, however, almost always h as-ti-.) I have adopted reflectio
that the two versions do not significantly diverge as to basic content (ex­ instead of reflexio because the former spelling is almost always given by
cept that the second version is shorter and tends to express the mathe­ MS C. I have written orthogonius instead of ortogonius, since the former
matics of the proofs in neater and more economic terms and its figures is exclusively used in MS C. Similarly I have adopted immo from C
are less crude than those in the first version). The principal shortening instead of imo, ymmo or ymo. I give ypotesis, for it is written so in all
found in the second version involves the long proemium included in the manuscripts except MS F, which normally writes ipotesis. I have already
first version. Furthermore, the second version completely omits the long commented on the fact that rnukefi is often miswritten as mukesi (and
appendix on steel that is added to Conclusion 10. For my text of the original occasionally muchefi or muksi) but everywhere I write the correct rnukefi.
version I have, where possible, relied on MS C, the oldest manuscript. I have also pointed out that the original author apparently wrote muniani
However that manuscript was badly burned and so large parts of the text instead of Gerard of Cremona’s munani and so I felt obliged to adopt the
in it are unreadable. Thus I have had to depend heavily on the text as former reading.
given in MSS O and F, which on most occasions seems to be closer to I must also remark on a few decisions I have made concerning readings
the text in C than the texts in MSS R and V. However, there are cases to be removed or added to the text.
where R or V supply a reading that is not given in OF but that was most (1) In one case I have removed from the text the second of two alternate
112 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 113

readings that seem to have come into the text from an alternate tradition are rarely included in MS C (and are not complete when they are given).
(see the Proemium, variant reading for line 12). They were omitted from MS F.
(2) I have rejected MS R's addition to the second definition (see the The marginal references to folio numbers that appear in my text are to
variant reading to line 10 of Descriptio prima etc.), for by it Regiomon­ MS O.
tanus seeks to limit a linea muniani to a conic curve, while Apollonius The Sigla
(and through him Gerard’s translation of the fragment from Apollonius)
would allow it to embrace all “ curved lines” (the original Greek term C = London, British Library, MS Cotton Tiberius B.IX, 231r-235r, 14c.
being καμπύλης, which would include the circumferences of circles paral­ This manuscript was badly burned and the sheets in this section have
lel to the base of the cone as well as the conic curves). been bound out of order. The proper order of the folios including this
(3) In the fourth conclusion, variant reading to line 2 , 1 have suggested tract is the following: 233, 234, 231, 232, 235, and it begins on 233v.
that perhaps the author originally had mukefi instead of parabola, since the It bears the title De speculis comburentibus on both 233v and 234r, per­
former is the expression he ordinarily uses throughout the text and, as haps both times in a later hand. For the connection of this part of the manu­
I have already stated, he almost certainly did not include the statement script with Simon Bredon and John Dee, see A. G. Watson, “ A Merton
from Witelo identifying the Arabic names for conic sections with the College Manuscript Reconstructed: Harley 625, Digby 178, fols. 1 -1 4 ,8 8 -
Greek terms (cf. Descriptio prima e tc ., variant reading to line 37). How­ 115; Cotton Tiberius B.IX, fols. 1-4, 225-35,” Bodleian Library Record,
ever, I must point out that the term parabola at this place of the text Vol. 9 (1976), pp. 207-16. Note that the text at the end of the tract is
appears in even the earliest manuscript C, and so, if it is an intrusion badly burned and torn on 235r, but I can read all but the last two lines and
into the text, it must have been a very early intrusion. hence I am quite sure that it does not go over to folio 235v. Note finally
(4) From the tenth conclusion, variant to line 62, it will be clear to the that the manuscript is dated “about the 14c” in A Catalogue o f the Manu­
reader that I have followed Red in adopting the positions of m and o scripts o f the Cottonian Library Deposited in The British Museum, Vol. 1
given in Fig. 4.12. It may well be, however, that the original text had (London, 1802), p. 37.
these two letters transposed, since they seem to be in the transposed O = Oxford, Bodleian Library, MS Canon. Misc. 480, 47r-54r, 15c.
positions almost everywhere in MSS COFV. But if I should now follow F = Florence, Bibl. Med. Laur., MS Ashburnham 957, 95r-110v, 15c.
the transposed positions in my text, certain other inconsistencies would V = Verona, Bibl. Capitolare, MS 206, lr-8v, 16c. For the first notice
appear in the text and diagram of the earlier tradition, and so for clarity’s of this copy, see J. L. Heiberg and Eilhard Wiedemann, “ Eine arabische
sake I decided to adopt the reading of Regiomontanus’ text. Schrift fiber die Parabel und parabolische Hohlspiegel,’'Bibliotheca Mathe­
A word should be said concerning the numbers of the conclusions. I have matica, 3. Folge, Vol. 11 (1910-1911), pp. 193-208. In this article the
given the numbers as they appear in Red and F (except for the last proposition, manuscript is partially collated for the Proem, the Definitions and Con­
which is unnumbered in F). I assume that these were also the numbers clusions 1 and 4 with Gogava’s ed (see below). The implication that the
understood by the scribe of manuscript C, for, though he gives only one tract depends on an Arabic original must be rejected.
number that I can read, that number is the correct one for Conclusion 9 R = Vienna, Nationalbibl., MS 5258, 27r-38v, 15c. This is in Regiomon­
and thus is past the points where the numbers of MSS O and V begin tanus’ hand. See E. Zinner, Leben und Wirken des Joh. Muller von Konigs-
to interrupt the normal order of conclusion numbers. The scribe of MS O, berg genannt Regiomontanus, 2nd ed. (Osnabrfick, 1968), p. 308. It is the
after numbering the first seven conclusions correctly, skips to the number source of Gogava’s edition. I have edited this version in Chapter 5 below,
” 10a” for the eighth conclusion and his numbers are askew thereafter, Text C.
except that, as in MSS F and V, he fails to number the last conclusion. ed = A. H. Gogava, editor, Cl.Ptolemaei . . . operis quadripartiti in
Incidentally the fact that all copies except Red omit numbering the last Latinum sermonem traductio. . . . Item, De sectione conica orthogona,
conclusion is no doubt because this extra proposition is not a part of the quae parabola dicitur: Deque speculo ustorio, libelli duo, hactenus de­
subject originally proposed by the author for his tract. Finally, we should siderati (Louvain, 1548), sig. P 4r-S 2r, i.e. 60v-70r. See the article of
note the peculiar numbering in MS V. The scribe omits the number of the Heiberg and Wiedemann cited in the description of MS V above and fur­
first conclusion, correctly gives the numbers for the second through the ther remarks in Chapter 5 below.
fourth, omits the number of the fifth, erroneously numbers the sixth through [There was apparently a copy of the Speculi almukefi compositio in
the ninth as the “ second,” “ third,” “ fourth,” and “ fifth,” and omits the library of the Sorbonne that was seen and used by Jean Fusoris
numbering the tenth and eleventh. in composing Chapter 3 of his Libellus de seccione mukefi, but I have not
For reproducing the diagrams I depend mainly on OF and Red. Diagrams located this copy. See the next chapter below— Text A.]
 SPECULI ALM UKEFI COMPOSITIO 115

The Composition of the Parabolic Mirror tionem frequentem dictarum conclusionum probationem ab aliquo didicis-
15 sem ad que (/) sive ad eorum (/) alterum (/) numquam potui ullo conamine
pervenire, tandem igitur multum motus ut dictarum probationem insisterem
47r / Speculi almukefi compositio primo ad earum alteram me converti. Et cum eius veritatem probandam
aliquamdiu studuissem occurrunt mihi probationes per quas dictam con­
[Proemium] clusionem saltem iuxta intellectum quem de ea conceperem demonstravi
20 notorie esse falsam et istam igitur tanquam falsam per longa tempora sprevi
Quia universorum quos de speculis ad datam distantiam comburentibus penitus et abieci. Verum postea diffuse considerans verisimile non fuisse
tractare perpendi, seu quorum vidi tractatus in scriptis, omnes et singuli quod dictus Apollonius et alii eum allegantes in falsitate dicte conclu­
duas supposuerunt conclusiones ab Apollonio Pergeo allegatas super qui- sionis tam enormiter deviassent— presertim cum dictum Apollonium al­
5 bus tota eorum intentio fundabatur, in tantum quod in perspectiva Vitel- leget Campanus in expositione 14' elementorum Euclidis, conclusione
lonis ipsemet auctor librum premittit in quo demonstrare nititur omnes 25 prima, quem ibidem Apollonium magnum vocat— supposui pro constanti
conclusiones que sibi videbantur necessarie atque preambule ad suum alium fuisse intellectum conclusionis illius, circa quem intellectum cum
propositum ostendendum ut totus suus liber sufficeret sibi ipsi, dictus studio diuturno; insistens finaliter cum magno tedio et labore ad unum
tamen Vitello duas Apollonii conclusiones in perspectiva sua supposuit verum eius intellectum deveni, cuius probatio dei sapientia influente michi
10 quas non probat sed eis utitur tanquam notis, licet veritatis earum notitia satis faciliter occurrebat. Ista igitur conclusione probata et probatione
me latuit valde diu, non obstante quod erga fratres de partibus longinquis 30 eius in scriptis redacta, circa probationem alterius conclusionis Apollonii
et alios diligentiam valde exquisitam exhibui per tempora diuturna ut vel donec eam deo instruente inveneram insistere non tardavi. Probationibus
eorum mediatione ad dictum librum Apollonii devenissem vel per opposi- igitur dictarum duarum conclusionum scitis et scriptis ne tam sollicitus
labor meus inutiliter perderetur sive omnino perdatur propono mediantibus
Tit. Speculi . . . compositio OF De speculis comburentibus CR Tractatus de
conclusionibus istis et aliis infra probandis docere facere unum speculum
speculis comburentibus V De sectione conica orthogona quae parabola dicitur
ed, et in paginis textus scr. ed De sectione parabola / almukefi (? abnukefi ?) O 35 combustivum quod ad quantamlibet comburet distantiam, que potest esse
almukesi F (cf. meam intro.) / post compositio add. OF secundum Rugerium dimidium semidiametri circuli cuiuscunque cuius aliqua portio quantum-
(O, Roggerium F) Bacon (F, bacon O) ordinis minorum sed delevi quia non cumque magna seu modica poterit instrumentaliter lineari. Et hec erit huius
est in C; cf. meam intro. tractatus intentio principalis. Ad quam intentionem perficiendam pariter
Proemium et probandam quasdam descriptiones et quasdam conclusiones premittam.
1 [Proemium] addidi
2-41 Cf. textum abbrev. in Red in cap. 5 inferius
2 universorum O universus V diversorum F / post speculis add. V concavis /
datam OF certam V 15 post conamine add. F modo aliquo
3 in scriptis tr. V post quorum 19 ea O eo F
4 duas . . . conclusiones OF composuerunt duas concavationes V / Apollonio 20 istam CF illam Ο I igitur C om. OF
Pergeo CV apolonio pereo O apollonio perseo F Apolonio pergeo R Apollonio 22 Apollonius C appolonius O appollonius F
Pergaeo ed 23 Apollonium C appolonium O appollonium F
5 -6 quod . . . auctor OF ut etiam ipse Vitello (?) auctor perspective unum V / 25 Apollonium CF appolonium O
perspectiva O prospectiva F / Vitellonis correxi ex Vitulonis in OF et Vitello 21 insistens C consistens OF
(?) in V 28 eius O eiusdem F et tr. F post intellectum / michi C m O mihi F
7 conclusiones OF concavationes V I videbantur OF videntur V 29-30 Ista . . . eius OF unius ergo probatione inventa et V
8 ostendendum OF ita V / suus liber OF tr. V 30 conclusionis Apollonii om. V / Apollonii C appolonii O appollonii F
9 Vitello V, (?)C Vitulo OF / conclusiones OF concavationes V et tr. V ante 31 eam . . . inveneram COF id monente invenirem V
Apollonii / Apollonii FV appolonii O / supposuit O disposuit F supponit V 32 igitur . . . conclusionum C igitur dictarum conclusionum OF itaque istarum
10 eis utitur OF tr. V / notis OF necessariis V / de notis scr. mg. O in fine 9‘ libri duarum questionum V / sollicitus C solicitus V subtiliter OF
et in fine 3“ (?) / licet OF et licet V 33 inutiliter COF om. V
11 me . . . diu OF latuit valde V 34 conclusionibus COF om. V I probandis COF prout docetur V / unum om. (?)C
11-28 erga. . . . sapientia OF adhiberem diligentiam magnam pro demonstratione 35 combustivum quod COF comburens V / ad quantamlibet CO a quantalibet F ad
obtinenda earundem tamen finaliter et tedio et labore ad unum verum quantam libuerit V / comburet COF om.V / que COF quod V
intellectum alterius earum deveni cuius probatio et quia V 36 semidiametri circuli COF diametri speculi V / aliqua COF alicuius V
12 post exhibui scr. OF aliter adhibui quod delevi; cf. meam intro. 37 hec COF om.V
13 Apollonii corr. ex apolonii in O et appolonii in F 38 principalis COF propinata V
39 quasdam' . . . conclusiones COF quasdam concavationes et descriptiones V
114
116 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UK EFI COMPOSITIO 117

40 Prima igitur est descriptio piramidis rotunde orthogonie seu erecte secun­ perficierum planarum secantium stet super alteram secundum angulos
dum quod eam ponit Euclides in undecimo elementorum suorum, et est ista. 20 rectos, tunc linea recta que est differentia communis est dictis duabus
superficiebus secantibus necessario vel secabit unum duorum laterum
Descriptio prima
trianguli et equidistabit alteri et tunc vocatur sectio mukefi vel secabit
[1] Piramis rotunda est transitus trianguli rectanguli altero suorum late­ unum duorum laterum trianguli et concurret cum altero in infinitum pro­
rum rectum angulum continentium fixo et triangulo ipso donec ad locum tracto, quod contingere potest dupliciter vel ex parte coni et tunc sectio
1 unde moveri cepit redeat circumducto. Si autem latus fixum / lateri cir- 25 quam continet linea muniani nominabitur sectio addita vel secabit ex
5 cumducto fuerit equale, erit piramis rectangula; si vero longius, acuti- parte basis intra vel extra et tunc— si non sit circulus— vocatur sectio
angula; et si brevius, erit obtusiangula [Fig. 4.1]. Axis autem ipsius pira­ diminuta [Fig. 4.3]. Item dicta linea que est differentia communis dictis
midis est latus fixum basisque sua circulus. superficiebus quarum altera est sectio mukefi vocatur axis seu sagitta
sectionis mukefi [Fig. 4.4], cuius extremum versus conum piramidis voca-
Descriptio 2a 30 tur caput sectionis mukefi et caput linee muniani. Et omnes linee protracte
orthogonaliter super dictam sagittam ab uno latere linee muniani usque
[2] In omni rotunda piramide orthogonia secta per superficiem planam
ad alterum eius latus erunt equidistantes et dividuntur per dictam sagittam
io non transeuntem per caput piramidis differentia communis dicte super­
in duo equalia et vocantur singule linee sic protracte linee ordinis illi
ficiei secanti et dicte piramidi vocatur sectio quam continet linea muniani
sagitte. Omnes isti termini ab Apollonio summuntur (/). Et est notandum
[Fig. 4.2]. Hec secunda descriptio ab Apollonio allegatur.
35 quod, licet hic inter sectiones non transeuntes per caput piramidis triplex
sectio describatur, non tamen indigebimus ad principale nostrum proposi­
Prima presuppositio [= Descriptio 3a]
tum nisi sectione mukefi, ut patebit.
[1] Quando secatur piramis rotunda orthogonia et rectangula cum dua-
15 bus superficiebus planis quarum una per caput et per centrum basis transit 2a presuppositio [= Descriptio 4a]
secans piramidem secundum triangulum, ut infra probabitur, et altera per
[2] Licet in sectione mukefi quelibet linea ordinis possit vocari basis
caput eius non transit immo secat eam cum superficie quam continet
40 recta seu latus rectum sectionis mukefi intercepte inter dictam lineam
linea muniani iuxta descriptionem premissam, sic quod una dictarum su-
ordinis et caput sectionis, non tamen ab Apollonio sic vocatur. Immo hec
40 igitur CO itaque F
40-41 Prima . . . ista COF quarum prima est V
19 stet CFRV stat O
40 piramidis CO pyramidis F hic et paene ubique (cf. meam intro.) / orthogonie
20 est tr. OF post communis
C ortogonie OF
22 vocabitur F I de sectio mukefi scr. mg.O quid est sectio mukefi / mukefi: parabola
41 eam ponit C tr OF / undecimo CO secundo F
ed hic et paene ubique
Descriptio prima—2a presuppositio 23 duorum om. C / altero Red, (?)C alio OF, et add. Red latere
1 Descriptio prima OF om. C 25 muniani Red, (?)C miniani OFV / de sectio addita scr. mg. O quid est sectio
2 de descriptione prima scr. mg. O diffinito pyramidis et in manu parviori hanc ad[d]ita i post addita add. ed sive hyperbola
ponit (?) euclides in principio 11' elementorum / rectanguli om. V 26-27 de sectio diminuta scr. mg. O quid est sectio diminuta / post sectio
3 post fixo scr. injuste O et triangulum continendum fixo diminuta add. ed sive Ellipsis
4 moveri OFRVed movere C / autem COF tunc RV 27 differentia communis CR differentia correspondens V communis sectio OF
5 longius CRV longior OF 28 sagitta CO sagypta F hic et paene ubique
6 brevius CRV brevior OF 30,33 pertracte O
Axis . . . circulus COF om. RV, sed hah. RV Hoc satis patet in principio 30,31 muniani CRed miniani OFV
undecimi elementorum euclidis 34 Apollonio CVed apolonio OR apollonio F (variae lectiones huius nominis
8 Descriptio 2a OF secunda V 2a descriptio mg. O postea non datae erunt)
9 orthogonia CFR ortogonia O Orthogona ed 37 post patebit add. OF V (sed hic var. V non do) Et ulterius (et etiam in F)
10 post piramidis add. RV nec equidistantem basi (sed non in COF et cf. meam est notandum quod sectio rectangula, parabola et mukefi idem sunt. Item sectio
intro.) ambligonia, yperbole (/), mukefi addita idem sunt. Et consimiliter sectio
11 muniani CR om. V minuali OF Arabice dicta muniani ed (cf. meam oxigonia (exigonia in F), ellipsis et mukefi diminuta sunt idem secundum
intro.) auctores (et mg. add. O sectiones et in mg. inferius repet. O sectio
12 secunda COF om. RVed / Apollonio CFVed apolonio OR parabola et rectangula et mukefi sunt idem; sectio ambligonia, yperbola et
13 Prima presuppositio O om. CRed prima suppositio mg. O mg. F tertia V mukefi addita sunt idem; sectio oxigona [/], eclipsis [/] et mukefi diminuta
15 basis CRVed om. OF sunt idem); sed istud totum delendum est quia non est in C et cf. meam intro.
16 secans. . . . probabitur del. C 38 2a presuppositio O om. CRVed 2a suppositio mg. O mg. F
18 linea muniani CRed linea minuani OF lineam minuani V 39 posset C
118 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 119

est descriptio eius vera sectionis mukefi cuiuscunque descendente sagitta Secunda Conclusio
a capite sectionis quousque ex parte basis cum axe piramidis concurrerit
[2] OMNIUM DUORUM PUNCTORUM IN SUPERFICIE ROTUNDE
linea ordinis transiens per punctum medium dicte sagitte vocatur latus
PIRAMIDIS EXISTENTIUM LINEA RECTA INTER DUO PUNCTA
45 rectum seu latus erectum illius sectionis mukefi quantumcunque illa sectio
PROTRACTA, SI IN INFINITUM PROTENSA NON TANGAT
ex parte basis longa fuerit sive curta. Unde patet ex hoc de omni sec­
5 CONUM PIRAMIDIS, TUNC DICTA LINEA VADET NECESSARIO
tione mukefi quod illa sectio et quelibet eius pars intercepta inter aliquam
INTRA PIRAMIDEM.
lineam ordinis et caput sectionis sit inter partes linee muniani [ac] singule
Probatio. Sint in piramide abc dicta duo puncta d et e inter que pro­
tales partes sunt sectiones mukefi et omnes tales idem obtinent latus rec-
tendetur linea recta de que ulterius quantumcunque protensa non transiret
50 tum seu latus erectum. Et erit illud latus rectum sive erectum necessario
per conum piramidis [Fig. 4.5]. Dico igitur de ingredi piramidem.
duplum ad sagittam sectionis mukefi protensam a capite illius sectionis
io Protraham a cono a per puncta d et e duas lineas rectas usque ad
48r quousque / concurrat cum axe piramidis, ut in quarta conclusione proba­
circumferentiam basis, que sint adb et aec, et coniungam c, b per lineam
bitur deo dante. Ad istam descriptionem recti lateris sectionis mukefi
rectam, que necessario secabit circulum qui est basis, per secundam 311
non potui unquam devenire per aliquod quod inveni ab Apollonio allegatum
Euclidis. Triangulus igitur abc secat piramidem. Sed in illo triangulo est
55 sed solummodo per hoc quod de omni alia linea ordinis ostendi cum
linea de\ igitur linea de est intra piramidem, cum totus ille triangulus
magno studio et labore conclusiones Apollonii esse falsas et de ista tantum­
15 sit intra piramidem. Hoc etiam patet, signato puncto / in linea de et
modo esse veras. Istis igitur quatuor descriptionibus prenotatis sequentur
protracta linea a f usque ad cb, que ei occurrat in puncto g, hec enim tota
conclusiones, quarum prima erit ista.
est intra piramidem, quare et linea de est intra piramidem.
Prima conclusio Tertia conclusio
[1] IN OMNI ROTUNDA PIRAMIDE ORTHOGONIA SECTA IN [3] IN OMNI ROTUNDA PIRAMIDE ORTHOGONIA SECTA CUM
DUO EQUALIA CUM SUPERFICIE PLANA TRANSEUNTE PER SUPERFICIE PLANA EQUIDISTANTI BASI SECTIO COMMUNIS
CAPUT PIRAMIDIS ET PER CENTRUM SUE BASIS DIFFERENTIA CURVE SUPERFICIEI PIRAMIDIS ET SUPERFICIEI SECANTI
5 COMMUNIS DICTE PIRAMIDI ET DICTE SUPERFICIEI SECANTI 5 ERIT LINEA CONTINENS CIRCULUM CUIUS CENTRUM ERIT
ERIT TRIANGULUS RECTILINEARIS CUIUS BASIS ERIT DYA- IN AXE PIRAMIDIS.
METER BASIS PIRAMIDIS ITA SECTE. Probatur. Sit piramis rotunda abc [Fig. 4.6], cuius centrum basis sit
Ista conclusio statim se offert intellectui ex descriptione piramidis per d , et axis piramidis ad\ seceturque cum superficie equidistanti basi que
triangulum circumductum, ut in prima descriptione liquebat; et sic patet faciet in superficie piramidis lineam curvam per premissam, que sit linea
io veritas illius quam descriptio tertia supposuit. Et vocantur latera dicti io ef. Dico igitur lineam curvam e f esse circumferentiam circuli et habere
trianguli ypotenuse piramidis, ut patet per conclusionem primam de curvis centrum super axem piramidis.
superficiebus. Signentur enim in linea ef duo puncta e et f ubicumque volueris, per
que ab a cono piramidis ad circumferentiam basis protrahantur linee aeb
42 de descriptio . . . mukefi scr. mg. O descriptio mukefi et afc, et protrahatur axis ad, que ubi transierit per superficiem planam
46 curta CR,(?)V mg. O curva OF
47 aliquam CRVed quamlibet OF Secunda conclusio
48 partes CRVed partem OF / [ac] addidi 1 Secunda conclusio O mg. F om. C 2a mg. R Secunda V PROPOSITIO
49 obtinent OFRed optinent C retinent V SECUNDA ed
50-53 Et . . . dante om. C 5 vadet CORed vadit V ibit F
51 sectionis:2 tensionis (?) vel sectionis O 6 intra: inter F
53 sectionis om. OF 7 Sint Red sicut OF simul V
53-57 Ad . . . veras om. V 10 Protraham FRed, (?)V pertraham O
54 ab: ad O 13 Sed RVed s.e.d. OF
55 per: pro F 14,15,171·2 intra RVed infra OF
57 quattuor F / sequentur C sequuntur O secuntur FR
Tertia conclusio
Prima conclusio 1 Tertia conclusio OV mg. F, om. (?) C 3a mg. R PROPOSITIO TERTIA ed
1 Prima conclusio O mg. F om. CV l a mg. R PROPOSITIO PRIMA ed 6 axe RVed maxime OF
2 [1] addidi et etiam omnes alios numeros conclusionum / ortonia F 8 post axis injuste hab. OF basis
7 ita secte: intersecte F, (?)0 10 igitur lineam O ergo linea F
11 ypotenuse CR ypotemusis O ypotenusis F ypotemisi V hypotenusae ed 13-14 protrahantur . . . afc om. F
120 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 121

15 secantem signetur punctus g, et protrahantur linee ge et g f et db et dc. dictum triangulum, ymaginor circulum transire per puncta g et h divi­
Erit igitur per secundam 6U Euclidis proportio bd ad eg atque cd ad fg dentem piramidem equidistantem basi piramidis, cuius diameter sit khl, erit
sicut linea da ad ga. Sed db et dc sunt equales quia semidiameteri basis. igitur quadratum linee hg equale quadrangulo contento ex kh et hi, per 30m
Igitur ge et g f erunt equales. Et consimiliter probabitur de quibuslibet 25 tertii et 8am sexti Euclidis. Cum igitur quadratum oc linee, que est
lineis aliis protractis a puncto g ad lineam curvam ef. Igitur per nonam semidiameter basis piramidis erecta orthogonaliter super diametrum bd,
20 tertii Euclidis linea ef est linea continens circulum et centrum eius est resultet eadem ratione ex ductu bo in o d , sed linee hl et od sunt equales cum
punctus g, quod fuit probandum. sint equidistantes et inter ho et Id equidistantes per 33am primi Euclidis,
igitur quadratum oc resultat ex ductu hi in ob. Et quia quadratum hg
30 resultat ex ductu eiusdem hl in hk, sed productorum et producentium
Quarta conclusio eadem est proportio, ut elici potest a 16a et 17a T Euclidis in numeris et ut
elici potest a prima Θ Euclidis in quantitatibus, eadem igitur erit proportio
[4] IN OMNI SECTIONE PARABOLA LATUS ERECTUM DUPLUM
bo ad kh sicut quadrati oc ad quadratum hg. Sed eadem est proportio bo ad
EST NECESSARIO AD SAGITTAM, ISTUDQUE LATUS ERECTUM
kh sicut f o adfh per 2am61Euclidis. Igitur eadem est proportio/o adfh sicut
AD OMNEM PERPENDICULAREM EXEUNTEM A SAGITTA
35 quadrati oc ad quadratum hg. Et consimiliter probari potest de omnibus
5 USQUE AD CURVUM LATUS SECTIONIS SE HABEBIT SICUT
perpendicularibus exeuntibus a duobus punctis sagitte usque ad latus
ILLA PERPENDICULARIS SE HABET AD PARTEM SAGITTE IN-
curvum figure mukefi quod linea sagitte intercepta inter perpendicularem
48v TERCEPTAM INTER DICTAM PERPENDI/CULAREM ET CAPUT
remotiorem et verticem figure mukefi se habet ad lineam sagitte inter­
PROPOSITE SECTIONIS.
ceptam inter perpendicularem propinquiorem et verticem sicut quadratum
Verbi gratia, in piramide abd [Fig. 4.7] secta sectione mukefi que
40 perpendicularis remotioris ad quadratum perpendicularis propinquioris.
io sit pfc, cuius sagitta fo concurrat in puncto o cum axe piramidis ao et
Cum igitur ex premissis o f ad h f sicut quadratum oc ad quadratum hg, sed
per h punctum medium sagitte transeat linea ordinis phg orthogonaliter
o f erit dupla ad h f cum ex ypotesi h sit medius punctus o f, igitur quadratum
super sagittam, que quidem linea ordinis vocatur latus erectum dicte sec­
linee oc erit duplum ad quadratum linee hg. Igitur per corelarium 17e 61
tionis per quartam descriptionem. Tunc dico quod linea pg est dupla ad
Euclidis linea oc se habet ad lineam hg in medietate duple proportionis,
sagittam fo et quod idem pg se habet ad perpendicularem mn exeuntem
45 linee autem oc et o d sunt equales cum sint semidiametri eiusdem circuli,
15 a sagitta ad latus curvum sectionis sicut illa perpendicularis mn se habet
scilicet basis piramidis; igitur od se habet ad hg in medietate duple
ad m f partem sagitte interceptam inter dictam perpendicularem et caput
proportionis. Sed et ad se habet ad od in medietate duple proportionis per
sectionis.
8am61et per corelarium 17e 6‘. Igitur ad est dupla ad hg. Sed et eadem ad est
Probatur. Secet triangulus abd propositam piramidem in duo media
dupla ad o f per secundam sexti, eo quod db sit dupla ad o b . Igitur eadem est
transiens per eius caput et per eius centrum sic quod ille triangulus stet
50 proportio ad ad o f et ad hg. Igitur hg et o f sunt equales per 7am 5UEuclidis.
20 super sectionem mukefi secundum angulos rectos, prout vult 3a descriptio.
Sed pg est dupla ad hg. Igitur pg est dupla ad fo . Patet igitur prima pars
Cum igitur in proposito necessario stabit linea gh perpendiculariter super
22 ymaginaliter F
15 punctus g C tr. OV igitur punctus F 24-25 SO1" tertii OR 33 F tertiam tertii V trigesimam 3 ed
16 Erit om. F 26 erecta FRVed erecte O
17 linea om. C 29 quia RVed om. COF
19 curvam CRVed om. OF / nonam: consequentiam F 31 ut1 . . . et2 om. Red / elici . . . a O eligi . . . a F probant (?) C probat V /
Quarta conclusio 16a et 17a 7': decima octava septimi V / ut2 om.F
1 Quarta conclusio OV mg. F om. C 4a mg. R PROPOSITIO IIII. ed 32 elici F dici V / post quantitatibus add, Red continuis / igitur C om. Red etiam
2 parabola COFRVed mukefi (?) cf. meam intro. OF ergo V
3 latus CRVed om. OF 34 est C om. OFRVed
4 post perpendicularem add. COF inter caput sectionis et dictum latus erectum 38-39 remotiorem . . . perpendicularem CRVed om. OF
sed cum RVed illud delevi 38 mukefi CR parabola Ved
9 Verbi gratia O om. Red ut igitur FV / abd CV abc OF adb Red 40 perpendicularis2: perpendicularem F
10 pfc CV pfg F pcf Red', et hab. O pfc correctum ex dfe (?) / fo CRVed om. OF / 43 erit (?)C, Red est OF
currat F 44 lineam CV om. OF
13 pg CRed phg OF pge (pg est?) V 46 scilicet COF sicut V
18 duo: secundo F 49 dupla1: regula F
20 3a CRVed 2a OF 50-51 Igitur (ergo in V) . . . hg CV (et cf. Red inferius cap. 5) om. OF
21 necessario om. OF 51 pg2: per g O / fo CRVed of OF
122 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 123

conclusionis proposite cum pg sit latus erectum sectionis et fo eius sagitta. SCRIPTOQUE CIRCULO TRANSEUNTE PER CAPUT SECTIONIS
Et ex consequenti sequitur quod latus erectum sectionis sit necessario 5 ET PER EXTREMITATES DICTE LINEE ORDINIS, CENTRUM CIR­
quadruplum ad medietatem sagitte interceptam inter medium punctum CULI SIC DESCRIPTI ERIT NECESSARIO IN SAGITTA, INSUPER
55 sagitte, quod est h, et caput sectionis, ita quod linea pg necessario est ET DIAMETRI EIUS TRANSEUNTIS PER CAPUT SECTIONIS ET
quadrupla ad lineam fh. PER CENTRUM DICTI CIRCULI PARS INTERCEPTA INTER DIC­
Quantum vero ad secundam partem proposite conclusionis probandam, TAM LINEAM ORDINIS ET CIRCUMFERENTIAM CIRCULI EX
quod videlicet/?# latus erectum sectionis se habet ad perpendicularem mn 10 PARTE BASIS PIRAMIDIS ERIT NECESSARIO EQUALIS ERECTO
sicut illa perpendicularis mn se habet ad mf, describo primo circulum LATERI SECTIONIS.
60 transeuntem per puncta m et n dividentem piramidem equidistantem basi, Probatio. Quod enim circulus potest per dicta tria puncta sic protrahi
cuius dyameter sit linea rms. Cum igitur ms et hl sint equales per 33lam] patet per doctrinam 5e 4* Euclidis. Et quod centrum eius erit necessario
primi, ut superius est argutum, et ex ductu ms in mr habetur quadratum in sagitta liquere potest ex correlario prime 31 Euclidis. Et quod pars
49r linee mn atque ex ductu hl equalis / linee ms in hk resultat quadratum hg, 15 dyametri circuli sic protracti intercepta inter verticem et perpendicularem
igitur ut superius eadem est proportio hk ad mr, et per consequens hf ad mf que est medietas protracte linee ordinis se habet ad illam perpendicu­
65 sicut quadratum linee hg ad quadratum linee mn. Et si sic, igitur per larem sicut illa perpendicularis ad residuum dyametri patet per 8am 6U
quintam decimam 51 Euclidis eadem est proportio pg, quadruple ad hf ut Euclidis. Sed per conclusionem premissam pars dyametri intercepta inter
supra ostensum est, adra/sicut quadrati/?#, quadrupli ad quadratum/?# per verticem et eandem perpendicularem se habet ad illam perpendicularem
corelarium 17l' 6l1 Euclidis, ad quadratum linee mn. Cum igitur per idem 20 sicut illa perpendicularis ad latus erectum mukefi. Igitur latus erectum
corelarium proportio quadrati pg ad quadratum mn sit dupla ad et pars dyametri antedicta intercepta inter illam perpendicularem et cir­
70 proportionem pg ad lineam mn, igitur proportio linee pg ad lineam m f est cumferentiam circuli ex parte basis piramidis necessario sunt equales,
dupla ad proportionem linee/?# ad lineam mn per suppositionem 10am 5U quod fuit probandum. Sit igitur ista conclusio primum preambulum ad
Euclidis. Igitur per idem eadem est proportio linee pg, que est latus propositum.
erectum sectionis, ad lineam mn sicut linee mn ad lineam mf, quod
secundario fuit probandum. Patet igitur conclusio secundum utramque sui Sexta conclusio
75 partem, cuius secunda pars fuit conclusio Apollonii circa cuius proba­
[6] IN SECTIONE MUKEFI PROTRACTA EIUS SAGITTA EX
tionem valde diu insteti (!). Consequenter igitur circa probationem alterius
PARTE UTRIUSQUE EXTREMI QUANTUMLIBET SIGNATAQUE
conclusionis eius insistam; sed prius alias preambulas conclusiones premit-
LINEA RECTA DICTAM SECTIONEM IN ALIQUO PUNCTO IN-
tam ut oportet.
5 TRINSECO (/ EXTRINSECO) LATERIS SUI CURVI CONTIN­
Quinta conclusio GENTE, PROTRACTA QUOUSQUE EX PARTE CAPITIS CONCUR­
[5] PROTRACTA IN SECTIONE MUKEFI QUACUMQUE LINEA RERIT CUM SAGITTA, SI A PUNCTO CONTINGENTIE DESCEN­
ORDINIS INTER LATUS EIUS ERECTUM ET CAPUT IPSIUS DE- DERINT DUE LINEE USQUE AD SAGITTAM QUARUM ALTERA
SIT PERPENDICULARIS SUPER SAGITTAM ET RELIQUA PER-
52 eius CRVed est OF 10 PENDICULARITER EXEAT A LINEA CONTINGENTIE, TUNC
54 post quadruplum scr. OF ad latus seu et delevi EADEM ERIT PROPORTIO LATERIS ERECTI SECTIONIS AD
55 ita . . . est CV itaque linea pg sit OF (cf. Red) PARTEM SAGITTE INTERCEPTAM INTER EXTREMA DICTARUM
58,59 mn CRVed mukefi OF
60 m et η: n et u F
60-63 n . . . equali (/ equalis) rep. et dei. O 7 et1 om. F
61 ms: ns O hic sed hub. O ms in 49r 10-11 erecto lateri CRed erectio lateri V erecti lateris F erectio lateris O (et forte
62 et om. F dei. O -o)
63 mn: mu OF I equalis CV equali OF 13 4* (?) O, R quarti ed quadranguli V quinti F / eius C om. OFRVed
65 mn: mu F 14 post potest scr. et dei. O per dicta tria puncta sic protrahi
67 ad mf V om. OF ad fm Red / sicut: sive F 20 mukefi COFV sectionis mukefi Red / latus illud erectum C
68 mn VRed mu O mk F
Sexta conclusio
69,731·2 mn: mu F 1 Sexta conclusio O mg. F secunda conclusio V 6a mg. R PROPOSITIO
76 post alterius scr. et dei. O probationis VI. ed
78 ut oportet OF om. Red quarum prima est V 4 -5 intrinseco COFV extrinseco Red
Quinta conclusio 5 lateris sui curvi CRed om.hic OF et scr. OF lateris sui post protracta
1 Quinta conclusio O mg. F om. V 5a mg. R PROPOSITIO QUINTA, ed 7 -8 descenderint CRVed ascenderint OF
124 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 125

PERPENDICULARUM SICUT PARTIS SAGITTE INTERCEPTE IN­ TUNC PARS SAGITTE INTERCEPTA INTER DICTAM PERPEN­
TER PERPENDICULAREM PROPINQUIOREM CAPITI ET PUNC- DICULAREM ET CAPUT SECTIONIS ATQUE LINEA INTERCEPTA
15 TUM CONCURSUS LINEE CONTINGENTIE CUM SAGITTA AD INTER IDEM CAPUT ET PUNCTUM CONCURSUS LINEE CON-
PARTEM SAGITTE INTERCEPTAM INTER EANDEM PERPENDIC­ io TINGENTIE CUM SAGITTA SIC ULTERIUS PROTRACTA NECES­
ULAREM ET VERTICEM SECTIONIS. SARIO ERUNT EQUALES.
Probatio. Sit sectio mukefi abc' [Fig. 4.8], super cuius sagittam a f de­ Verbi gratia, sit ut in presenti figura [vide iterum Fig. 4.8] sectio mukefi
scendat perpendicularis de, et in puncto d contingat linea recta que sit abc, quam in puncto d contingat linea dg protensa quousque in puncto
49v 20 dg sectionem mukefi, que quidem linea dg / concurrat cum sagitta ul­ g concurrerit cum sagitta fa ulterius lineata. Tunc dico quod linea ea et
terius protracta in puncto g, et a puncto d exeat perpendicularis super 15 linea ag sunt omnino equales, est enim vera eius conversa, scilicet, si
lineam dg quousque concurrat cum sagitta in puncto h . Quia igitur angulus ag et ae sunt equales, linea gd transiens per punctum d continget sec­
gdh est rectus a quo descendit linea de perpendiculariter super basim tionem in d puncto.
trianguli gdh, erit proportio he ad ed sicut ed ad eg per octavam 6U Probatio. Quia si non, igitur secabit sectionem per conclusionem secun­
25 Euclidis. Sed per ante premissam latus erectum sectionis se habet ad ed dam; et si sic, igitur vel inter d et a vel inter d et basim sectionis. Sed
sicut ed ad ea. Igitur quadratum de tam resultat ex ductu lateris erecti 20 quod non inter d et a probatur; quia si sic, secet in puncto k, a quo
sectionis in ea quam ex ductu he in eg. Igitur per 13am 6li Euclidis eadem descendat linea kl perpendiculariter super sagittam; quia igitur ga et ae
est proportio erecti lateris sectionis ad he partem sagitte interceptam posite sunt equales, erit ga maior quam al. Esto quod per lineam gm,
inter extrema e et h duarum perpendicularum de et he sicut eg ad ea, eruntque gm et el equales atque al et am equales, et per consequens
30 hoc est, sicut partis sagitte intercepte inter perpendicularem ed et punctum ml erit dupla ad al. Sed ut patuit ex probatione conclusionis quarte pro-
concursus linee contingentie cum sagitta ad partem sagitte interceptam in­ 25 portio ae ad la est dupla proportionis ed ad Ik, et sicut ed ad Ik ita
ter eandem perpendicularem de et caput sectionis, quod fuit probandum. eg ad Ig per quartam 6* Euclidis. Igitur proportio ea ad la est dupla
Istas duas ultimas conclusiones inserui non tanquam necessarias ad proportionis eg ad Ig, sed sicut ea ad la ita duplum unius ad duplum
nostrum propositum, verum quia notabiles videbantur. Modo vero conse- alterius per 15am 5U Euclidis. Igitur proportio eg ad Im est dupla pro­
35 quenter incedam ad probationem alterius conclusionis Apollonii, que est ista. portionis eg ad Ig·, et, si sic, igitur per decimam diffinitionem 5li Euclidis
30 sicut eg ad Ig ita Ig ad Im. Cum igitur eg tota se habet ad Ig totam
Septima conclusio sicut idem Ig pars totius eg ad Im partem totius Ig, igitur per 19am 5li
Euclidis el, que est residuum totius eg, se habet ad mg, que est residuum
[7] QUAMCUNQUE SECTIONEM MUKEFI LINEA RECTA totius Ig, sicut eg ad Ig. Sed eg maior est quam Ig, quia totum maius
CONTIGERIT QUE LINEA SIC CONTINGENS PROTRAHATUR est sua parte. Igitur el est maior mg, et per consequens si eis addantur
QUOUSQUE EX PARTE CAPITIS CONCURRERIT CUM SAGITTA 35 equalia, scilicet al ad le et am ad mg, erit per communem animi con­
5 ULTERIUS QUANTUMLIBET PROTENSA ET A PUNCTO CONTIN­ ceptionem ag minor ae; sed posite sunt equales. Igitur idem est minus
GENTIE DESCENDERIT PERPENDICULARIS SUPER SAGITTAM,
15 cum sagitta CRVed om. OF 8 et caput sectionis CRVed om. OF
19 perpendicularis CRed perpendiculariter OFV / recta om. OF 13 in1 . . . quousque CFV (cf. Red) om. O
20 concurrat CRVed contingat OF 14 fa CRVed pa OF I ulterius . . . ea CFV om. O
20-21 ulterius CRVed om. OF 15 post ag add. F et ae / omnino equales tr. F
23 gdh CRVed dgh OF I rectus CRVed rectogonus OF 15-16 est . . . equales CFV (cf. Red) om. O
25 ante OFRVed om. C 15 vera CRVed una F
26 ed: eg O / tam: cum F / resultat CV resultet O resultet tam F 16 continget C contingeret F
26-27 ex . . . quam CFV om. O / lateris erecti tr. C 18-19 secundam CRVed om. OF
27 he om. O 19 sectionis CRVed erit sectio communis OF
29-30 et2 . . . ed CF om. O 22 per lineam: pro linea F
33-35 Istas. . . . ista COF om. RVed 25 ae: ea F / et RVed om. OF / sicut . . . Ik FRVed om. O / ita: in F
33 conclusiones CF considerationes O 26 quartam V (cf. Red) secundam O duplam F / proportio V (cf. Red) om. OF
34 verum C sed OF / post notabiles add. OF mihi 26-27 ea . . . sed O om. F
27 eg: ag O
Septima conclusio 30 lg2 om. O
1 Septima conclusio O mg. F 7a mj?. R Tertia conclusio V PROPOSITIO 35 erit OFRVed igitur C
VII. ed 36 minor Red maior OFV maior erit C I minus RVed maius OF
126 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 127

et equale eidem, quod est impossibile; et hoc sequitur si gd secet figuram 65 Ionii ab eam allegantibus non probata. Et ideo circa probationem earum
mukefi inter a et d. Igitur hoc est falsum. studui multum diu et eam tandem cum dei adiutorio adinveni, cui sit
Quod si dicatur quod gd secabit figuram mukefi non inter d et a sed honor et gloria per seculum seculorum.
40 inter d et c esto quod in puncto o, et tunc descendat linea on perpen-
diculariter super sagittam an, et ponantur ag et ae equales sicut prius, Octava conclusio
et quantum distat n ab e ex parte basis tantum distet / p a g supra g. [8] IN OMNI SECTIONE PARABOLA SIGNATO PUNCTO IN MEDIO
Erunt igitur na et pa equales, et per consequens np erit dupla ad na et SAGITTE IPSIUS, QUI QUIDEM PUNCTUS ERIT ETIAM IN MEDIO
eg dupla ad ea. Arguatur igitur, sicut liquet ex quarta huius et eius pro- LATERIS EIUS ERECTI, TUNC A QUOCUNQUE PUNCTO LA-
45 batione, quod proportio linee na ad lineam ea est dupla proportionis
5 TERIS SUI CURVI DESCENDERIT RECTA LINEA VERSUS BASIM
linee no ad lineam ed, igitur et dupla proportionis linee ng ad lineam eg EQUIDISTANTER SAGITTE, SI AB EODEM PUNCTO IN CURVO
per quartam 6“ Euclidis. Sed sicut na ad ea ita duplum unius ad duplum LATERE EXISTENTE EXIERINT ALIE DUE LINEE RECTE
alterius per 15am 5a Euclidis. Igitur proportio np ad eg est dupla propor­ QUARUM ALTERA EX UTRAQUE PARTE PROTENSA CON­
tionis ng ad eg. Igitur per diffinitionem decimam quinti Euclidis sicut TINGAT SECTIONEM ET RELIQUA PROTENDATUR QUOUSQUE
50 np ad ng ita ng ad eg. Cum igitur np totum se habet ad ng totum sicut 10 IN PUNCTO PRIUS SIGNATO CONCURRERIT CUM SAGITTA,
idem ng pars totius np ad eg partem totius ng, igitur per 19am quinti CAUSABUNTUR DUO ANGULI EQUALES QUORUM ALTER
Euclidispg, que est residuum totius np, se habet ad ne, que est residuum CAUSABITUR EX LINEA CONTINGENTE A PARTE CAPITIS SEC­
totius ng, sicut np ad ng. Cum igitur np sit maior quam ng quia totum TIONIS ET EX ILLA PROTENSA USQUE AD PUNCTUM SIG­
maius sua parte, igitur pg erit maior quam ne. Consequens falsum quia NATUM ET RELIQUUS CAUSABITUR EX LINEA PROTRACTA
55 ex ypotesi posite fuerunt equales et nichil potest esse equale alteri et 15 EQUIDISTANTER SAGITTE ET EX LINEA CONTINGENTE A
maius eodem. Sed hoc sequitur si linea gd secet figuram mukefi inter
PARTE ALTERA VERSUS BASIM.
d et c; igitur falsum est gd secare figuram mukefi inter d et c et prius Verbi gratia, sit in sectione parabola abc [Fig. 4.9] sagitta ak protensa
ostensum est quod non secat eam inter d et a. Igitur nullicubi secabit; quousque concurrerit cum axe piramidis cuius punctus medius sit e , erit
igitur necessario continget in puncto d si ae et ag sunt equales. Et si sic, igitur et punctus e medius punctus erecti lateris sectionis, ut patet ex
60 igitur et econtra: si gd contingat sectionem in puncto d, tunc ae et ag erunt 20 quarta descriptione huius. Signetur igitur punctus d in curvo latere sec­
equales, quod fuit probandum. Patet igitur probatio huius conclusionis, tionis, a quo exeat linea dl equidistanter sagitte ak, et in eodem puncto d
que fuit conclusio Apollonii a Vitellone et aliis perspectiviis tanquam contingat lineagdh dictam sectionem, protrahaturque linea de. Tunc dico
principium allegata sed a nullis eorum probata. Patet etiam probatio quarte quod angulus edg et angulus Idh necessario erunt equales.
conclusionis, cuius secunda pars fuit consimiliter conclusio eiusdem Apol- Probatio. Angulus enim deg vel est acutus vel rectus vel obtusus. Sit,
25 gratia exempli, acutus primo, et a puncto d ducatur df perpendicularis
40 c OFV (et etiam in C?) basim Red / o: c F / on: cn F super sagittam, que per 32am primi Euclidis cadet inter puncta a e t e .
40-41 perpendiculariter tr. F post descendat Et protrahatur ea quousque concurrerit cum linea contingente in puncto
41 sagitta F I ponatur F 50v g, eritque angulus fg/d acutus per 32am primi Euclidis, et erit per pre-
42 g2: e F
43 Erunt igitur C igitur autem O igitur erunt F / na1: η O
missam ag equalis af. Quia igitur linea ae est divisa in puncto / et illi
44 Arguatur igitur C tr OF Arguitur igitur V / sicut: sic (?)C 30 additur ag equalis alteri parti dividentium, erit per 8am 21 Euclidis quad-
47 quartam V (cf. lin. 26 superius) secundam OF
49 ng FRed np COV / eg FRed ng COV 67 per seculum seculorum C om. OF
51 IO3™: septuagesimam nonam O
53 np ad ng CFRVed ng ad np O Octava conclusio
54 quia CV cum OP 1 Octava conclusio scripsi, om. C viii. conclusio mg. F 10a conclusio O 8a mg. R
55 posite V positi COF / fuerunt C fuerint V sunt OF / potest esse: est V Quarta conclusio V PROPOSITIO VIII. ed
55-56 equale . . . eodem CV maius altero et equale eidem OF 13-14 signatum CRVed om. OF
57 c 1 COFV basem Red 15 contingente COF contingencie RV contactus ed
58 est OFV om. C 18 post piramidis add. Red in k / erit CV om. OF
60 econtra CRVed regula OF / ae et ag: ag et ae F 19 medius punctus CV tr. OF
61-67 Patet . . . gloria COF Patet ergo conclusio Appolonii ante probata et allegata V 27 ea: ae F I concurrerit C concurrat OFV
62 Vitellone C Vitulone OF / prospectiviis F 29 linea ae CV tr. OF
63 nullis C nullo OF / probatio tr. F post conclusionis 30 dividentium CRFV dividendum O segmentorum ed
128 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALMUKEFI COMPOSITIO 129

ratum linee eg equale ei quod fit ex ductu linee ea in lineam ag vel in sibi 60 quam linea ae. Sed per premissam af et ag sunt equales. Igitur ag est
equale af quater et quadrato linee f e . Sed linea ea est quarta pars erecti maior quam ae. Fiat igitur per 3an' primi Euclidis linea ani equalis ae.
lateris sectionis parabole ex prima parte 4e huius. Igitur per primam 6' Remanet igitur gm linea equalis linee fe, et per consequens linea em
et per primam 21 Euclidis illud quod fit ex ductu a f in ae quater est equale addita utrobique erit linea fm equalis linee ge. Quadratum igitur linee
35 ei quod fit ex ductu af in latus erectum sectionis semel, quod quidem fm est equale quadrato linee ge. Quia igitur linea fa est divisa in puncto
latus est linea bc per descriptionem 4amhuius. Illud ergo quod fit ex ductu 65 e, cui addita est linea am equalis uni dividentium, scilicet linee ae, patet
af in bc cum quadrato fe est equale quadrato linee eg. Sed per 2am partem igitur per 8am 2* Euclidis quod illud quod fit ex ductu linee af in lineam
4t0 huius patet quod illud quod fit ex ductu af in latus erectum valet quadratum am, vel in sibi equale ae, quater cum quadrato linee/p est equale quadrato
df, cum sit perpendicularis super sagittam, duo vero quadrata d f et fe linee fm vel linee eg que sunt equales. Sed illud quod fit ex ductu linee
40 valent quadratum de linee per penultimam primi Euclidis. Quadrata igitur fa in lineam ae quater, ut patet ex premissis, est equale ei quod fit ex
linearum eg et ed sunt equalia. Igitur linea ed est equalis linee eg. Igitur 70 ductu linee fa in latus erectum sectionis. Igitur illud quod fit ex ductu
per 5am primi Euclidis in trigono edg angulus edg est equalis angulo egd. fa in latus rectum sectionis cum quadrato fe linee valet quadratum linee
Sed linee dl et ae sunt equidistantes ex ypotesi. Igitur per 29am primi fm. Sed per secundam partem 4e huius illud quod fit ex ductu fa in latus
Euclidis angulus Idh extrinsecus est equalis angulo egd intrinseco. Angulus rectum sectionis valet quadratum fd. Igitur quadratum fd cum quadrato
45 igitur edg est equalis angulo Idh, quod fuit probandum. Et eodemmodo fe valet quadratum fm. Cum igitur per dulcarnon quadratum de valet duo
probabitur de qualibet linea equidistante sagitte et copulata cum puncto 75 quadrata df et fe, igitur quadratum de valet quadratum fm et per conse­
e ita quod e sit angulus acutus. quens valet quadratum sui equalis eg. Et si sic, duo quadrata de et eg
Quodsi angulus deg sit rectus [Fig. 4.10], adhuc probabitur propositum. sunt equalia et per consequens linee de et eg sunt equales, quare et
Quod e angulus est rectus, ideo linea de est perpendicularis super sagit- 5lr anguli edg I et egd sunt equales. Igitur et angulus Idh equalis uni erit
50 tam. Igitur per premissam linea ag est equalis linee ae. Sed linea ae, equalis alteri. Sunt igitur anguli Idh et edg equales, quod fuit probandum.
ut prius, est quarta pars erecti lateris sectionis per primam partem 4e 80 Sic igitur patet veritas conclusionis sive deg angulus acutus seu rectus
huius. Igitur lineage, que est dupla linee ae per premissam, est medietas fuerit vel obtusus.
erecti lateris sectionis. Sed ed est medietas erecti lateris sectionis cum
Nona conclusio
e sit medius punctus eius ex 4a descriptione huius et ex ypotesi. Igitur
55 de et ge linee sunt equales. Igitur per 5am primi Euclidis erunt anguli [9] IN SPECULO CONCAVATO CONCAVITATE SECTIONIS
edg et egd equales. Quare per 29am primi Euclidis erunt anguli Idh et egd MUKEFI OMNES RADII SOLARES SUPER DICTUM SPECULUM
equales, et per consequens edg et Idh sunt equales, quod fuit probandum.
60 premissa F
Si vero angulus deg sit obtusus [Fig. 4.11], ducatur linea perpendicu­ 65 am equalis: a inequalis F
laris df. Erit igitur punctus / inter puncta e et k. Linea igitur a f est maior 67 quater: quartus F
68 linee1 om.OF / illud C idem OF
32 quater: quadrato O 69 quater CRVed om. OF
33 parte RVed om. OF 70-71 Igitur . . . sectionis om. O
33-34 6‘ . . . 2' Red 21 vel per primam T Ο T vel per primam Θ F quinti et per 70 illud CV istud F
primam secundi V 71 rectum C erectum FV
34 illud O om. Red istud F id V / ae: ac V / quater om. O 72 illud C idem O id est F
37 bc: bf O 74 dulcarnon C dul. OF penultimam primi V
37-38 cum. . . . valet Red om. OF (et hoc non possum bene legere in C) cum ΙΑ- l bvalet2. . . . de om. V
75-76 valet . . . equalis C equale O valet quadratum fm et per consequens valet de
quarto (!) fe est equale quarto (/) linee eg. Sed per secundam partem quarte
huius patet quod id quod fit ex ductu af in bc videlicet V equale F
38 patet quod illud C om. Red / valet Red videlicet OFV 76 eg1: ag O
39 vero CRVed om. OF 77 et eg om. F
40 valent Red sunt equalia V sunt perpendicularia non equalia OF / per . . . 78 sunt OFV om. C
Euclidis V om. OF per penultimam primi Red 79 edg corr. mg. O ex egd in textu
44 angulus1RVed om. OF 80-81 Sic . . . obtusus COF om. V
46 coppulata F 80 post angulus hub. OF sit / seu C sive OF
AI acutus RVed om. OF Nona conclusio
48 sit V erit OF / adhuc V ad huc O ad hoc F 1 Nona conclusio scripsi Conclusio 9a mg. C 9U mg. R undecima conclusio O
49 Quod OF quia V V1III. conclusio F Quinta conclusio V PROPOSITIO IX. ed
56-57 Quare . . . equales' om. F 2 concavato CRVed concavo OF / post concavato hub. V cum
130 ARCHIMEDES IN THE MIDDLE AGES SPECULI A LM U K EFl COMPOSITIO 131

INCIDENTES EQUIDISTANTER SAGITTE AD UNUM ET IDEM quantitatem dicte semidiametri, et in secunda regula lineetur portio def
s PUNCTUM REFLECTENTUR, QUI QUIDEM PUNCTUS TAM ERIT 15 secundum quantitatem semidiametri brevioris priori, ut gratia exempli
MEDIUS SAGITTE QUAM ETIAM RECTI LATERIS SECTIONIS. per quantitatem pedalem. Deinde ab utraque regula aufferatur cum lima
Hec conclusio statim patet ad oculum ex premissa et ex isto principio quicquid fuerit extra gibbositatem dictarum portionum in regulis protrac­
perspective, quod angulus incidentie et angulus reflectionis necessario tarum. Deinde iaceat super superficiem orizontis corpus oblongum quad­
sunt equales. rangulum sex basium equidistantium et equidistantium linearum atque
io Premissis autem conclusionibus istis novem procedam ulterius ad prin­ 20 rectangulum cuius longitudo sit ut longitudo regule abc, et sit gratia ex­
cipale propositum prosequendum, pro quo requiritur fabricatio cuiusdam empli longitudinis octo pedum, latitudinis trium pedum et altitudinis unius
instrumenti quo calibs vel aliquod tale durum concavitate mukefi poterit pedis, et sit latus eius superius ghdf, et iaceat super basim ikac. Protraho
concavari. Pro doctrina igitur fabricandi huiusmodi instrumentum sit igitur super latus eius superius unam lineam equidistantem linee gh et
decima huius tractatus conclusio ista sequens. distantem ab ea per pedalem distantiam, scilicet secundum quantitatem
25 qua semidiameter circuli d e f est minor semidiametro circuli abc. Et sit
punctus medius illius linee e. Deinde reguletur def super latus superius
Decima conclusio corporis oblongi ita quod punctus medius curvature regule def contingat
[10] INSTRUMENTUM PERFICERE QUO CONCAVITATE PARA­ lineam e equidistantem linee gh in puncto e, et secundum istam regulam
BOLA CAVARI POTERIT MATERIA SOLIDA SPECULI COM- def sic curvatam protrahatur linea curva super latus superius corporis
BUSTIVI. 30 oblongi, que curvatura in illo latere vocetur def. Deinde ex transverso
5 Pro isto igitur considerari oportet primo ad quantam distantiam velis lateris superioris dicti corporis oblongi protraho lineam rectam transeun­
quod speculum flendum comburat, ut verbi gratia ad 20 pedalia. Et tunc tem per punctum e , que quidem linea erit equidistans utrique extremorum
oportebit ut habeas aliquam perticam sive virgam duple longitudinis, ut illius corporis, propter hoc quod punctus medius distat equaliter ab ex­
40 pedalium, cum qua tanquam semidiametro circuli portio poterit lineari. tremis. Deinde ab extremis illius linee medie transeant alie due linee ex
Habeantur igitur due regule plane quadrangulares eree seu ferree quarum 35 transverso laterum costarum illius corporis equidistantes ab eius extremis
io utraque sit tante longitudinis quanta debet esse speculi latitudo, sitque quousque concurrerint cum basi super quam iacet illud corpus. Et inter
utriusque regule latitudo tanta ut in ea ab uno eius extremo ad aliud siv extrema inferioria illarum duarum linearum sic des/cendentium protra­
lineari poterit portio circuli secundum quantitatem semidiametri supra- hatur linea recta ex transverso basis, que etiam linea erit equidistans
dicte. In altera igitur regularum lineetur portio abc [Fig. 4.12] secundum extremis dicte basis. Verbi gratia ubi dicta linea in basi tangit lineam
40 ik signetur punctus b , et vocetur linea transversalis basis linea bm. Elevetur
igitur basis et iaceat super latus ghfd oppositum basi. Deinde applicetur
4 equidistanter CRVed equidistantes OF regula abc super illud latus ikac ita quod punctus medius curvature regule
5 reflectentur CF reflectuntur O relucent Red relevant (?) V
8 reflectionis CO reflexionis FRVed abc contingat lineam ik in puncto b , et secundum illam regulam sic curva­
10-14 Premissis. . . . sequens COF om. RVed tam abc protrahatur linea curva super illud latus ikac, que curvatura in
10 novem CF xi O 45 illo latere vocetur abc. Fiunt igitur due curvature in lateribus oppositis
13 huiusmodi C hoc OF que sunt abc et def portiones duorum circulorum quorum centrum su-
14 decima CF undecima O
14 de portio mg. scr. O circuli alia
Decima conclusio
15 post quantitatem scr. O dicte sed non in aliis MSS / priori COF parum FV
1 Decima conclusio scripsi, om. CV X. conclusio mg. F 10a mg. R XII. con­
paulo ed
clusio O PROPOSITIO X. ed
16 Deinde: demum F I auferatur F / lima CRed, (7)0 linea FV
2 Instrumentum COF Instrumentum de calibe vel de (om. V) alia materia dura 17 fuerit CRVed om. OF
RVed
18 Deinde: demum OF
3 solida CRVed om. OF
22 pedis om. F
5 oportet O opportet F oporteret (?) C
23-24 linee gh et distantem CRVed om. OF
7 ut1 CRVed quod OF / sive C seu OF vel RVed 28 lineam e CV om. Red e lineam OF / gh RVed ik OF
8 post tanquam scr. C cum
29 curvatam correxi ex curvam in COFV (cf. lin. 44-55)
9 erree O
34 illius linee FV tr. O
11 ad CRVed in OF
39 ubi RVed om. OF
12 portio: proportio F
44 illud V idem OF
13 lineetur CFRVed lineatur O
46 et om. C
 SPECULI ALM UK EFI COMPOSITIO 133
132 ARCHIMEDES IN THE MIDDLE AGES

per n, que sit linea equidistans linee ibk, per cuius longitudinem et punc­
perioris directe supraponatur centro inferioris sic quod linea protracta
tum e transeat una superficies plana abscindens quicquid est extra illam
inter centra est perpendicularis super utrumque circulum. Per instru­
superficiem ex parte b. Differentia igitur communis illi superficiei plane
mentum igitur cuius acies sit linea recta abradatur ex costa laterali que
et superficiei curve que est abcdef est sectio mukefi. Nam si ymaginemur
so est ghki quicquid fuerit extra curvaturas abc et def. Relinquetur igitur
80 superficiem planam triangulam transeuntem per conum et centrum basis
in corpore illo oblongo una curva superficies que est def abc, que quidem
piramidis atque per lineam bon cuius latera sunt in superficie curva pira­
superficies sic curvata est portio curve superficiei piramidis rectangule
inter duos circulos equidistantes. midis dividentem piramidem in duo equalia, tunc linea en erit equidistans
alteri lateri dicti trianguli, linea enim bn protracta usque ad latus trianguli
Quod autem sit portio curve superficiei piramidis satis liquet, eo quod
ex parte n illa linea causabit cum dicto latere ex parte coni angulum
55 centra duorum circulorum quorum portiones constituunt latera curva
85 equalem medietati recti, eo quod illa linea erit equidistans dyametro
superficiei proposite sunt in linea perpendiculari super utrumque circu­
52r / basis piramidis. Sed angulus eno est medietas recti. Igitur en et dictum
lum. Et quod illa piramis sit rectangula probatio, iaceat enim corpus
latus equidistant per 28am primi Euclidis. Sic igitur habetur superficies
oblongum super basim eius priorem ita quod curvatura abc sit in latere
plana disposita in lateribus secundum sectionem mukefi. Illi igitur applice­
inferiori et curvatura def in latere superiori. Tunc a puncto e ymaginemur
tur lamina de calibe et formetur secundum formam dicte sectionis. Et
60 lineam perpendiculariter descendere super basim que necessario cadet
90 postea secundum eandem formam acuatur ita quod acuties (/ acies) eius
in linea bm transversali basis. Cadat igitur gratia exempli in puncto o,
sit linea curvata curvitate mukefi. Deinde cum illo instrumento cavetur
et sit perpendicularis eo. Cum igitur linea eo precise contineat pedale
alia lamina calibea vel ferrea et postea poliatur et habebis speculum com­
eo quod altitudo corporis oblongi fuit posita pedalis ex ypotesi et cum
burens ad tantam distantiam quanta est medietas semidiametri circuli
etiam linea ob sit pedalis eo quod e punctus distat a linea gh per pedale
cuius portio abc lineatur in prima regula superius designata.
65 ex ypotesi, igitur in triangulo eob duo latera eo et ob sunt equalia et
95 Quod autem ad tantam comburet distantiam patet ex nona conclusione
angulus o rectus. Igitur angulorum obe et beo uterque est medietas recti.
huius et ex communi omnium perspectivorum principio, quod scilicet in
Cum igitur linea oe sit equidistans axi piramidis, si transierit be linea
omni reflectione radii solaris vel alterius luminaris angulus incidentie et
quousque concurreret cum axe, esset angulus beo equalis angulo causato
angulus reflectionis necessario sunt equales. Et sic patet propositum quod
ex concursu linee be cum axe, scilicet medietas recti. Et per consequens
in principio huius opusculi est premissum.
70 totalis angulus coni piramidis esset rectus. Igitur est piramis rectangula
per primam descriptionem huius et superficies curva abcdef est portio
superficiei curve piramidis rectangule, quod fuit probandum. Conditiones boni calibis
Istis sic habitis super lineam bo protractam ulterius ex parte o signetur
Verum quia valde requiritur quod instrumentum concavandi de bono
punctus n, qui tantum distet ab o quantum o distat a b, ita quod o sit
et electo calibe fuerit atque duro, ideo nota tres conditiones boni calibis,
75 in medio inter b et n . Protrahatur igitur super basim ikac una linea transiens
quarum prima est lenitas superficiei exterioris cum continuatione bona

47 supraponatur R supponatur OF supponitur V

48 centra CRed centrum OFV 76 n: m F
49 acies OFVed acuties CR 77 e CRed om. V m F I illam CRVed istam OF
52 curvata correxi ex curva in MSS 79 ymaginaremur O imaginaremur F
57 probatio COF lac.V probatur Red / iaceat CRV iacet OF incumbat ed 82 dividente O
59 inferiori: interiori F / e om. F I imaginetur F 84 n (corr. ex m ?) . . . linea C enim illa linea OF n RVed
60 perpendiculariter CF perpendicularem ORVed / que CRVed qui OF 86 eno Red enm COF (cf. var. lect. lin. 62) amn V
61 transversali RVed transversalem OF / o Red m COFV 90 acuatur OFRVed acuetur C
62 eo1 correxi ex em in COFV (et om. Red) / eo2 Red em COF bm V (et ubique 91 curvata correxi ex curva
COFV hub. m pro o; forte in textu originali m et o transponebantur sed cf. 92 polliatur O
meam intro.) 95-99 Quod. . . . premissum om. Red (sed cf. textum Red in cap. 5)
63-65 et cum . . . ypotesi CV (cf. Red) om. OF 96 quod scilicet CO tr FV
65 post equalia injuste scr. OF tertio sed non in CRVed 97 reflectione O reflexione CFV / anguli V
66 obe correxi ex mbe in FV et mba in O 98 angulus reflectionis C reflexionis angulus OF reflexionis V / Et sic OF Sic igitur
68 angulus: triangulus OF C sic V
69 medietas COFV medietati (?)R
71 abcdef CRVed abc et def OF Conditiones boni calibis
1-104 Conditiones. . . . polietur om. Red (sed cf. additionem mathematicam in
72 superficei curve CRV tr. OF
75-77 Protrahatur . . . e om. O Red)
1 Conditiones . . . calibis mg. O om. CFV
75 igitur om. F / ikca F
134 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 135

5 partium, scilicet sine fissuris. Secunda est facilitas frangendi et tertia est indigeris ad opus tuum. Deinde consolidentur hec masse adinvicem bine
claritas coloris in loco fractionis. Unde prima et tertia eius munditiam 40 et bine in hunc modum. Igniatur aliqua consolidandarum et ignita per
designant. Sed media ipsius duritiem attestatur. Fit autem purgatio seu ferrum scindens exasperetur ut cum alia melius poterit incorporari. Deinde
depuratio calibis per hunc modum. Accipiatur petia ferrea oblonga ad superponatur una alteri et ambe in ignem mittantur et igniantur atque
modum virge ferree aliquantulum late latitudine, scilicet unius pollicis. bene consolidentur. Deinde extrahantur et adinvicem malleentur et firmiter
10 Applicentur petie calibis invicem fracte, bene adinvicem sine vacuo situate. consolidentur. Et sicut dictum est de hiis duabus similiter fiat de aliis
Deinde aspergatur tota hec massa cum aqua fabrili cui admixta erit quedam 45 duabus et tribus. Quibus factis confringantur iterum hec masse et iterum
maneries terre crocee qua utuntur omnes fabri pro consolidatione calibis malleentur et consolidentur, et hoc totiens fiet quousque calibs sit suffi­
et ferri, nam sine tali intinctione nullo modo possent petie calibis nec cienter mundatus. Tunc ad unam massam omnes petie redigantur que
cum ferro nec sibi invicem consolidari. Virga vero ferrea ideo exigitur per ignem debet mollificari et diligentissime malleari et ad figuram intentam
15 quatenus super eam fiat fabricatio et malleatio pro consolidatione petiarum redigi competenter ne sit vesica aliqua vel fissura seu inmunditia ex neg-
calibis adinvicem et etiam propter conservationem calibis ne per ignem 50 ligentia malleationis. Deinde indurari debet. Circa quod nota quod tres
aliquid de puro consumatur sicut conservatur argentum in sua depuratione sunt modi calibem indurandi, quorum primus efficit calibem durissimum
per plumbum sibi admixtum. Deinde ponatur hec massa sic aspersa in et frangibilissimum et hoc fit per appositionem eius in aquam frigidam
igne et hec in fabrica docentur et aperte preparata, cuius operatio est post magnam ignituram cum incendit videlicet calor igniture ad albedinem
20 quod fiat in fundo quedam apertura in qua transcinduntur virge ferree se qualis coloris est sol inter orientem et medium celi. Secundus modus
orthogonaliter intersecantes ad modum fenestrarum ferrearum, et per hanc 55 efficit calibem durum atque frangibilem sed non ita excellenter ut prius
aperturam possit inmunditia metalli inferius descendere et fabrica munda et hoc fit per appositionem eius in aquam frigidam cum rubuerit in igne
conservari. Et postquam fuerit illa massa bene ignita ita quod sit quasi ad modum cerase mature, ipso postea licet parum calefacto quousque
alba extrahatur et malleetur bene et fortiter ut consolidentur petie calibis. appareat in eo color ferri purpureus, et iste modus conpetit instrumentis
25 Et iterum ponatur in igne et dum est in igne aspergatur frequentqr arena scindentibus ferrum aut calibem mollem aut latonum aut aliud genus duri
et zabulo pro meliori consolidatione facienda. Et iterum malleetur bene 60 metalli, et hic modus est conpetens speculo fiendo de calibe; tamen cum
et fortiter quousque facta fuerit firma consolidatio et sufficiens. Verum- reponitur ab igne intingatur petia circumferentialis exterior ad quantitatem
tamen in qualibet extractione ab igne considerari debet si poterit inmundi­ duorum digitorum aqua predicta sic ®, ne debitam concavitatem admittat
tia aliqua inveniri, que discerneretur per nigredinem aliquam apparentem ad quam prius per malleationem diminutum erat. Et post extractionem
30 in corpore metalli infra superficiem eius ignitam, que si inventa fuerit eius ab aqua calefiat solum eius medium; tamen iste modus calefaciendi
per instrumentum scindens quo utuntur fabri abscindatur. Post hoc scinda­ 65 inducit diversitatem colorum in superficie speculi post eius polituram.
tur massa sic consolidata dum est bene ignita in plures petias per idem Tertius modus indurationis in hoc differt a secundo, quod ultima cale­
instrumentum partibus abinvicem non omnino separatis vel abscisis sed factio expectat colorem in calibe purpureum qualis est color viole, et iste
aliqualiter sibi invicem adherentibus. Deinde induretur hec massa per hunc modus calefactionis conpetit cultellis et instrumentis ordinatis ad ligna
35 modum. Reponatur in igne quousque rubuerit ad modum cerase mature scindenda. Alium insuper modum indurandi calibem ponit Albertus in de
et postea submergatur in aqua donec ignis extinguatur et calor consume-
52v tur. Deinde confringatur / cum malleo in petias in locis incisionum pre- 39 tuum V unum OF
dictarum. Et per hunc modum fiet alia massa et tertia et quarta quot 41 incorporari OF consolidari et incorporari V
42 superponatur O supponatur V supponantur F
4 3 - 44 Deinde . . . consolidentur FV om. O
5 scilicet om. F / Secunda OFV alia (?) C 43 malentur V
7 ipsius C eius OF / seu C om. V vel OF 4 4 - 45 similiter . . . duabus F om. O sic dicendum est de aliis duabus V
9 virge bis F / latitudine om. F / pollicis C policis OFV 45 et tribus OF om. V / fractis F / confringantur V confragantur OF
13 intinctione F, (?)C intentione O fractione V 46 malleentur F maleentur OV I fiet V om. OF
14 sibi OFV ad C 47 que OF et V
15 malleatio CF maleatio OV 48 malleari CF male ari OV
18 hec massa CV tr. OF 49 vestica F / fixura F
24 malleetur CF maleetur OV 50 malleationis CF maleationis OV
26 malleetur F maleetur OV (et non possum legere C hic) 52 aquam CFV aqua O
31 abscindatur OF abscindatur et abscindatur V 56 per OV propter F
35 igne FV ignem O 59 mollem (?) C om. OF molem V
37 confringatur V confrangatur OF / malleo F maleo OV / locis OV locum F 63 malleationem CF maleationem OV
136 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 137

70 mineralibus, libro secundo, tractatu 4° (/ 3°), capitulo 2°, de causa ymagi- sanguinem et ponunt super superficiem illam de pulvere calcis et accipiunt
num apparentium in lapidibus, ita dicens, “ que autem experti sumus hic too instrumentum ferreum ad hoc preparatum et forbiunt sicut prius quousque
dicimus. Induratur enim et depuratur calibs sepius donec fere habeat ydola appareant et quousque nichil de vestigiis Iapidum precedentium
albedinem argenti, et tunc ex eo formantur instrumenta sculpta angulis appareat quovis modo. Licet igitur ex premissis quomodo fabricari aut
convenientibus et subtilibus, et tunc exprimitur succus rafani et permisce- quomodo debeat indurari instrumentum quo speculum combustivum
75 tur cum succo rafani aqua que extrahitur de lumbricis terre contusis et causabitur aut insuper quomodo dictum speculum polietur. Sic igitur ad
expressis per pannum ita quod tantum sit de uno quantum de altero, et 105 nostri principalis propositi complementum (/ completionem) deveniemus
postea candens instrumentum extinguitur in aqua illa bis vel ter vel (/ devenimus). Verum licet optatum nostrum propositum ad optatum
plures quotiens oportet quod efficitur ita durum quod radit gemmas et finem deduximus, quia tamen ex sectione addita sequitur quedam con­
incidit aliud ferrum sicut plumbum.” Hec Albertus. clusio que prima facie videtur mirabilis, ideo dictam conclusionem hic
80 Quantum vero ad polituram in speculo nostro fiendam valet emerillus adiungere dignum duximus, et est ista.
qui est quidam lapis habens colorem ferreum sicut adamas bonus aut
habens colorem citrinum obscurum et est similis cilicibus repertis in aquis [Undecima] conclusio
correntibus et debet pulvericari in mortario eneo et pulvis colari per pan­
num lineum vel cribrari per cribrum de pilis spisse contextum. Qui quidem [11] DUAS LINEAS QUARUM ALTERA EST RECTA ET ALTERA
85 pulvis debet misceri cum aqua et totum commixtum poni debet super CURVA VEL QUARUM UTRAQUE CURVA CURVITATE CON­
plumbum et sic per plumbum ita humefactum poliri. Unde primo potest SIMILI POSSE PROTRAHI QUE QUANTO AMPLIUS PROTRA-
poliri cum grosso pulvere emerilli, et deinde cum subtiliori qui auferat 5 HUNTUR AMPLIUS APPROPINQUANT NEC TAMEN CONCUR­
r radios per pulverem grossum causatos. Ad idem / valet emerillus usitatus RENT ETIAM SI IN INFINITUM FUERINT SIC PROTRACTE.
molitus super lapidem, quem utuntur aurifabri et vocant idem pocheam Verbi gratia, capiatur piramis rotunda orthogonalis, quam piramidem
90 cuius libra ut communiter valet trium solidorum. Est etiam alia quedam secet in duo equalia triangulus abc [Fig. 4.13] transiens per caput pira-
pochea que vocatur colcur qua utuntur pararii et cristallarii cuius libra midis et per centrum sue basis, quod sit d, prout probat prima huius,
ut communiter valet viginti solidorum, licet optime poliat cum ligno aut io posito quod dicta piramis iaceat super superficiem orizontis sic quod caput
super laminam ex plumbo et stagno factam. Item forbitores gladiorum eius sit versus boream et eius basis versus austrum, et iaceat sic quod
ad forbiendum habent duplex genus lapidum quorum unus est altero sub- trianguli abc secantis eam in duo media latus ab iaceat super orizontem.
95 tilior. Et primo liniunt superficiem forbiendam cum sanguine et forbiunt Erit igitur dictus triangulus secundum se totum in circulo meridiano.
eam cum lapide grossiori et frequenter sanguinem sup[per]ponunt. Deinde Ymaginor igitur aliam superficiem planam secantem piramidem predictam
cum lapide subtiliori et sanguine forbiunt sicut prius quousque ydola ali­ 15 equedistanter triangulo abc. Et sit communis sectio curve superficiei pira-
quantulum appareant. Deinde tergunt totam illam superficiem et auferunt72
101 appareant OF apparuerint materie V
72 depuratur OF distellatur C maleatur V 102-104 Licet . . . polietur om. V
74 sucus F
102-103 aut . . . indurari F debeat O
75 suco F
104 polietur F pollietur O
78 opportet F 104-109 Sic. . . . ista OF Sequitur conclusio mirabilis prima facie que quanvis
79 incidit F incidet O inscidit C scindit V propositio non pertineat tamen ex sectione muchefi videtur sequi V Sequitur
80 nostro fiendam CV tr. OF conclusio mirabilis prima facie que licet propositio non pertineat (inserviat ed)
83 correntibus V, (?)C recentibus F, (?)0 / debet pulvericari CV tr. OF I eneo CV ipsa tamen ex (e ed) sectione addita (hyperbole ed) sequi videtur Red
ereo OF
107 addita R mukesi O mukefi F muchefi V hyperbole ed
86-87 potest poliri V, (?)C tr. OF
87 et CV om. OF Undecima conclusio
88 idem CV illud OF 1 [Undecima] conclusi o scripsi; om. OV Conclusio mg. F 1la mg. R PROPOSITIO
89 quem V quibus OF / idem V illud OF / pocheam (?)C pochee OF pothe V XI. ed
91 colcur OF colur V / cristalarii O 2 est OF om. RVed
92 licet OF hoc V / poliat F polliat O polit V / ligno OF ligno mundo V / aut F ut 4 amplius OF plus RVed
O ac V 5 nec tamen OF et tamen nunquam RVed
93 laminam FV lamina O 6 in om. O
94 unus V unum OF I orthogonalis RV ortogonalis OF orthogona ed
94-95 altero subtilior F alio subtilior O subtilior alio CV 10 super om. O
97-100 quousque. . . . prius repet. O II borream F I austrum FRVed austram O
97-98 aliquantulum F aliquantum O manifeste V 14 igitur OF enim unam V ergo unam R
98 appareant FV appareat O 15 communis FRed comunis V tertia O
138 ARCHIMEDES IN THE MIDDLE AGES SPECULI ALM UKEFI COMPOSITIO 139

midis et illius superficiei plane sic equidistantis triangulo linea eghf ita illarum duarum superficierum linea recta mrt. Erit igitur linea mn equi­
quod e sit eius punctus propinquior capiti piramidis. Erit igitur linea eghf distans linee ac et est in eadem superficie plana cum linea curva eghf,
necessario curva per secundam huius cum non transeat per caput pira­ 50 que quidem linee, scilicet mn et eghf, licet in infinitum protrahantur ex
midis. Que si in infinitum protrahatur ex parte/ super piramidem in infini- parte basis piramidis nunquam concurrent et tamen plus et plus appropin­
20 tum elongatam ex parte basis, et ac latus trianguli cum elongatione pira­ quabunt ex parte basis.
midis in infinitum elongetur, tunc dico quod linea curva e f et linea recta Probatio. Et probo primo quod nunquam concurrent quia superficiem
ac in infinitum protracte continue plus appropinquabunt et plus et nun­ planam superpositam lateri ac trianguli abc nullus tanget punctus piramidis
quam tamen concurrent. Et quod nunquam concurrent patet quia insunt 55 nisi qui fuerit in linea recta ac. Sed in illa superficie est linea recta mn
in duabus superficiebus equidistantibus planis ex ypotesi. Igitur semper cuius nullus punctus tanget lineam ac. Igitur nullus punctus linee mn
25 distabunt ad minus per tantum quanta est perpendicularis protracta ab tanget aliquem punctum piramidis, nec econtra. Sed linea curva eghf est
una superficie ad aliam. Et quod continue plus appropinquabunt et plus in superficie curva piramidis. Igitur linea eghf nunquam concurret cum
probatur, nam in linea curva eghf signo duo puncta g, h, per que per- linea mn, quod fuit primo probandum.
transeant duo circuli secantes piramidem equidistanter basi. Et ubi dicti 60 Modo probo quod continue plus et plus appropinquabunt et hoc ex
circuli secant lineam ac signo puncta/: et /. Est igitur linea recta protracta parte basis. Nam quibuscumque duobus punctis in linea eghf signatis,
53v 30 inter h et / brevior linea recta protracta inter g et / k, quo probato liquet ut o, p, ita quod o sit propinquior cono piramidis, si per illa duo puncta
propositum. Hoc autem probo sic. transeant duo circuli dividentes piramidem equidistanter basi, qui quidem
Corda dupli arcus portionis circularis intercepte inter puncta h et / est circuli tangant lineam ac ulterius protractam in duobus punctis q et r,
equalis corde dupli arcus portionis circularis intercepte inter puncta g 65 erit sinus rectus portionis oq equalis sinui recto portionis pr propter hoc
et k cum utraque corda sit dupla ad perpendicularem interceptam inter quod duo puncta o et p sunt in superficie plana equidistante triangulo
35 superficiem trianguli abc et superficiem sibi equidistantem transeuntem abc, in cuius latere ac sunt duo puncta q et r. Igitur cum pr sit portio
per puncta g, h, f. Sed circuli illi sunt inequales quia hl est maior, gk maioris circuli quam portio oq et per consequens minoris curvitatis, patet
minor, et quia minor ideo curvior. Ergo sinus versus portionis gk est quod sinus versus portionis pr sit minor sinu verso portionis oq. Sed duo
maior sinu verso portionis hl et sinus recti eorum (!) equales quia medie­ 70 puncta o, p distant a linea mn solummodo per quantitatem sinuum verso­
tates cordarum equalium. Igitur per dulcarnon linea recta inter g et k est rum dictarum portionum. Igitur punctus p minus distat a linea mn quam
40 longior quam linea recta inter h et /, quod fuit probandum. punctus o, quod fuit probandum. Patet igitur conclusio de duabus lineis:
Sed potest dici quod hoc non est mirabile eo quod ille due que sunt quarum una est recta et alia curva.
ac et eghf non sunt in eadem superficie plana. Ideo probabo etiam con­ Quodsi ymaginemur aliam piramidem similem et equalem priori iacere
clusionem de duabus lineis existentibus in eadem plana superficie. Nam 75 super priorem piramidem sic quod linea ac sit communis utrique et yma­
capio eandem superficiem de qua prius fiebat mentio que transit per puncta ginemur superficiem transeuntem per e, g, h , f equidistanter triangulo abc
45 e, g, h, f equidistanter triangulo abc et extendatur illa superficies extra extendi ulterius quousque secuerit superiorem piramidem eodem modo
piramidem quousque orthogonaliter secuerit unam aliam superficiem quo secat inferiorem, tunc sectio communis dicte superficiei et curve
planam iacentem super latus ac trianguli abc, et sit communis sectio superficiei superioris piramidis erit linea curva consimilis curvature cum
80 linea curva eghf, que due linee curve continue plus et plus appropin­
16 equidistans O / linea RVed om. OF / ita RV et OF ut ed quant eo quod utraque plus et plus continue appropinquabit ad lineam
22 plus1 tr. RV post appropinquabunt
22 et2 RVed om. OF 49 in om. O
23 concurrent COF concurrant RVed / patent F / insunt OF sunt RVed 50 protrahantur (?)C, V protendantur OF
27 probatur RVed probacio O probatio F 51-52 piramidis . . . basis CFV om. O
29 et om. F / protracta V om. OFRed 53 probo C probatur OFV
30 recta RVed om. OF / probato correxi ex protracto in OF, (?)V 60 probo V probacio O probatio F patet R et tr. OFR post basis; om. C hic et non
36 hi OFV transiens per hl Red possum eum legere post basis
38 maior ORVed minor F 66 equidistante corr. ex equedistante in CV et equidistanter in OF
38-39 medietates om. F 68 maioris RVed minor OF
39 dulcarnon OF penultimam primi RVed 73 quarum una est curva et alia recta V quarum alia (una ed) est recta alia curva
42-43 Ideo . . . superficie C (cf. Red) om. OFV Red recta et curva OF
44 eandem CRV om. OF / fiebat mentio OF scilicet RVed / que: quando F 75 utrique RVed utriusque OF
46 orthogonaliter RV ortogonaliter OF 77 superiorem RVed priorem OF I eodem RVed eo OF
46-47 superficiem planam tr. F 81 eo quod . . . appropinquabit (appropinquat R) RV (cf. ed) om. OF
140 ARCHIMEDES IN THE MIDDLE AGES
THE COMPOSITION OF THE PARABOLIC MIRROR 141

intermediam, que est mn, et tamen nunquam concurrent eo quod neutra

I first turned myself to the first of them. And when I had studied for some
concurret cum mn linea intermedia. Sic igitur patet veritas conclusionis
time the truth of it that required proof, proofs occurred to me by which
proposite quantum ad utramque eius partem. Et nota quod ad probationem
I demonstrated that the said conclusion was absolutely false, at least
85 huius conclusionis assumpta est sectio mukefi (! dei.) addita sicut satis
according to the understanding which I had of it, and I therefore thor­
liquet sedule intuenti. Et licet de istis consectionibus piramidis possent
oughly disdained and rejected this conclusion as false. But after exten­
plures conclusiones mirabiles educi, quia tamen de facili possem incidere
sively considering that it was improbable that the said Apollonius and the
54r in conclusiones quas Apollonius magnus in libro suo de co/nicis figuris
others citing him had gone so far astray [as would be evidenced] in the
inseruit, cum nunquam viderim eius librum quanquam exactissimam dili-
falsity of the said conclusion— particularly since Campanus cites the said
90 gentiam apposuerim eum vidisse, ideo presenti opusculo cum laude dei
Apollonius in his comment to Proposition 1 of Book XIV of the Elements
predicta sufficiunt.
of Euclid and in the same place speaks of him as “ the Great Apollonius”
— I took as a fixed supposition that there was another way to understand
the conclusion. Standing fast with great zeal in this matter of another
The Composition of the Parabolic Mirror understanding, I finally arrived after great tedium and labor at the one,
true understanding of it. And, with the wisdom of God flowing into me,
Proem its proof occurred to me easily enough. With this conclusion proved and
its proof set down in writing, I did not delay in considering the proof
Because each and every one of all those I have judged to have treated of the other conclusion of Apollonius until, with God’s disposition, I had
of mirrors that burn [something] at a given distance, or [at least] of those found it. With the proofs of the said two conclusions known and written
whose tracts I have seen in writing, assumed two conclusions adduced down and in order that such solicitous labor of mine would not be use­
by Apollonius of Perga, upon which conclusions their whole intention was lessly lost, or might be completely lost, I propose by means of these and
founded, inasmuch as the author, Witelo himself, in his Perspectiva other conclusions to be proved below to teach how to make a burning
presents an initial book [Bk. I] in which he strives to demonstrate all mirror that would burn at any distance, which distance can be half the
the conclusions which seemed to him necessary and prefatory to that radius of any circle, some segment of which— however great or small
which he had to demonstrate so that his whole book would be sufficient in size— could be drawn by an instrument. And this will be the principal
unto itself and yet the said Witelo has assumed the two conclusions intention of this tract. For this intention to be achieved adequately and
of Apollonius without proof and uses them as if known, although the to be proved, I shall premise certain definitions and certain conclusions.
knowledge of the truth of them has escaped me for a long while, and not Therefore, the first definition is of a right or erect cone and is that which
withstanding that I have spent much zeal over a great period with brothers Euclid posits in Book XI of his Elements. It follows.
from distant parts and others in order that I might by their mediation
come upon the said book of Apollonius or that by frequent comparison First Definition
of the said conclusions I might learn a proof from some one to which
[1] A cone is the [figure produced by the] transit of a right triangle
(or to one of which) I could never have come by any effort, and finally
when one of its sides containing the right angle is held fixed and the
therefore greatly moved to establish the proof of the said [conclusions]83456
triangle is rotated until it returns to the place from which it began to be moved.
83 mn: n F / intermedia correxi ex media in Red et intercepta in OFV If the fixed side is equal to the rotated side, it will be a right-angled cone;
83-90 Sic. . . . vidisse OF om. Red et cf. V inferius in lin. 86-90 if it is longer, it will be an acute-angled cone; and if it is shorter, it will
84 utramque RV utramlibet OF be an obtuse-angled cone [see Fig. 4.1]. Moreover the fixed side is the axis
85 mukefi delendum est; cf. meam intro. of this cone and the circle its base.
86-90 Et . . . vidisse om. V et cf. Red superius in 83-90
86 consectiones F et sectiones O
88 conicis correxi ex conitis in O et cognitis in F Second Definition
90 eum F cum O
90-91 ideo . . . sufficiunt OF et hec ad presens dicta breviter sufficiant V hec [2] In every right cone cut by a plane surface not passing through the
breviter sufficant R, et post sufficiunt scr. O et cetera. Explicit et F EXPLICIT apex of the cone the section common to the said cutting surface and to
FELICITER, et post sufficiant scr. V et sic est finis ad laudem omnipotentis the said cone is called “ the section which the curved line contains.”
dei eiusque matris gloriosissime matris virginis marie / predicta O dicta F
[See Fig. 4.2.] The second definition is presented by Apollonius.
142 ARCHIMEDES IN THE MIDDLE AGES THE COMPOSITION OF THE PARABOLIC MIRROR 143

First Presupposition [= Third Definition] conclusion. I was never able to come to this definition of the latus rectum
of a parabola from anything which I found cited by Apollonius but only
[1] When a right-angled right cone is cut by two plane surfaces, one by the circumstance that I demonstrated with great zeal and labor that
of which passes through the apex and the center of the base producing the conclusions of Apollonius were false for any other line of order and
a triangular section (as will be proved below) while the other does not
that they were true for this one alone. Therefore, with these four defini­
pass through the apex but rather cuts the cone by a surface which a tions noted beforehand, [certain] conclusions will follow, of which this
curved line contains (following the preceding definition) so that one of
will be the first.
the said cutting planes stands at right angles to the other, then the straight
line which is the section common to the said two cutting planes by neces­ First Conclusion
sity will either cut one of the two sides of the triangle and will be parallel [1] IN EVERY RIGHT CONE BISECTED BY A PLANE SURFACE
to the other and then it is called a “parabola,” or it will cut one of the PASSING THROUGH THE APEX OF THE CONE AND THE CENTER
two sides of the triangle and meet with the other indefinitely produced. OF ITS BASE THE SECTION COMMON TO THE SAID CONE AND
This can happen in two ways, either in the direction of the apex [of the SAID CUTTING SURFACE WILL BE A RECTILINEAR TRIANGLE
cone] and then the section contained by the curved line will be called a WHOSE BASE WILL BE THE DIAMETER OF THE BASE OF THE
“ hyperbola,” or in the direction of the base (inside or outside [of the CONE SO CUT.
cone]) and then, if it is not a circle, the section is called an “ ellipse.” This conclusion immediately presents itself to the intellect from the
[See Fig. 4.3.] Also the said line which is the section common to the said description of the cone by means of a rotated triangle, as was clear in the
surfaces (of which one is the parabola) is called the “ axis” or the “ arrow” first definition. And thus the truth of that which the third definition sup-
of the parabola [see Fig. 4.4], whose extreme toward the apex of the cone posits is evident. And the sides of the said [axial] triangle are called the
is called the “ vertex of the parabola” and the “ vertex of the curved line.” “ hypotenuses of the cone,” as is evident by the first conclusion of the
And all the lines protracted orthogonally to the said axis [of the parabola] [Book] on the Curved Surfaces.
from one side of the curve to the other will be parallel and are bisected
by the said axis, and the individual lines so protracted are called “ lines Second Conclusion
of order to that axis.” All of these terms are supposed by Apollonius.
And it should be noted that, although we have defined here the three [2] IF A STRAIGHT LINE IS DRAWN BETWEEN ANY TWO
sections which do not pass through the apex of the cone, still we require POINTS LYING IN THE SURFACE OF A CONE AND IT DOES NOT
for our principal objective only the parabola, as will be evident. TOUCH THE APEX OF THE CONE EVEN WHEN PROTRACTED
INDEFINITELY, THEN THE SAID LINE NECESSARILY WILL
PROCEED INSIDE THE CONE.
Second Presupposition [= Fourth Definition] Proof. In cone abc let the said two points be d and e, between which
[2] Although any line of order in a parabola could be called the “ right a straight line de is drawn, which line, regardless of how far it is extended,
base” or “ latus rectum” of the parabola intercepted between the said would not finally pass through the apex of the cone [see Fig. 4.5]. I
line of order and the vertex of the section, it is not so called by Apollonius. say, therefore, that de proceeds inside the cone.
Rather, here is his true definition of the latus rectum of a parabola. With I shall draw from apex a through points d and e two straight lines up
an axis descending from the vertex of the section until it meets the axis to the circumference of the base, which lines [we] let be adb and aec,
of the cone in the direction of the base, the line of order which passes and I shall join c and b by a straight line which will necessarily cut the
through the middle point of the said axis [of the section] is called the base circle, by III.2 of Euclid. Therefore, triangle abc cuts the cone.
“ latus rectum” or “ erect side” of that parabola, however long or short But line de is in that triangle; therefore, line de is inside the cone since
that section is in the direction of the base. Whence it is evident from this the whole triangle is inside the cone. This is also evident when [any]
for every parabola that that section and any segment of it intercepted point / is designated in line de and line af is drawn up to cb, which [we]
between some line of order and the vertex of the section lies between let it meet in point g, for this whole [line] is inside the cone. Hence line
the [two] sides of the curved line, and such individual segments are them­ de is also inside the cone.
selves parabolas and have the same latus rectum or erect side. And that Third Conclusion
latus rectum or erect side will necessarily be double the axis of the parab­
ola, i.e., the axis extended from the vertex of the section until it meets [3] IN EVERY RIGHT CONE CUT BY A PLANE SURFACE PAR­
the axis of the cone, as will be proved (God granting it) in the fourth ALLEL TO THE BASE THE SECTION COMMON TO THE CURVED
144 ARCHIMEDES IN THE MIDDLE AGES THE COMPOSITION OF THE PARABOLIC MIRROR 145

SURFACE OF THE CONE AND TO THE CUTTING SURFACE WILL so that it divides the cone and is parallel to the base of the cone and the
BE A CIRCUMFERENCE WHOSE CENTER WILL BE IN THE AXIS cutting circle’s diameter is AT?/, therefore hg2 = kh ·hi, by III.30 and VI.8
OF THE CONE. of Euclid. Therefore, since oc2 = bo-od (oc being a radius of the base
Proof. Let there be a cone abc [see Fig. 4.6], whose base center is of the cone, a radius erected perpendicular to diameter bd) and hi = od
d and whose axis is ad. And let it be cut by a [plane] surface parallel (the lines being parallel and lying between parallels ho and Id by 1.33 of
to the base, a surface which will thereby form a curved line, say ef, on Euclid), therefore oc2 = hl-ob. And because hg2 = hl-hk and the ratios
the surface of the cone, by that premissed [in the second definition]. of products and their multipliers are the same (as can be deduced from
I say that curved line e f is the circumference of a circle and has its center VII. 16 and VII. 17 of Euclid for numbers and VI. 1 of Euclid for [con­
on the axis of the cone. tinuous] magnitudes), so bo / kh = oc2 / hg2. But also bo / kh = fo / fh
For let two points e and / be designated in [curved] line e f anywhere by VI.2 of Euclid. Therefore fo / fh = oc2 / hg2. And it can be proved
you wish. Let lines aeb and afc be drawn through these points from in the same way concerning any ]two] perpendiculars proceeding from
a , the apex of the cone, to the circumference of the base, and let axis two points on the [parabola’s] axis to the curved side of the parabola
ad be drawn, with the point where it passes through the cutting plane that the axial segment intercepted between the farther perpendicular and
surface designated as point #. Let lines ge, g f and db, dc be drawn. the vertex of the parabola has itself to the axial segment intercepted
Therefore, by VI.2 of Euclid, bd / eg = cd I fg = da / ga. But db —dc, between the nearer perpendicular and the vertex as the square of the
since they are radii of the base. Therefore, ge = gf. [The same thing] farther perpendicular is to the square of the nearer perpendicular. Hence,
will be proved in the same way for any other lines protracted from point since o f I hf = oc2/ hg2from the foregoing and o f = 2 hf from the hypothesis
g to curved line ef. Therefore, by III.9 of Euclid, [curved] line e f is the that h is the midpoint of of, therefore oc2 = 2 hg2. Therefore, by the
circumference of a circle and its center is point#, which was to be proved. corollary to VI. 17 of Euclid, oc l hg = V2/ l ,] an doc = od (these lines being
radii of the same circle, namely the base of the cone). Therefore, od /
Fourth Conclusion hg = v5 / 1. But ad / od = V5 / 1, by VI.8 and the corollary to VI. 17.
Therefore, ad = 2 hg .But ad = 2 of by VI.2 because db = 2 o b . Therefore
[4] IN EVERY PARABOLA THE LATUS RECTUM IS NECES­
ad / of = ad / hg. Therefore hg and of are equal by V.7 of Euclid, and pg = 2
SARILY DOUBLE THE AXIS [THAT EXTENDS FROM THE VER­
hg. Therefore, pg = 2 fo. Therefore, the first part of the proposed
TEX OF THE PARABOLA TO THE AXIS OF THE CONE], AND
conclusion is evident since pg is the latus rectum of the section and fo is its
THAT LATUS RECTUM WILL HAVE ITSELF TO ANY PERPEN­
axis. And from this conclusion it follows that the latus rectum of the section
DICULAR PROCEEDING FROM THE AXIS [OF THE PARABOLA]
is necessarily quadruple the semiaxis intercepted between the midpoint of
TO THE CURVED SIDE OF THE SECTION AS THAT PERPENDICU­
the axis—point h — and the vertex of the section, so that necessarily line
LAR HAS ITSELF TO THE SEGMENT OF THE [PARABOLA S]
AXIS INTERCEPTED BETWEEN THE SAID PERPENDICULAR pg = 4fh.
AND THE VERTEX OF THE PROPOSED SECTION. Now as for the second part of the proposed conclusion that has to be
For example, [construct] cone abd [see Fig. 4.7] cut by a parabola proved, namely that the latus rectum pg / perpendicular mn = perpendicu­
pfc, whose axis fo meets, in point o, the axis of the cone ao, and through lar mn / mf, I first describe a circle that passes through points m and n,
h (the midpoint of the axis [of the parabola]) let there be passed a line divides the cone, and is parallel to the base circle; and let line rms be the
of order phg perpendicular to the axis, which line of order [we] let be diameter of the cutting circle. Therefore, since ms = hi by 1.33, as argued
called the latus rectum of the said section by the fourth definition. Then above, and ms ■mr —mn2 and hl-hk —hg2 (hi being equal to ms) so, as
I say that line pg ~ 2 ax. seg .f o , and that the same pg has itself to per­ above, hk I mr = hf I mf = hg2 / mn2. And, if this is so, therefore, by
pendicular mn proceeding from the axis to the curved side of the section V.15 of Euclid,/?# / mf = pg2 / mn2, pg being 4 hf as demonstrated above,
as that perpendicular mn has itself to axial segment m f intercepted be­ and pg2 being 4 hg2 by the corollary to VI. 17 of Euclid. Since, therefore,
tween the said perpendicular and the vertex of the section. by the same corollary, pg2 / mn2 = (pg / mn)2, so pg / mf = (p# / mn)2,
Proof. Let triangle abd bisect the proposed cone by passing through
its apex and the center of its [base] so that that triangle stands at right
' "oc / hg = v5 / 1" translates the expression “ linea oc se habet ad hg in medietate
angles to [the plane of] the parabola, as the third definition wishes. There­ duple proportionis. " For such usage of “ half a double” in connection with ratios, see M.
fore, since I propose that line gh necessarily will stand perpendicular Clagett, Nicole Oresme and the Medieval Geometry of Qualities and Motions (Madison,
to the said triangle and I imagine a circle passing through points g and h Wise. 1968), p. 474.
146 ARCHIMEDES IN THE MIDDLE AGES THE COMPOSITION OF THE PARABOLIC M IRROR 147

by the tenth definition of Book V of Euclid.2 Therefore, by the same TANGENT TO THE SAID SECTION IN SOME EXTRINSIC POINT
reasoning, line pg, which is the latus rectum of the section, is to line mn OF ITS CURVED SIDE IS EXTENDED UNTIL IT MEETS THE AXIS
as line mn is to line mf, which was the second thing to be proved. There­ EXTENDED IN THE DIRECTION OF THE VERTEX, IF TWO LINES
fore the conclusion is evident in both of its parts. The second part was DESCEND FROM THE POINT OF TANGENCY TO THE A X I S -
the conclusion of Apollonius whose proof I considered over such a long ONE LINE BEING PERPENDICULAR TO THE AXIS AND THE
time. Therefore, I shall follow up by considering the proof of his other OTHER PERPENDICULAR TO THE TANGENT: THEN THE RATIO
conclusion; but first I shall advance other prefatory conclusions as need be. OF THE LATUS RECTUM OF THE SECTION TO THE SEGMENT
OF THE AXIS INTERCEPTED BETWEEN THE EXTREMES OF THE
Fifth Conclusion SAID PERPENDICULARS IS THE SAME AS THE RATIO OF THE
[5] WITH ANY LINE OF ORDER DRAWN IN A PARABOLA BE­ SEGMENT OF THE AXIS INTERCEPTED BETWEEN THE PER­
TWEEN ITS LATUS RECTUM AND ITS VERTEX AND WITH ANY PENDICULAR CLOSER TO THE VERTEX AND THE POINT OF
CIRCLE DESCRIBED THAT PASSES THROUGH THE VERTEX JUNCTURE OF THE TANGENT AND THE AXIS TO THE SEGMENT
OF THE SECTION AND THE EXTREMITIES OF THE SAID LINE OF THE AXIS INTERCEPTED BETWEEN THAT SAME PERPEN­
OF ORDER, THE CENTER OF THE CIRCLE SO DESCRIBED WILL DICULAR AND THE VERTEX OF THE SECTION.
NECESSARILY BE IN THE AXIS [OF THE PARABOLA], AND IN Proof. Let there be a parabola abc [see Fig. 4.8], on whose axis af
ADDITION IN REGARD TO THE DIAMETER THAT PASSES [we] let descend perpendicular de, and let straight line dg be tangent
THROUGH THE VERTEX OF THE SECTION AND THE CENTER to the parabola in point d, which line dg [we] let meet the axis further
OF THE SAID CIRCLE, ITS SEGMENT THAT IS INTERCEPTED extended to point g , and from point d we let a perpendicular to line dg
BETWEEN THE SAID LINE OF ORDER AND THE CIRCUMFER­ go out until it meets the axis in point h. Therefore, since Agdh is a right
ENCE OF THE CIRCLE IN THE DIRECTION OF THE BASE OF angle from which line de descends perpendicularly upon the base of tri­
THE CONE WILL NECESSARILY EQUAL THE LATUS RECTUM angle gdh, therefore he / ed = ed / e g , by VI.8 of Euclid. But, by the
OF THE SECTION. fourth conclusion, the latus rectum of the section has itself to ed as ed
Proof. For it is evident that the circle can be drawn through the said is to ea. Therefore, the square of de results as well from the product
three points by the doctrine of IV.5 of Euclid. And that its center will of the latus rectum of the section and ea as from the product of he
necessarily lie in the axis can be made evident from the corollary to III. 1 and eg. Therefore, by VI. 13 of Euclid, the ratio of the latus rectum of
of Euclid. And that the segment of the diameter of the circle so drawn the section to the axial segment he intercepted between the extremes e
that is intercepted between the vertex and the perpendicular, which per­ and h of the two perpendiculars de and he is the same as the ratio of
pendicular is half of the protracted line of order, has itself to that per­ eg to ea, i.e. as the ratio of the axial segment intercepted between per­
pendicular as that perpendicular is to the rest of the diameter is evident pendicular ed and the point of juncture of the tangent and the axis to the
by VI.8 of Euclid. But by the preceding conclusion the segment of the axial segment intercepted between the same perpendicular de and the
diameter intercepted between the vertex and the same perpendicular has vertex of the section, which was to be proved.
itself to that perpendicular as that perpendicular is to the latus rectum of I have inserted these last two conclusions not as ones necessary to our
the parabola. Therefore, the latus rectum and the aforesaid segment of the proposition [which follows] but as ones that seemed notable. Now it is
diameter intercepted between that perpendicular and the circumference convenient to proceed to the proof of the second conclusion of Apol­
of the circle in the direction of the base of the cone are necessarily equal, lonius, which is this.
which was to be proved. Therefore, let this conclusion be the first one
prefatory to the proposed conclusion [of Apollonius].
Seventh Conclusion
Sixth Conclusion
[7] IF A STRAIGHT LINE IS TANGENT TO ANY PARABOLA,
[6] WHEN IN A PARABOLA ITS AXIS IS EXTENDED IN BOTH AND IF THE TANGENT IS DRAWN AS FAR AS WE LIKE IN THE
DIRECTIONS BY AS MUCH AS WE LIKE AND A STRAIGHT LINE DIRECTION OF THE VERTEX UNTIL IT MEETS THE AXIS EX­
TENDED FARTHER BY ANY AMOUNT, AND IF FROM THE POINT
2 "pg2 / mn2 = (pg / mn)2 translates the expression “ proportio quadrati pg ad quad­ OF TANGENCY A PERPENDICULAR DESCENDS TO THE AXIS,
ratum mn sit dupla ad proportionem pg ad lineam mn." In medieval tracts on proportion,
the authors continually say that one ratio is “ double” another when they mean that the
THEN THE AXIAL SEGMENT INTERCEPTED BETWEEN THE
ratio is “ taken twice," that is to say “ squared.” Compare the preceding footnote and my SAID PERPENDICULAR AND THE VERTEX OF THE SECTION
Archimedes in the Middle Ages, Vol. 2, p. 464, 24rH (“dupla proportio” ). WILL BE NECESSARILY EQUAL TO THE LINE INTERCEPTED
148 ARCHIMEDES IN THE MIDDLE AGES THE COMPOSITION OF THE PARABOLIC MIRROR 149

BETWEEN THAT SAME VERTEX AND THE POINT OF JUNCTURE greater than its part), therefore pg > ne. But the conclusion is false since
OF THE TANGENT AND THE AXIS THUS EXTENDED FARTHER. they were posited equal by hypothesis and no magnitude can be equal
For example, as in the present figure [see again Fig. 4.8], let there to another and also greater than it. But this follows if line gd intersects
be a parabola abc', and let line dg be tangent to the parabola in point d the parabola between points d and c. Therefore it is false that gd inter­
and extended until it meets in point g the axis fa that has been farther sects the parabola between d and c. And it was demonstrated earlier that
extended. Then I say that lines ea and ag are completely equal, since it does not intersect it between d and a . Therefore it will intersect nowhere;
its converse is true, namely, if ag and ae are equal, line gd passing through therefore, of necessity, it will be tangent [to it] in point d if ae and ag
point d will be tangent to the section in point d. are equal. And, if so, the converse [is true]: if gd is tangent to the section
Proof. For, if not, it will therefore cut the section, by the second con­ in point d, then ae and ag will be equal, which was to be proved. There­
clusion; and, if so, then [it will intersect it] between d and a or between fore evident is the proof of this conclusion, which was the conclusion
d and the base of the section. But that [its intersection is] not between of Apollonius adduced by Witelo and other perspectivists as a principle
d and a is proved; for, if so, let the intersection be in point k, from which but proved by none of them. Also evident is the proof of the fourth con­
we let line kl descend perpendicularly to the axis. Because, therefore, clusion, whose second part was similarly a conclusion of the same Apol­
ga and ae are posited equal, so ga will be greater than al. Let it be greater lonius but not proved by those adducing it. And therefore I have studied
by line gm, and gm and el will be equal, and al and am are equal. Conse­ the proof of these conclusions for a great while and finally have discovered
quently ml = 2 al. But, as was evident from the proof of the fourth con­ it with the help of God, to Whom let there be honor and glory through
clusion, ae / la = (ed / Ik)2, and ed / Ik = eg / Ig, by VI.4 of Euclid.1 the ages.
Therefore ea / la = (eg / Ig)2. But ea I la - 2 ea / 2 la, by V.15 of
Euclid. Therefore, eg / Im = (eg / Ig)2. And, if thus, therefore, by the Eighth Conclusion
tenth definition of [Book] V of Euclid, eg / Ig = Ig / Im. Since whole
eg / whole Ig = Ig I Im (Ig being a part of the whole eg and Im a part of the [8] WHEN IN EVERY PARABOLA A POINT IS DESIGNATED
whole Ig), therefore, by V.19 of Euclid, el / mg = eg I Ig (el being the IN THE MIDDLE OF ITS AXIS, WHICH POINT WILL ALSO BE IN
remainder of the whole eg and mg the remainder of the whole Ig). But THE MIDDLE OF ITS LATUS RECTUM, AND THEN FROM ANY
eg > /g , because a whole is greater than its part. Therefore, el > m g . POINT OF ITS CURVED SIDE A STRAIGHT LINE PARALLEL TO
Consequently, if equals are added to them, i.e. al to le and am to mg, THE AXIS DESCENDS TOWARD THE BASE, IF FROM THAT SAME
then by Euclid’s axiom ag < a e . But they were posited equal. Therefore POINT EXISTING IN THE CURVED SIDE TWO STRAIGHT LINES
we have a magnitude that is both less than and equal to the same magni­ PROCEED— ONE EXTENDED IN BOTH DIRECTIONS THAT IS
tude, which is impossible. But this follows if gd intersects the parabola TANGENT TO THE SECTION AND THE OTHER PRODUCED UN­
between a and d. Therefore this [last assumption] is false. TIL IT MEETS THE AXIS IN THE POINT PREVIOUSLY DESIG­
But if it is said that gd will intersect the parabola not between d and NATED, TWO EQUAL ANGLES WILL BE FORMED, OF WHICH
a but between d and c , say in point o , then let line on descend perpendicu­ ONE IS FORMED BY THE TANGENT IN THE DIRECTION OF THE
larly to axis an, and let ag and ae be posited equal as before, and the VERTEX OF THE SECTION AND BY THE LINE PRODUCED TO
amount that n is distant from e in the direction of the base let p be the THE DESIGNATED POINT AND THE OTHER IS FORMED BY THE
same amount distant from g above g. Therefore, na = p a , and conse­ LINE DRAWN PARALLEL TO THE AXIS AND THE TANGENT IN
quently np = 2 na and eg - 2 ea. Therefore let it be argued that, as it is THE DIRECTION OF THE BASE.
clear from the fourth conclusion of this [work] and its proof that line For example, let there be in parabola abc [see Fig. 4.9] axis ak ex­
na / line ea = (line no / line ed)2, therefore na / ea = (line ng / line eg)2, tended until it meets with the axis of the cone [in k]. Let the midpoint
by VI.4 of Euclid. But Am lea = 2na / 2 eg, by V.15 of Euclid. Therefore, of the axis be e\ and therefore point e will also be the midpoint of the
np / eg = (ng / eg)2. Therefore, by the tenth definition of [Book] V of latus rectum of the section, as is evident from the fourth definition of this
Euclid, np I ng = ng I eg. Since the whole np / whole ng = ng / eg (ng [work]. Therefore let point d be designated in the curved side of the
being a part of the whole np and eg a part of the whole ng), therefore, section. Let line dl be drawn from this point parallel to axis ak, and let
by V. 19 of Euclid, pg / ne = np / ng (pg being the remainder of the whole line gdh be tangent to the said section in this same point d, and let line
np and ne the remainder of the whole ng). Since np > ng (a whole being de be drawn. Then I say that Cedg = Aldh.
Proof. Adeg is acute, right or obtuse. For example, let it first be acute,
See above, note 2 of the translation of the Fourth Conclusion.
and from point d let line d f be drawn perpendicular to the axis, which
150 ARCHIMEDES IN THE MIDDLE AGES THE COMPOSITION OF THE PARABOLIC MIRROR 151

perpendicular falls between points a and e by 1.32 of Euclid. And let ea if so, d e 2 = eg2 and consequently de = eg, and therefore L e d g — L e g d .
be extended until it meets the tangent in point g , and L f g d will be acute Therefore, L ld h , which is equal to one, will be equal to the other. There­
by 1.32 of Euclid; and, by the preceding [conclusion], ag = af. Because fore, L ld h = L e d g , which was to be proved. Thus the truth of the con­
line ae is divided in point/ and there is added to it ag, which is equal to clusion is evident whether L d e g is acute, right or obtuse.
one of the segments, so by II.8 of Euclid (1) eg 2 = 4 ea-ag + f e 2 or eg2
= A e a - a f + f e 2 (since ag = af). But line ea is one fourth part of the latus Ninth Conclusion
rectum of the parabola by the first part of the fourth [conclusion] of this
[9] IN A MIRROR HOLLOWED OUT WITH PARABOLIC CON­
[work]. Therefore, by VI. 1 and II. 1 of Euclid, 4 af-ae = α/Tatus rectum
CAVITY ALL SOLAR RAYS PARALLEL TO THE AXIS THAT FALL
= af-bc (since the latus rectum is line be, by the fourth definition of
ON THE MIRROR WILL BE REFLECTED TO ONE AND THE SAME
this [work]). Therefore, (2) af-bc + f e 2 — eg2. But, by the second part
POINT, WHICH POINT WILL BE THE MIDDLE OF THE AXIS AS
of [conclusion] 4 of this [work], it is evident that (3) a f - latus rectum
WELL AS THE MIDDLE OF THE LATUS RECTUM.
= d f 2 [or af-bc = d f 2], since d f is perpendicular to the axis. And (4) d f 2
This conclusion is immediately evident to the eye from the preceding
+ f e 2 = d e2, by the penultimate of [Book] I [of Euclid]. Therefore [by
[conclusion] and from the following principle of perspective: the angle
equations (4), (3) and (2) together] e g 2 = ed2. Therefore line ed = line eg.
of incidence and the angle of reflection are necessarily equal.
Therefore, by 1.5 of Euclid, L e d g in triangle edg is equal to L e g d . But
With these nine preceding conclusions [proved] I shall proceed further
lines dl and ae are parallel by hypothesis. Therefore, by 1.29 of Euclid,
to the principal proposal to be sought. For this there is required the
extrinsic L ldh is equal to intrinsic L eg d . Therefore, L e d g = L ld h , which
fabrication of a certain instrument by which steel or some such hard
was to be proved. The [same] will be proved in the same way for any
substance can be hollowed out in [the form of] a parabolic concavity.
line that is parallel to the axis [and is reflected by the section such that
Let the tenth conclusion of this tract (which follows) be the one that
its reflected line is] joined with point e so that e is an acute angle.
outlines the doctrine of fabricating an instrument of this sort.
But if /-deg is a right angle [see Fig. 4.10], the proposition will still
be proved. Because /-e is a right angle, therefore line de is perpendicular
Tenth Conclusion
to the axis. Therefore, by the preceding [conclusion], line ag = line ae.
But line ae, as before, is !4 part of the latus rectum of the section, by the [10] TO MAKE AN INSTRUMENT BY WHICH THE SOLID MA­
first part of the fourth [conclusion] of this [work]. Therefore, line ge, TERIAL OF A BURNING MIRROR CAN BE HOLLOWED OUT IN
which is double line ae by the preceding [conclusion], is Vi the latus rectum [THE FORM OF] A PARABOLIC CONCAVITY.
of the section. But ed is Vi the latus rectum of the section since e is its It is necessary, for this instrument, first to consider at what distance
midpoint, by the fourth definition of this [work] and by hypothesis. There­ you wish the mirror that is to be constructed to burn, as for example
fore, lines de and ge are equal. Therefore, by 1.5 of Euclid, A edg = L eg d . at 20 feet. And then it will be necessary for you to have a staff or rod
Hence, by 1.29 of Euclid, L ldh = L e g d . Consequently L e d g = L ld h , that is twice as long, namely 40 feet, with which as the radius of a circle
which was to be proved. the segment of a circle can be drawn.1 Then take two plane quadrangular
But if Ldeg is obtuse [see Fig. 4.11], let the perpendicular line d f be rules of copper or iron which are just as long as the width of the mirror,
drawn. Therefore, point/ will be between points e and k. Therefore, line and let the width of each rule be such that the segment of a circle with
a f > ae. But, by the preceding [conclusion], a f = ag. Therefore, ag > ae. the above mentioned radius can be drawn in it from one extreme to the
Therefore, by 1.3 of Euclid, let line am be made equal to ae. Therefore, other. In one of the rules let the segment abc be drawn with the said
line gm remains equal to line/<?; and consequently, with line em added radius [see Fig. 4.12]. And in the second rule let segment d e f be drawn
to each, line fm will equal line ge. Therefore,/ra2 = g e 2. Since line fa with a shorter radius than before, for example one a foot shorter. Then
is divided in point e and line am (equal to one of the segments, namely, from each rule let everything which is beyond the curvatures of the said
to line ae) has been added to it [i.e. to fa], therefore, by II.8 of Euclid, segments protracted in the rules be removed with a file. Then let there
it is evident that 4 af-am + f e 2 = 4 af-ae + f e 2 = f m 2 = eg2. But, as is
evident from the preceding [conclusion],fa -4ae = fa latus rectum. There­ Tenth Conclusion
fore, /« Tatus rectum + f e 2 = f m 2. But by the second part of the fourth
1 See below, Chap. 5, Text C, bracketed addition to Conclusion 10, for Regiomontanus'
[conclusion] of this [work], fa Tatus rectum = f d 2. Therefore, f d 2 + f e 2 explicit calculation of the focal distance (from the vertex of the parabolic section) as a
= f m 2. Since, by dulcarnon [“ The Possessor of two horns,” i.e., the little more than 27 feet when the radius of the base circle of the segment of the cone is
Pythagorean theorem], de2 —d f 2 4- f e 2, therefore de2 = fm 2 = eg2. And 40 feet.
152 ARCHIMEDES IN THE MIDDLE AGES THE COMPOSITION OF THE PARABOLIC MIRROR 153

lie upon the surface of the horizon an oblong, quadrangular and rectangu­ foot by hypothesis and since line ob is also one foot because point e is
lar body of six parallel bases and parallel edges whose length is as long distant from line gh by one foot by hypothesis, therefore in triangle eob
as the length of rule abc. Let it be, for example, eight feet long, three the two sides eo and ob are equal and angle o is a right angle. Therefore,
feet broad and one foot high. Let the upper face of it be ghdf and let it each of the two angles obe and beo is one half a right angle. Therefore,
lie on base ikac. Then I draw on its upper face [ghdf] a line parallel since line oe is parallel to the axis of the cone, if line be were to be extended
to gh at one foot distance from gh, evidently distant from it by the same until it would meet with the axis, the angle beo would be equal to the
amount that the radius of circle def is less than the radius of circle abc. angle formed by the meeting of line be with the axis, namely, one half
And let the midpoint of that line be e. Then let [rule] def be adjusted a right angle. And consequently the whole angle at the apex of the cone
on the upper face of the oblong body so that the midpoint of the curvature would be a right angle. Therefore the cone is a right-angled cone by the
of rule def touches in point e the line [through] e parallel to gh. And, by first definition of this [work] and the curved surface abcdef is a segment
using the rule def so curved, let a curved line be drawn on the upper face of the curved surface of a right-angled cone, which has to be proved.
of the oblong body, which curve in that face [we] let be called def. Then With these things accepted let point n be designated in line bo extended
across the upper face of the said oblong body I shall draw a straight line further in the direction of o, which point n we let be as distant from o
passing through point e, which line will also be equidistant to the extremes as o is distant from b , so that o is the midpoint between b and n . Then
of the body since the midpoint is equally distant from the extremes. Then let a line be drawn on the base ikac that passes through n and is parallel
from the ends of the middle line [through e] let two other lines be drawn to ibk. Through this line and point e let there pass a plane surface cutting
across the side faces of the body and parallel to its extremes until they out whatever is outside of that surface in the direction of b. Then the
meet the base on which the body lies. And between the lower extremes section common to that plane surface and the curved surface abcdef is a
of these two lines so descending let a straight line be drawn across the parabolic section. For if we imagine a plane triangular surface passing
base, which line will also be equidistant from the extremes of the said base. through the apex and the center of the base of the cone as well as through
For example, where the said line in the base touches line ik let it be designated line bon and containing sides in the curved surface of the cone and bisecting
point b, and let the transverse line of the base be called bm. Let the the cone, then line en will be parallel to one side of the said triangle,
base be turned up and lie on the face ghfd opposite the base. Then let for with line bn extended to the side of the triangle in the direction of
rule abc be applied to that face ikac so that the midpoint of the curve n that line will form with the said side (in the direction of the apex) an
of the rule abc touches line ik in point b\ and, by using the rule abc so
angle equal to one half a right angle because that line [bn] will be parallel
curved, let a curve be drawn on face ikac, which curve in that face [we]
to the diameter of the base of the cone. But Aeno is one half a right
let be called abc. Then two curves have been produced in opposite faces,
angle. Therefore en and the said side are parallel by 1.28 of Euclid. There­
which curves are abc and def, segments of two circles, of which the center
fore, we thus have a plane surface disposed in the sides according to the
of the upper one is directly supraposed to the center of the lower one
form of a parabolic section. To this section let us therefore apply a plate
so that a line drawn between the centers is perpendicular to each circle.
of steel and let it be shaped according to the form of the said [parabolic]
By an instrument whose sharp edge is a straight line let there be shaved
section. Then afterwards, keeping the same shape, we let it be sharpened
off from the side face ghki whatever is outside of the curves abc and
so that its cutting edge is a line curved with a parabolic curvature. Then
def. Therefore, there will be left in that oblong body a curved surface
which is def abc, which surface so curved is a segment of the curved with this instrument another plate of steel or iron may be hollowed out,
then polished so that you will have a burning mirror [that burns something]
surface of a right-angled cone between two parallel circles.
at a distance equal to half the radius2 of the circle whose segment abc
That it is a segment of the curved surface of a cone is clear enough,
was delineated on the first rule singled out above.
for the centers of the two circles whose segments constitute the curved
sides of the proposed surface are in a line perpendicular to each circle. Moreover that it will burn [something] at such a distance is evident
And that the cone is a right-angled cone is proved, for let the oblong body from the ninth conclusion of this [work] and from the axiom accepted by
lie on its prior base so that the curve abc is in the lower face and the all the perspectivists: in every reflection of a solar ray or a ray of some
curve def is in the upper face. Then let us imagine that from point e other luminous source the angle of incidence and the angle of reflection are
a line descends perpendicularly to the base, a line which necessarily falls necessarily equal. And thus is evident the proposal which has been set
on transverse line bm of the base. Let it fall [on it], for example, in point forth in the beginning of this opusculum.
o, and let eo be a perpendicular. Therefore, since line eo contains pre­
cisely one foot because the height of the oblong body was posited as one 2 See the preceding note.
154 ARCHIMEDES IN THE MIDDLE AGES THE COMPOSITION OF THE PARABOLIC MIRROR 155

Conditions of a Good Steel pieces to be consolidated be fired, and when fired let it be roughed by the
But, because it is strongly required that the instrument for hollowing cutting iron so that it may be better incorporated with another piece. Then
out [mirrors] be of steel that is good, choice and hard, therefore note let one piece be placed on the other and both put into the fire, and let
the three conditions of a good steel. The first of these is smoothness of them be fired and well consolidated. Then let them be removed [from the
the exterior surface with fine continuity of the parts, that is, without fire] and hammered together and firmly consolidated. And in the manner
fissures. The second is frangibility and the third is clearness of color at just described for these two pieces, the same thing can be done for two
the place of fraction. Whence the first and the third designate [its] purity or three other pieces. When done, again let these masses be broken up
and the second testifies to its hardness. Let the purging or refining of and again hammered and consolidated, and this procedure can be repeated
steel be done by the following method. Take an oblong piece of iron like as many times as need be until the steel is sufficiently purified. Then let
an iron rod that is a little wide, say, one thumb wide. Let pieces of steel all the pieces be reduced to a single mass, which ought to be softened
be broken up together and arranged without any space between them. by fire and carefully hammered and reduced to the intended shape in such
Then let this whole mass be sprinkled with smith’s water mixed with a a way that there is no blister, fissure, or impurity that arises from care­
certain kind of earth of saffron which all the smiths use for consolidating less hammering. Then it ought to be hardened. Concerning this, note that
iron and steel, for without such wetting the pieces of steel could not be there are three methods of hardening steel. The first of these produces the
consolidated in any way with the iron or with each other. Now therefore, hardest and most brittle steel, and this is done by placing it in cold water
the iron rod, to the degree that it has been forged and hammered, is needed after extensive firing to the point that the heat of firing produces a whiteness
for the consolidation together of the pieces of steel and also for the pres­ of the color of the sun between rising and its high point in the heaven. The
ervation of the steel lest some of its pure parts be consumed by the fire second method produces a hard and brittle steel but not so excessively as
(just as silver is preserved in its refining by lead mixed with it). Then let before, and this is done by placing it in cold water after it reddens in the fire
this mass, sprinkled in this way, be placed in the fire (and these things like a ripe cherry; [this is] after it has been allowed to be somewhat heated
are taught and openly prepared in the smith’s shop). The procedure is until there is in it the purple color of iron; and this method is suitable for
this. Let a certain aperture be made in the bottom [of a crucible], in which instruments that cut iron or soft steel or masonry or another kind of hard
aperture iron rods are set crisscross, intersecting each other orthogonally metal, and this method is suitable for making a mirror of steel. However,
like the bars of iron windows, and by means of this aperture the impuri­ when it is taken away from the fire, let the exterior circumferential piece be
ties of the metal can run out below and the pure material can be preserved. wetted with the aforesaid water to an amount of two fingers thus: ® , so that
And after this mass has been well fired so that it is almost white, let it it will not introduce any [un]due concavity to that which has been
be removed [from the fire] and let it be hammered quite strongly so that diminished earlier by hammering. And after its removal from the water let
the pieces of steel might be consolidated. And let it again be placed in only its middle part be heated. But this method of heating produces a
the fire and while it is in the fire let it be frequently sprinkled with sand diversity of colors in the surface of the mirror after its polishing. The third
and gravel to bring about better consolidation. Then again let it be ham­ method of hardening differs from the second in that the final heating has as
mered quite strongly until the consolidation becomes firm and sufficient. its object a purple color like the color of violet in the steel. And this method
However, in [the course of] any removal [of it] from the fire it ought to of heating is suitable for knives and instruments designed to cut woods.
be considered whether some impurity can be found [in it]. This impurity Still another method of hardening is put forth by Albertus Magnus in On
would be recognized by some blackness appearing in the body of the metal Minerals, Bk. II, Tract. 3, Chap. 2 on the cause of images appearing in
below its fired surface. If it is found, let it be cut out by the cutting instru­ stones. He speaks as follows. “ But those things which we have tried, we
ment which smiths use. Thereupon let the mass so consolidated be cut, describe here. For let steel be hardened and purified often until it almost
while it is well fired, into several pieces by means of the same instrument, has the color of silver, and then there are fashioned out of it instruments for
with the parts not completely separated or cut away but adhering some­ cutting with suitable and subtle angles. And then some juice of radish is
what to each other. Then let this mass be hardened by this method. Let squeezed and mixed in equal amount with a water extracted from
it be placed again into the fire until it is red like a ripe cherry, and after­ earthworms that have been crushed and squeezed through a cloth, and
wards let it be submerged in water until the fire is extinguished and the afterwards the glowing instrument is extinguished in that water twice or
heat consumed. Then let it be hammered into pieces at the aforesaid inci­ thrice or as many times as necessary for it to be made hard enough to
sion spots. And by this same method a second mass and a third and a fourth scratch gems and to cut another iron as [if it were] lead.” So much Albert.
(indeed as many as you need for your work) may be completed. Then As for polishing to be done to our mirror, emery is valuable. This is a
let these masses be consolidated two by two in this way. Let one of the stone having an iron-like color like that of a good adamant or having a
 THE COMPOSITION OF THE PARABOLIC MIRROR 157
156 ARCHIMEDES IN THE MIDDLE AGES
circle. Hence, I imagine another plane surface cutting the aforesaid cone
dark citrine color, and it is similar to flints found in running streams.
and parallel to triangle abc. And let line eghf be the common section
And it ought to be pulverized in a bronze mortar and the powder strained
of the curved surface of the cone and that plane surface parallel to the
through a linen cloth or sifted through a sieve tightly woven of hairs.
triangle so that e is its point nearer to the apex of the cone. Therefore
This powder ought to be mixed with water and the whole mixture ought
line eghf will necessarily be curved by the second [conclusion] of this
to be placed on a lead [strip] and thus the polishing [is done] by means
[work] since it does not pass through the apex of the cone. If it (the curve)
of the lead so sprinkled. Whence it can be polished first with a gross emery
is extended indefinitely in the direction of / on the surface of the cone
powder and then with a finer powder to remove the streaks caused by
when the latter is prolonged indefinitely in the direction of the base,
the gross powder. For the same [operation] is valuable the emery used
and side ac of the triangle is prolonged indefinitely with the prolongation
and ground on a stone [slab], which emery the goldsmiths use and call
of the cone, then I say that curved line ef and straight line ac (both
pochea and which is commonly valued at three solidi per pound. There is
protracted indefinitely) continually approach more and more closely to
also a certain other pochea called colour which the jewelers and crystal
[each other] without, however, ever meeting. And that they never meet
workers use and which is commonly valued at twenty solidi per pound,
is evident because they lie in two parallel planes by hypothesis. There­
although it polishes best with a wood [base] or on a sheet made of lead
fore they will always be distant [from each other] at least by the length
and tin. Also furbishers of swords use, for furbishing, two kinds of stones,
of the perpendicular protracted from one surface to the other. And that
of which one is finer than the other. And first they line the surface to be
they continually approach [each other] more and more closely is proved,
furbished with blood and they furbish it with the grosser stone, and they
for in curved line eghf I assign two points g and h, through which we
frequently put [more] blood on it. Then, with the finer stone and with
let be drawn two circles that cut the cone and are parallel to the base.
blood, they furbish it as before until images appear somewhat. Then they
And where the circles cut line ac I assign points k and /. Therefore,
wipe clean the whole surface and remove the blood. Then they put on
the straight line protracted between h and / is shorter than that protracted
the surface powder of lime and they take the iron instrument prepared
between g and k. When this is proved, the proposition will be clear.
for this and furbish as before until images appear and until no vestige of
I prove it as follows.
the preceding stones remains at all. It will be clear, therefore, out of what
The chord of double the arc of the circumferential segment intercepted
has been presented how an instrument ought to be fabricated and how
between points h and / is equal to the chord of double the arc of the circum­
hardened so that with it a burning mirror will be shaped, and, in addition,
ferential segment intercepted between points g and k since each chord
how the said mirror will be polished. So we have, therefore, come to the
is double the perpendicular intercepted between the surface of triangle
end of our principal proposal. But though we have brought our desired
abc and the surface parallel to it that passes through points g, h and/.
proposal to its desired end, because a certain conclusion that seems on
But those circles are unequal because [the circle through] hi is greater
the face of it marvelous follows from the [construction of a] hyperbola,
and [that through] gk is lesser and therefore the lesser is more curved.
therefore we have thought it worthy to add the said conclusion here,
Therefore, the versine of segment gk is greater than the versine of seg­
and it is this.
ment hi, and their right-sines are equal because they are halves of equal
chords. Therefore, by dulcarnon [“ The Possessor of Two Horns” i.e.,
Eleventh Conclusion the Pythagorean theorem], the straight line between g and k is longer
than the straight line between h and /, which was to be proved.
[11] TO BE ABLE TO PROTRACT TWO LINES— ONE STRAIGHT
But it can be said that this is not marvelous because the two lines
AND THE OTHER CURVED OR EACH CURVED WITH SIMILAR
ac and eghf are not in the same plane surface. Therefore, I shall prove
CURVATURE— SUCH THAT THE FARTHER THEY ARE PRO­
TRACTED THE CLOSER THEY APPROACH [EACH OTHER], the conclusion as well for two lines that lie in the same plane surface.
For I take the same surface of which mention was made before which
WITHOUT, HOWEVER, [EVER] MEETING, EVEN IF THEY ARE
passes through points e, g, h, f and is parallel to triangle abc, and let
SO PROTRACTED INDEFINITELY.
that surface be extended outside of the cone until it orthogonally cuts
For example, let a right cone be taken, which cone triangle abc [see
another plane surface that lies on side ac of triangle abc, and let the
Fig. 4.13] bisects by passing through the apex of the cone and the center
common section of these two surfaces be straight line mn. Therefore,
of its base d, as the first [conclusion] of this [work] proves, it having
line mn will be parallel to line ac and is in the same plane surface as
been posited that the said cone lies on the surface of the horizon so that
curved line eghf. Lines mn and eghf, though extended indefinitely in the
its apex is toward the north and its base toward the south, and let it lie
direction of the base of the cone, will never meet but will approach [each
so that side ab of triangle abc (which latter bisects the cone) lies upon
other] more and more closely in the direction of the base.
the horizon. Therefore, the said triangle will be completely in a meridian
158 ARCHIMEDES IN THE MIDDLE AGES
CHAPTER 5
Proof. And I prove in the first place that they will never meet because
no point of the cone will touch the plane surface superimposed on side
ac of triangle abc except the points which lie in straight line ac. But
straight line mn, no point of which touches line ac, is in that surface.
Therefore, no point of line mn will touch any point of the cone, and con­
versely. But curved line eghf is in the curved surface of the cone. There­
fore, line eghf will never meet line mn, which was the first thing to Conic Sections in the Fourteenth and
be proved.
Now I prove that they will continually approach [each other] more
and more closely in the direction of the base. [This is so] for when any
Fifteenth Centuries: Jean Fusoris,
two points o and p are assigned in line eghf so that point o is closer
to the apex of the cone and if through these two points we let pass two Giovanni Fontana, and
circles that divide the cone and are parallel to the base, these two circles
touching line ac (which is further protracted) in the two points q and r,
then the right-sine of segment oq will be equal to the right-sine of segment
Regiom ontanus
pr because the two points o and p are in the plane surface that is parallel
to triangle abc, in side ac of which are the two points q and r. Therefore,
The main directions taken in the treatment of conic sections from the
since pr is the segment of a greater circle than is segment oq and conse­
early fourteenth into the sixteenth century were established by the tracts
quently of lesser curvature, it is evident that the versine of segment pr
I have analysed in the first four chapters. We must first examine a brief,
is less than the versine of segment o q . But the two points o and p are
anonymous work extant in three manuscripts.1It is without title in the two
distant from line mn precisely by the lengths of the versines of the said
earliest manuscripts (both 14c.), but in the third manuscript (17c) it is en­
segments. Therefore point p is less distant from line mn than is point o,
titled De sectione optimi speculi comburentis: vocatur Mukefi. It com­
which was to be proved. Therefore, the conclusion is evident for the two
mences with the observation, as had Alhazen’s De speculis comburentibus,
lines: one of which is a straight line and the other a curved line [in the
that the strongest combustion results from a mirror whose surface is in the
same plane].
form of a sectio mukefi (i.e., a parabola),2 meaning, of course, a para-
But if we imagine another cone similar and equal to the former to lie
on the former cone so that line ac is common to each and we imagine 1 The three manuscripts are Paris, BN lat. 7434, 102r-103v, 14c; Lincoln, Cathedral
that the surface passing through e, g, h, f and parallel to triangle abc Chapter Library MS 238 (A.7.9), 13r-18r, 14c; and Oxford, Bodleian Library, Corpus
is extended farther until it cuts the upper cone in the same way it cuts Christi Collge MS 260, pp. 147-50, 17c (with a diagram of a paraboloidal mirror occupying
page 149). The only complete manuscript is that of Paris. The Lincoln manuscript is nicely
the lower, then the common section of this said surface and the curved written; but, in passing from 15v to 16r, the scribe has omitted almost half of the discussion
surface of the upper cone will be a curved line with curvature similar on preparing the tool for making a burning mirror. The Corpus Christi manuscript has
to that of curved line eghf and the two curved lines continually approach slightly less than half of the first part of the tract which quotes definitions from the fragment
more and more closely [to each other] because each will continually ap­ of the Arabic version of the Conics translated by Gerard of Cremona. Because the Paris
proach more and more closely to the intermediary line mn, and yet they manuscript is the only complete one and because it appears to be the earliest, I shall
quote from it almost exclusively in the succeeding notes.
will never meet [each other] because neither will meet the intermediary 2 Paris, BN lat. 7434, 102r: “ Quoniam in opere speculorum precipuorum comburencium
line mn. So, therefore, the truth of the conclusion of the proposition necesse est habere seccionem mukefi secundum quam debet fieri regula per quam totum
is evident in each of its parts. And note that for the proof of this conclu­ speculum rectificari debet et fieri, ideo nunc ut occurrit propono dicere qualiter possit talis
sion a hyperbola has been assumed, as will be clear enough to any one seccio inveniri. Quamvis enim per seccionem seu porcionem circuli possit fieri speculum
comburens ut communiter fieri consuevit, tamen tale speculum non est forcioris combus­
diligently considering it. And though concerning these conic sections
tionis que possit esse in eadem quantitate speculi. Immo multo forcior esset combustio
further marvelous conclusions could be deduced because I could easily si in eadem quantitate fieret secundum seccionem mukefi, quoniam in speculo facto secundum
reduce them to the conclusions which the great Apollonius inserted in his porcionem circuli non congregantur radii ad unum eundem punctum nisi de circumferencia
book On Conic Figures, still, since I have never seen his book in spite tantum unius circuli, ut demonstratur alibi. Sed in speculo facto secundum seccionem mukefi
of looking for it with the greatest diligence, the afore-stated conclusions congregantur radii a tota superficie ad unum punctum, ut demonstratum est in libro de
will, praise God, suffice for the present opuscule. speculis factis secundum hanc seccionem, et ideo ceteris paribus speculum puta quantitate
et bona factione tale speculum valde forcius combureret quam aliud, quamvis posset fieri

159
160 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 161

boloidal surface produced by the rotation of a parabola. After a brief This last-mentioned treatise continued to exert its influence in the four­
proem, the author presents verbatim extracts from the fragmentary transla­ teenth century, and most importantly on the Libellus de seccione mukefi of
tion by Gerard of Cremona of the beginning of the Arabic text of the Conics Jean Fusoris. Fusoris was born about 1365 in Giraumont (Ardennes), took
of Apollonius: the first three numbered definitions that concern the cone3 a bachelor of arts degree at Paris, following it with a master of arts
and the first four definitions added by the Arabic editor that concern the degree, and much later a master of medicine degree (1396), after which he
definition of the linea munani and the generation and naming of the conics was a regent master at Paris until 1400-1401.6 Having earlier learned and
(■sectio mukefi, sectio addita and sectio diminuta).4 Their Greek names practiced his father’s craft as a pewterer, and having become interested in
are not given. After completing these definitions taken from Gerard of astronomy, he built astronomical instruments, achieving considerable suc­
Cremona’s fragment, the author turns to the preparation of an instrument cess in that undertaking. I shall pass over his later career and his difficulties
for hollowing out the paraboloidal mirror. Here his chief source was, no with an action for treason (for supposedly aiding the English cause) merely
doubt, the tenth conclusion of the Speculi almukefi compositio that I have to note that he died in 1436. The date of his Libellus de seccione mukefi is
edited in the preceding chapter. At any rate, the author cuts away a block not known, but I would suppose that it dates before (probably considerably
to form a segment of a cone whose upper surface is a parabola.5Though he before) 1400-1401, his last year as a regent master at Paris.7
does not go on to say how this surface was to be used in preparing a cutting The Libellus is extant uniquely in a mathematical codex prepared by
or hollowing-out instrument, presumably he meant for the upper surface of Fusoris: MS Bibliotheque de Dijon, Anciens Fonds 441 (266), 204r-206r.8
the segment to be laid on a metal sheet on which the surface could be The Libellus is several times mentioned by Fusoris in notes that he added to
traced. The section traced could then be cut out and its edge sharpened, as his own summary copy of Witelo’s Perspectiva that occupied folios 154r-
it was in the earlier treatise. This account has no mathematical interest; 20 lr of the same codex. It should be observed that in his version of Witelo’s
and, since it abandons diagrams and is entirely verbal, it is not nearly so Perspectiva, Fusoris did not include the proposition numbers for the prop­
clear as the relevant part of the Speculi almukefi compositio. I hardly need ositions of Book I, while he did include them for the propositions of Book
say that its gross incompleteness and general ineptness mark it definitely as IX. Before examining in detail Fusoris’ Libellus, I shall quote in full his
the child rather than the parent of the Speculi almukefi compositio. pertinent notes on Propositions IX.39-IX.41 of the Perspectiva, where he
takes a justly critical view of Witelo’s text:9
magnum speculum secundum porcionem circuli quod combureret forcius quam parvum
factum secundum seccionem mukefi, ut si esset speculum secundum porcionem circuli « E. Wickersheimer, Dictionnaire biographique des medecins en France au moyen a g e,
cuius diameter esset trium pedum vel 4 et aliud factam secundum seccionem mukefi esset Vol. 1 (Paris, 1936), pp. 403-04. Cf. L. Mirot, “ La Proces de maitre Jean Fusoris,” Memoires
tantum [unius] pedis illud forcius combureret et intencius forte et extencius, et hoc propter de la Societe de VHistoire de Paris et de /’Ile-de-France, Vol. 27 (1900), pp. 137-287,
magnitudinem superficiei a qua plures radii refranguntur quam a parva quantitate quamvis and, more recently, E. Poulle, Un Constructeur d ’instruments astronomiques au XVe siecle
infiniti utrobique." I have added [unius] from the Corpus Christi MS. Jean Fusoris (Paris, 1963), pp. 1-6. See also Poulle’s article on Fusoris in the Dictionary
3 Paris, BN lat. 7434, 102r: “ Primo ergo videndum est quid sit seccio mukefi, ad cuius o f Scientific Biography, Vol. 15 (New York, 1978), pp. 162-63.
cognicionem requiritur videre quid sit piramis. Sicut ergo dicitur in anxiomatibus (!) Apol­ 7 1 suspect that Fusoris was still at the university when he wrote this work, primarily
lonii:‘Cum continuatur. . . . Then follow the definitions from Gerard's fragment which because he mentions consulting a manuscript at the Sorbonne (see below, Text A, Chap. 3).
are almost completely identical to Definitions [1] —[3] given above in Chap. 1, note 5. 1 Note also that Fusoris in his notes to Witelo’s Perspectiva (see my note 9, section [5])
should note the error that the scribe makes in giving Definition [3]. He says “ Et nomino calls himself beiane (which I have translated as “ ninny” ), customarily written as beanus
basim ortogoniam . . ." For basim he should have written piramidem. (or in French as bejaune), which often signifies a “ new student. But it has the further
4 Ibid., 102r-v. These definitions are exactly as in Chap. 1, notes 18 and 21, above. meaning of “a rude and unlearned fellow.” I suspect that it is the latter meaning which
The quotation from Gerard's fragment ends: “ Ista est in libro Appollonii (/) de piramidibus Fusoris intends in his self-deprecating reference. See the informative entry Beanus in Du
que permittit cum quibusdam aliis libro suo, ex quibus patet quid sit seccio mukefi.” Cange’s Glossarium.
Ibid., 102v- 103v: “ Qualiter autem hanc sectionem inveniemus et per ipsam regulam 8 For this manuscript, see Catalogue general des manuscrits des bibliotheques publiques
faciemus hic est dicendum ut occurrit. Primo ergo fieri debet pars aliqua magne piramidis de France: Departments, Vol. 5: Dijon (Paris, 1889), p. 104. This catalogue does not (but
super quam ipsam inveniemus. Sic ergo fiat ad presens. . . . " I refrain from giving the should) distinguish this tract (which ends on folio 206r) from a succeeding treatise on mirrors
constructional details for two reasons. They are not pertinent to our mathematical investi­ and other material (that may or may not be by Fusoris). The latter treatise continues to
gation, and I have already given the clearer account found in Conclusion 10 of the Speculi folio 212r. The first scholar to indicate that our tract was written by Fusoris was D. C.
almukefi compositio for making a segment of a cone whose upper surface is contained by a Lindberg, A Catalogue o f Medieval and Renaissance Optical Manuscripts (Toronto, 1975),
parabolic curve. The little tract ends (103v): “ linea ergo per ista tria puncta ducta est p. 57.
medium linee munani continentis sectionem mukefi.” A later hand adds: "Item de sectione 9 MS Bibl. de Dijon, Anc. fonds 441 (266), 196v mg.: “ [1] pro intellectu istarum duarum
mukefi de qua hic agitur — (?) in tractatu de speculis qui est _ (3) ante perspectivam et etiam sequentis 4 1‘‘ vide in folio sequenti.” Ibid., mg.: “ [2] Nota tu fusoris quod ille
fratris Rogeri in fine illius tractatus." I expect this is a reference to the Speculi almukefi 3 figure illius 41 ‘‘ propositionis sunt male facte; etiam videtur quod commentator non intel-
compositio, which I edited in the preceding chapter and which was associated in some ligebat ipsam propositionem. Ideo (?) de ista propositione (?) vide in folio (et dei. f hic)
manuscripts with Roger Bacon. sequenti.” Ibid., 197r, line 9: “ [3] Ex predictis apparet quod predicte figure tres illius propo-
162 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 163

sitionis non sunt bene facte quia in ipsis semper linea ah est maior quam linea eh. De [1. On Propositions IX.39-IX.41]
residuo, qualiter debet fieri speculum, vide in uno parvo tractatu quem habes qui incipit For the understanding of these two propositions [39 and 40] and also of the
‘de subtiliori quidem geometre etc.' Et hec est melior et fortior figatur omnium figurarum following, 41, consult the following folio.
radios solares ad unum punctum agregantium quoniam a tota superficie et a quolibet puncto
illius solares radii ad punctum unum agregantur. Adverte, tu Fusoris, quod falsum est illud [2. On the Figures of Proposition IX.41]
quod dicitur in commento istius propositionis, scilicet quod linea ae sit 4ta pars linee Ig You, Fusoris, note that the three figures of Proposition 41 [see Figs. 3.24-
quia certe Ig est plusquam quintupla ad lineam ae. Unde ex 2a figuratione apparet in com­ 3.26] are badly made. It seems also that the Commentator [i.e., Witelo] did not
mento quod linea be est dupla ad lineam ae quia ipsa est equalis linee eh, et de hoc vide in understand this proposition. Therefore, concerning this proposition consult
folio sequenti seu statim. [4] Qualiter autem componatur speculum tale, ipse docet capere the immediately (?) following folio.
ferrum scindens factum secundum formam seccionis parabole et ponere petiam in tornatorio
et mediante illo ferro facere figuram speculi. Pro aliquali declaratione illius media (?) adverte [3. On the same Proposition IX.41]
quod verbi gratia sit pyramis hqf [Fig. 5.0(a)] cuius angulus qui rectus, quam pyramidem From the above-mentioned it appears that the aforesaid three figures [Figs.
secet aliqua superficies plana equidistanter lateri hq in punctom , k , a , r [vel] s , vel quocunque 3.24-3.26] are not well made, because in them line ah (/ eh) is always greater
placuit, verbi gratia, in puncto a. Tunc constituetur seccio mukefi lag, cuius cuspis a, dya- than line eb [though the lines should be equal]. As for the rest—how the mirror
meter da et linea recta Idg, mediante qua seccione fiet reflexio ad punctum tantum distans ought to be made—consult a small tract, which you have, that begins: ‘De
ab a quantum est 4a pars linee Ig [mg.: illud est falsum: vide in pagina sequenti], que est subtiliori quidem geometre etc.” It is better and more effective if the solar rays
linea recta predicte seccionis, que quidem Ig est equalis linee hf quia ambe sunt dyametri
of all the figures [are shown to] aggregate at a single point since from the whole
eiusdem circuli, puta basis predicte piramidis, et sic fiet reflexio secundum longitudinem
linee ef, que etiam est equalis linee ac, si esset protracta. Si autem placeret quod fieret
surface and from any point of it all the solar rays are [in fact] aggregated at a
reflexio ad maiorem distantiam, tunc oporteret quod intersectio predicte piramidis fieret single point.
magis remote (!) a cuspide, qui, verbi gratia, in puncto s, et tunc fieret reflexio ad quanti­ You, Fusoris, observe that what is said in the comment on this proposition
tatem linee st et sic linea recta illius seccionis transiret per punctum /; et sic per istam [by Witelo] is false, namely that line ea is the fourth part of line Ig, because
figuram tu poteris cognoscere ubi deberet situari cuspis seccionis que deberet reflecti ad certainly line Ig is more than quintuple line ae. For it appears in the second
certam distantiam datam. Data igitur distantia designare seccionem mukefi que ad talem figuration [Fig. 3.25] that line be is double line ae because the former is equal
distantiam reflecteret radios patet predicta nunc. Recorderis etiam de tuo modo ymaginandi to line eh. And concerning this, consult the following folio, or rather [my ob­
qualiter protrahatur seccio mukefi, quia sit triangulus dqf, qui revolvatur circa axem dq. servations that follow] immediately.
Deinde sit d'a una lamina plana latonis fixa et quiescens [Fig. 5.0(b)], Deinde ymagineris
unam petiam calibis acutam que colabit supra laminam rsf altificando in revolutione pre- [4. On Fashioning a Parabolic Mirror and Protracting a Parabolic Section]
dicti trianguli, que petia relinquet vestigium supra predictam laminam, quod vestigium erit But, as for how such a mirror is fashioned, he [the Commentator in Proposi­
seccio mukefi. Item nota quod tu potes, si placet, facere triangulum predictum talis figure, tion IX.44] teaches how to take a cutting iron made according to the form of a
scilicet quod petia calibis acuta colaret supra lineam nm, quia scias quod parvum spacium parabolic section and to place the piece on a turning device and by means of the
inter n [et] / sufficiet pro valde magna seccione mukefi protrahanda. Causa est quia sec- [cutting] iron to produce the figure of the mirror.
ciones mukefi modius sunt circulares; ymmo quasi sunt linee recte, saltim ille que reflec­
For some sort of clarification of this [but following Witelo’s intent] observe
terent ad magnum spacium. Due (! Inde?) ego faciam instrumentum isto modo et iste est
bonus modus, apparet mihi pronunc. Nota quod cuiuslibet seccionis mukefi locus concursus that there be, for example, a cone hqf [see Fig. 5.0(a)], whose [vertical] angle
radiorum est in puncto medio sue sagitte, ut tu demonstrasti in libello tuo de seccionis is a right angle. Let some plane surface cut this cone parallelwise to side hq, in
mukefi protractione. [5] Nota tu, beiane fusoris, linea curva lag est vera seccio mukefi, point m, k, a, r, [or] s, or in any point one pleases, e.g., in point a. Then a para­
e vero est punctus concursus radiorum in ipsa, Ig vero est dyameter pyramidis illius sec­ bolic section lag will be formed, whose vertex is a, whose diameter is da and
cionis [Fig. 5.0(c)]. Supposita igitur figuratione sicut est facta, (197v) dico primo quod 39a whose straight line is Idg. By means of this section reflection will take place at a
propositio huius noni libri habet veritatem generaliter sive capiatur linea contingens prope point a certain distance from a that is one-fourth part of line Ig (margin: this
a sive in puncto g sive in medio ubicunque fuerit, prout clare probavi in libello de pro­ [observation of Witelo’s] is false; consult the following page), which is the
tractione seccionis mukefi, etiam nihil supponendo ab appollonio (/) pergeo. Secundo dico straight line of the aforesaid section and is equal to line hf because they are
quod 40a propositio non habet veritatem de linea Ig, scilicet quod eb sit medio loco pro­
diameters of the same circle, namely, of the base of the aforesaid cone. And so
portionalis inter lineam Ig et lineam ea, neque etiam est verum quod linea zo sit medio
loco proportionalis inter Ig et za. Sed ipsa 40a propositio bene habet veritatem de linea feb,
reflection will take place according to the length ef,10 which is also equal to
que scilicet transit per punctum concursus radiorum, ita quod linea be est medio loco pro­
ac, if the latter were drawn. But if it were preferred that reflection take place
portionalis interfb et ea, et etiam linea zo est medio loco proportionalis inter/δ etza, prout at a greater distance, then it would be necessary that the intersection of the
probavi ubi supra in libello. Tertio dico quod non est verum illud quod assumitur in com­
menti principio propositionis 41°, scilicet quod linea Ig sit quadrupla ad lineam ea, quia certe dividatur sagitta da per equalia in puncto e, tunc e erit punctus concursus radiorum illius sec­
ipsa est plusquam quintupla ad illam. Sed bene verum est quod linea fb est quadrupla ad cionis.” Notice that both piramis and pyramis are written by the author. I have added the
lineam ea. Et sic apparet quod in istis duobus commentis istarum duarum propositionum est prime sign to lamina d'a [see Fig. 5.0(b)]. I have not maintained here the author’s para­
magnus defectus, qui defectus forte processerit magis ex vitio scriptorum quam com­ graphing. But his paragraphing is followed in my translation above.
mentatoris. Si autem ampliora de ista materia volueris videre, vide in libello tuo de protrac­ 10 As I observe in the discussion of these comments below, Fusoris is in error here in
tione seccionis mukefi vera et de compositione speculi mukefi. Notandum generaliter quod believing that the focal distance from the vertex is ef or its equal ac. In fact, ac = V T focal
locus concursus radiorum in seccione mukefi est in puncto medio sue sagitte ita quod distance. The focal distance, then, is rather da / 2.
164 ARCHIMEDES IN THE MIDDLE AGES
THE FOURTEENTH AND FIFTEENTH CENTURIES 165
aforesaid [plane and] cone take place at a greater distance from the apex, e.g.,
in point s, and then the reflection would take place according to the length of It is to be noted generally that the focus of the rays of a parabolic section is
line st and so the straight line of that section would pass through point /; and so in the midpoint of its axis, so that, if axis da is bisected in point e , then e will be
by this figure you will be able to know where the vertex of the section ought to the focal point of the rays of the section.
be situated so that reflection would be at a certain distance. Therefore with the
The style of Fusoris’ notes is clearly a very personal one; sometimes he
distance given, [the problem] to designate the parabolic section which would
reflect rays to such a distance is evident by the statement just made.
uses the first person (as he does in his Libellus) and sometimes he ad­
You [Fusoris] may also recall your method of imagining how a parabolic sec­ dresses reminders to himself in the second person, as when he says “ Fu­
tion is to be drawn. For let there be a triangle dqf', which [we] let be rotated soris, you ninny . . .” or when he says “ You, Fusoris, note. . . or
about axis dq. Then let there be a plane lamina of latten d'a [see Fig. 5.0(b)] “ You may also recall your method. . . or “ as you have demonstrated
that is fixed and at rest. Then you may imagine a small, sharp piece of steel in your Booklet . . . ” and so on. In the first passage he directs himself or
which will ride on lamina rsf by changing its height during the rotation of the the reader to the succeeding discussion. In the second, he reminds himself
aforesaid triangle. This piece will leave a trace on the aforesaid lamina [d'a] that the figures for Witelo’s Proposition IX.41 are badly executed, and he
and this trace will be a parabolic section. Also note that you can, if you wish, is quite correct in this observation, so far as I can tell from the manuscripts
make the aforesaid triangle of such shape that the sharp piece of steel will ride of the Perspectiva available to me. Hence Fusoris’ observation that the
on [a small] line nm , because you may know that the space between n and/ will Commentator (Witelo) did not understand the proposition seems justified.
suffice for the drawing of a very large parabolic section. The reason is that
In passage [3] he specifies the faults in the figures. He first observes that
parabolic sections are greater in measure than [comparable] circular sections.
line eh (by a slip he says ah\) is always greater than eb in the diagrams,
Nay, they are almost straight lines, at least those that would reflect at a great
[focal] distance. Thence (?) I shall make the instrument in this way; and this is a when, in fact, the lines should always be equal, as he stresses in his Libel­
good way, it appears to me for the present. lus. In the course of this comment Fusoris mentions Alhazen’s De speculis
Note that the place of concourse [i.e., focus] of the rays of any parabolic sec­ comburentibus, identifying it by its incipit. The second error that he notes
tion is in the midpoint of its axis [intercepted between the vertex of the section is the more crucial one, since it points to a basic lack of understanding of the
and the point of intersection of the axes of the section and the cone], as you latus rectum of the parabola. In none of the diagrams is Ig (the designated
have demonstrated in your Booklet on the Protraction o f a Parabolic Section. latus rectum) actually quadruple line ea. If we look at Fig. 3.25, for exam­
[5. The Correct Understanding of Propositions IX.39, IX.40 and IX.41] ple, it is clear that Ig is more than quintuple ea, for be, supposedly equal to bh
Fusoris, you ninny, note that the curved line lag is a true parabolic section (but not so in the diagram), should be twice ea (and gd, one half of Ig, is ob­
[see Fig. 5.0(c)], while e is the focal point of the rays in it and Ig is the diameter viously greater than be). He will correct all of this in Passage [5], to which I
[of the base] of the cone of the section. Supposing the figuration as made [in shall allude in a moment. Meanwhile, in Passage [4], Fusoris produces an er­
Fig. 5.0(c)], I say, in the first place, that Proposition IX.39 has general truth roneous discussion on the determination of the focal distance from the vertex,
whether the tangent [to the section] touches near a or in point g or anywhere an extension of an error made in the tenth conclusion of the Speculi almukefi
between, as I have clearly proved in the Booklet on the Protraction o f a Para­ compositio. Fusoris tells us essentially that the focal distance [from the
bolic Section, even without supposing anything from Apollonius of Perga. In vertex] in all parabolic sections produced from a right-angled cone is equal
the second place, I say that Proposition IX.40 does not have truth with respect to the perpendicular distance to the axis of the cone from the vertex of the
to line Ig, namely, that eb is the mean proportional between line Ig and line
section. This is equivalent to saying (as he does in the second chapter of the
ea. Nor is it true that zo is the mean proportional between Ig and za. But Prop­
osition IX.40 does indeed have truth with respect to line feb, namely, the line
Libellus ) that the diameter of the base circle of the cone whose center is the
which passes through the focal point of the rays, so that line be is the mean point of intersection of the axis of the section and the axis of the cone is
proportional between fb and ea; and also, line zo is the mean proportional quadruple the desired focal distance. We should note, however, that the
between fb and za, as I have proved in my booklet, in the place mentioned anonymous author of the Speculi almukefi compositio made a rather dif­
above. In the third place, I say that what is assumed in the beginning of [Wite- ferent assertion: the radius of the circle of the base of the segment (a circle
lo’s] comment to Proposition 41 is not true, namely, that line Ig is quadruple that in his example does not contain the point of intersection of the axes of
line ea, because it is certainly more than quintuple that line. But it is true that the section and the cone as its center) is twice the desired focal distance,
Ymefb is quadruple line ea. And thus it is apparent that in these two comments and he seems to imply that this is also true for all other base circles. In the
to these two propositions there is great defect, which defect has perhaps arisen case envisioned by Fusoris it is clear from an examination of Fig. 5.0(a)
more from corruption by the scribes than by the Commentator. Now if you that the radius of the base circle is not double the focal distance but is
wish to see more about this matter, consult your Booklet on the True Protrac­
rather 2 vTtim es that distance. One could perhaps defend Fusoris by say­
tion of a Parabolic Section and on the Composition of a Parabolic Mirror.
ing that he merely meant that the focal is dependent on the quarter-diam­
166 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 167

eter, i.e., the focal distance increases directly as the quarter-diameter in­ rect determination of the proper latus rectum of a parabola from the last
creases, and, though the approximation of one to the other is a gross one, of these treatises, though he never applied the name latus rectum to this
it at least gave him some idea of the value of the focal distance when he parameter. It was for him simply the line of order that passed through the
started with a cone of a particular size. I see the need for some kind of focus, the latter determined as the midpoint of the axial segment between
defense of Fusoris only because he quite clearly understood the true geo­ the vertex of the section and the intersection of the axis of the section and
metrical relationships of these lines when he stated in the second chapter of the axis of the right-angled cone from which the section was formed. We
his Libellus that the axis da (which is double the focal distance) is the shall also see that Fusoris developed a method of describing a parabola as
diagonal of a square of which the side is one quarter of the base diameter. the locus of the totality of points that reflect to a single focus the totality of
Passage [4] also refers briefly to Fusoris’ own compass-like instrument rays parallel to the axis, the focus being located at a given distance from the
for tracing a parabola on a metal lamina that is presented in greater detail in vertex of the section. This, then, is a method that essentially freed the
the third chapter of his Libellus (see Text A below), and we shall discuss it generation of a parabola from the cone in favor of a two-dimensional
later. But I would judge from his statements concerning the instrument generation that depended only on the focus and its distance from the ver­
both here and in the Libellus that Fusoris had not actually made the instru­ tex of the section (Text A, cap. 4). Earlier in the tract he presented, how­
ment. Finally, in Passage [5], he redraws the diagram of a parabolic sec­ ever, the method of constructing a material cone, which he cut down to
tion [Fig. 5.0(c)], that allows him to refute the errors of the figures of Witelo’s produce a conic block with a parabolic upper surface, a surface that al­
text for Proposition IX.41 and to point out how Propositions IX.40-IX.41 lowed him to trace on a metal sheet a parabolic curve. The metal sheet was
can be correctly interpreted in terms of his own figure. Thus he correctly then cut along the line of the parabolic curve, its edge being thereupon
locates the latus rectum as the line of orderfeb through the focus e, though sharpened so that the parabolic shaped metal piece might serve as a cut­
he does not call it the latus rectum. Proposition IX.40 of the Perspectiva ting tool for hollowing out the paraboloidal mirror (Text A, cap. 3). This
will be true if the lineft> rather than Ig is used as the latus rectum, and hence method he admits to having taken from a work he found in the library of the
the proof of Proposition IX.41 that depends on Proposition IX.40 will also Sorbonne, a work that obviously must have been the Speculi almukefi
be correct if again we accept line fb rather than line Ig as the latus rectum. compositio. Furthermore, we shall see that he devised two different in­
Finally, he once again stresses that the focus is the midpoint of axis da . And struments of his own for describing a parabolic section, the one the al­
incidentally it is clear from Fig. 5.0(a) that by the axis da he intended the ready mentioned compass-like instrument (Text A, cap 3) and the other an
axial segment between the vertex of the section and the intersection of the instrument composed of interconnected metal rules, the resultant of whose
axes of the section and the cone, though he does not state this in a general simultaneous movements would be a parabola (Text A, cap. 5). Finally, we
way as did the author of the Speculi almukefi compositio. should note that, having described the generation of a parabola, starting
Fusoris’ Libellus de seccione mukefi, which in the colophon he entitles only with the focus and its distance from the vertex, he deduced some of the
Libellus de compositione et protractione vera seccionis mukefi (and in his well-known properties of the parabola (Text A, cap. 4). All in all, this is an
notes to Witelo’s Perspectiva as Libellus de seccionis mukefi protractione impressive performance that more than excuses an erroneous statement
or Libellus de protractione seccionis mukefi or Libellus de protractione concerning the relationship of the focal distance to the diameter of a right-
vera seccionis mukefi et de compositione speculi mukefi), has never been angled cone from which it may be cut.
published or analysed. But because of the importance of it to our investiga­ Now let us make a more leisurely analysis of the Libellus, keeping an eye
tion of conic sections in the Middle Ages, I have appended the text and a on its text below. In the brief proem, Fusoris observes that the precise
translation of it below (Text A). (In editing Text A, I have retained Fusoris’ description (i.e., drawing) of a parabola is most difficult and so the objec­
orthography, except that I am not sure that I have always been able to de­ tive of “ libellus iste tertius” is to teach such a precise description of the
cide correctly whether Fusoris has written -ti- or -ci- in such words as parabola in the manner of the continuous description of the circumference
distantia, protractio, etc. Still, it is clear that he always wrote seccio, as of a circle [by a compass]. The phrase “ libellus iste tertius” is clear in the
indeed I have done in the text below (Cf. note 9). But this is of no great manuscript but it may be an error for “ libelli istius tertium,” that is, “ the
moment, since it is sometimes evident that scribes used -cc- when in fact third chapter of this booklet,” for indeed the third chapter does show how
they intended -ct- to be understood.) My detailed analysis will show that to describe the parabola by a compass-like instrument, as we have already
Fusoris was well acquainted not only with Witelo’s Perspectiva, as has noted. But if the phrase as it stands is what was intended by Fusoris, then
already become evident from our consideration of his notes to that tract, I am not sure what were the two other works after which this was the third.
but also with Alhazen’s De speculis comburentibus and with the influential Perhaps he simply meant that this was the third work in his mathematical
Speculi almukefi compositio. It will be evident that Fusoris learned the cor­ codex (it is preceded in the Dijon manuscript by versions of the Elements
168 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 169

of Euclid and Witelo’s Perspectiva). Fusoris then goes on to say that the tion ebk, hence angle tbh = angle ebk. Further, he adds, if bk is bisected
work “ will contain four or five chapters.” Still, only the first three chapters and from the point of bisection a perpendicular is erected, it will pass
are so marked in the manuscript, but I have added headings for Chapters 4 through focus e; and finally he asserts that the axis da is the diagonal of a
and 5 in what seemed to me to be appropriate points of division. The square whose side is one-fourth the diameter of the base of the right-angled
author’s hesitation here at the beginning of the work and a series of tenta­ cone of which ba[v] is the section. All of these statements are interrelated
tive statements in the fifth chapter, like “ it appears to me for the present,” and can be proved or made clear by means of Propositions IX.39, IX.40 and
seem to indicate that the version of the Libellus in the Dijon codex is an un­ IX.41 of the Perspectiva (no doubt as clarified by the proper reinterpreta­
revised first draft. tion of them found in Fusoris’ notes to the Perspectiva). Finally, the second
The first chapter identifies seccio mukefi as the Arabic name of the parab­ chapter ends with the observation that we need not make a mirror that uses
ola and seccio parabola and seccio rectangula as the Latin names. The the whole section (i.e. the whole paraboloidal surface produced by the ro­
source for this statement was Proposition 1.98 of Witelo’s Perspectiva tation of the parabola on its axis), but we can use only a portion of the sur­
(see above, Chap. 3, note 14). He also takes the Euclidian description of the face near the vertex or in the middle or toward the base of the section. In
right-angled cone from Proposition 1.89 of the same work (ibid., note 8), to the last two cases an annular mirror will be formed. Such partial mirrors he
which he adds a physical example of the generation of such a cone by the will have found described in Proposition 5 of Alhazen’s De speculis com­
rotation of a right triangle in a waxy mass. He also indicates that such a burentibus and Witelo’s Proposition IX.44.11
cone is like a toy top. Fusoris continues the dependence on Proposition The third chapter of Fusoris’ Libellus, which, he says, embraces the
1.98 of the Perspectiva in his description of the parabolic section of a cut­ principal intent of his tract, namely, to give the actual description or pro­
ting plane parallel to one side of the axial triangle, and indeed he closes the traction of the parabolic section, contains two parts. The first part briefly
purely descriptive account by giving the full enunciation of Witelo’s prop­ describes the method of making a cutting instrument that will hollow out
osition. He says that he does this because he does not know the proposi­ the paraboloidal mirror. This method he saw in a work on the seccio mukefi
tion number. We have already seen that Fusoris’ version of the Perspectiva that he found in the library of the Sorbonne. There can be no doubt that
contained no proposition numbers for the propositions of Book I, though it this work was the Speculi almukefi compositio which I have edited in the
gives numbers for those of Book IX, and Fusoris later in the Libellus gives preceding chapter and that Fusoris was summarizing its tenth conclusion
the numbers of the three propositions of Book IX with which he is con­ briefly. In doing so, Fusoris makes the same error concerning the focal
cerned. distance that he made in the fourth passage of his notes to the Perspectiva
The second chapter of the Libellus is also largely explanatory. It arises when he tells us to make the diameter of the base circle of the right-angled
from a knowledge of Proposition IX. 41 of Witelo’s Perspectiva (see above, cone four times the focal distance from the vertex of the parabola that will
Chap. 3, note 42), perhaps of Proposition 1 of Alhazen’s De speculis become the upper surface of the segment of the cone that is his objective.
comburentibus (see above Chap. 1, note 33) and particularly of the eighth This statement differs from the similar error committed by the anonymous
conclusion of Speculi almukefi compositio (see the text in Chap. 4). I say author of the Speculi almukefi compositio, as I have already observed
‘’particularly” of the last, for instead of simply referring the focal distance above. Be that as it may, Fusoris quickly dismisses the method he has just
ae to a falsely placed latus rectum Ig (as in Witelo’s text), and instead of summarized as being “ tedious and difficult to describe” in favor of his own
merely locating the focus e on the axis a distance from the vertex of lA the instrument. The description of his own instrument occupies the second
length of an unlocated latus rectum (as in the case of Proposition 1 of Al­ part of his third chapter. In the beginning of that description [see Fig. 5.3]
hazen’s De speculis comburentibus), Fusoris, like the author of Speculi he has a right triangle adc that is to be rotated about axis ad and he cor­
almukefi compositio, locates the focus e at the midpoint of the axial seg­ rectly states that ac is twice the desired focal distance (from the vertex of
ment intercepted between the vertex of the parabola and the intersection of the section). (For other statements that bear on the question of the deter­
the axis of the parabola with the axis of the right-angled cone from which it mination of the focal distance, see Text A below, Translation, cap. 2, note
was formed. Fusoris’ account begins by announcing that, if e is the mid­ 1; cap. 3, notes 1-3.) Now let us look at Fusoris’ compass-like method
point of the axis d a , it will be the ‘‘concourse of the rays for such a section” of describing a parabola. He first takes the aforementioned right triangle,
(see Fig. 5.2). He then says that if we extend da so that ak = da and sup­ which has been made of metal and held at points a and d so that it is able to
pose that kb is a tangent to the section at b and draw ray tb parallel to d a , it
can be proved that this ray is reflected to point e . He states (without further 11 See J. L. Heiberg and E. Wiedemann, ‘‘Ibn al-Haitams Schrift uber parabolische Hohl-
proof) that eb and ek will be found to be equal, that the angle of incidence spiegel,” Bibliotheca mathematica, 3. Folge, Vol. 10 (1909-10), pp. 230-31. Cf. Chap. 3
tbh is equal to angle k , and, since angle k is also equal to the angle of reflec­ above, n. 43.
170 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 171

rotate about ad. Then a plate of steel, iron or bronze is affixed to the im­ angle of incidence mgl = the angle of reflection egi (proved in the same
mobile surface pi under the triangle, pi being perpendicular to ac (or its ex­ way as before). (9) But z is any point on the axis and thus we have in effect
tension). This plate on pi is held firm so that, when triangle adc is rotated, all points of the curve, each of which will reflect its ray to the given focus.
a sharp-pointed stylus, so held to side ac by cramps that it is allowed to And so the curve is determined. [In fact, Fusoris need not have initially
ride up and down on ac, traces a parabola on plate pi. The focal distance of determined that point b was on the curve, for he could have simply started
this parabola will then be Vi ac, as I have said. So far as I know, this is the with any point z on the axis and proceeded as in steps (6)-(9)].
first compass-like device for tracing a parabola found in the Latin West.12 Fusoris follows this exposition with a series of corollaries that are condi­
Now the fourth chapter is the most interesting and important one for my tions or properties of the determined curve (and indeed are well-known
investigation. In it, Fusoris first notes that Witelo had assumed without properties of the parabola). The first property embraces Conclusion 4 of
proof two propositions from Apollonius’ Conics in order that Proposition the Speculi almukefi compositio (=Prop. 1.11 of the Conics of Apollonius
IX.41 of the Perspectiva might be proved. This would appear to be an echo and IX.40 of Witelo's Perspectiva) and the converse of Conclusion 5 of the
of the similar statement made by the author of the Speculi almukefi com­ Speculi almukefi compositio. Fusoris takes the curve as just constructed,
positio in the proem of his work. Fusoris then declares, no doubt again in noting that its focus is e and its line of order through the focus is leg (see
imitation of the author of the earlier tract, that he (Fusoris) has not seen Fig. 5.5). Then he states for line leg, i.e. for line Ig, that eg2 = lg-ea, which
Apollonius’ work. But instead of then setting out to prove the two Apol­ assertion he does not prove but which is immediately evident from his prior
lonian propositions, as his predecessor had, he takes an entirely different construction of the curve. For in that construction he took ah = ea (i.e.,
tack. He supposes rather that he has not yet found the parabolic section and eh = 2ea) and eg = eh. Thus the circle with radius eh passes through g,
that he has nothing from Apollonius. He starts, then, only with the objec­ and from the property of a circle eg2 = 2ea-2ea. But 4ea = Ig. Therefore,
tive of finding a curve that will reflect all rays parallel to its axis to a given eg2 = lg-ea, which may be rewritten as Ig / eg = eg / ea. Now he lays out
focus at a given distance from a given vertex of the unknown curve. ed = Ig and then bisects ad in point o. With o as the center and ao as the
As I have said, the resultant curve is indeed the parabola. And so he has radius he draws a circle. It will pass through point g, since that circle must
in this process freed himself from the generation of a parabola by cut­ have an ordinate that is the mean proportional between Ig and ea and we
ting a cone, his procedure being confined completely to two dimensions. have just proved that that mean proportional is eg. The same two conclu­
His method is, then, one more step toward the acceptance of the focus as a sions are then applied to any perpendicular zb, namely that zb2 = lg-za or
key element in the definition of conic sections. Fusoris’ procedure con­ Ig I zb = zb I za (the basic property of the parabola proved in Conclu­
sists of the following steps after assuming a given focus e at a given focal sion 4 of the Speculi almukefi compositio), and that, with zc assumed equal
distance ae from the vertex a (Fig. 5.4). (1) Extend ae indefinitely toward to Ig and zc bisected in point i, a circle with radius ai must pass through
d [and also, in the other direction, beyond a] and mark off ah = ae. (2) point b. This latter conclusion, is, as I have said, the converse of Conclu­
From e erect on ad a perpendicular eb = eh, and then draw a straight line sion 5 of the Speculi almukefi compositio, which, in terms of the figure
hbk. (3) Also draw tb parallel to d a ; tb may then be considered as a ray fall­ given here, holds that if we have the parabola lag with latus rectum leg
ing on point b. (4) Now ray tb would be reflected to e along line eb because through focus e , and if we draw a circle throughf , a and b , its circumference
angle tbk (the angle of incidence) equals angle ebh (the angle of reflection). must cut the axis again at point c such that zc = Ig. Fusoris’ more general
This latter equality is so because angle tbk = angle h (two parallel lines be­ conclusions concerning his so-called first condition can be proved in the
ing cut by a third to produce the equal angles) and angle h = angle ebh be­ manner I have already outlined for the special case of ordinate eg.
cause line eb was made equal to line eh). (5) Thus we now have two points The second property is the tangential property enunciated in Conclu­
on the desired curve, a and b. (6) Other points are to be determined as fol­ sion 7 of the Speculi almukefi compositio (=Prop. 1.35 of the Conics of Apol­
lows. Take any point z in line ae [in fact, in line ad] and take point i on the lonius and Prop. IX.39 of Witelo’^Perspectiva). Though it is not proved by
extension ae such that ai = az. (7) Erect zq as an indefinitely long per­ Fusoris, its proof is again immediately evident from his method of con­
pendicular to ad. (8) Then with e as the center and ei as the radius draw a cir­ structing the curve, where this equality of za and ah (Fig. 5.6) is assumed
cumference that will intersect zq in g, and so g will be a third point on the for each and every one of the tangents.
curve, for if we connect i andg by a straight line and draw line mg parallel to The third property is that expressed as Conclusion 8 of the Speculi al­
ad, then we consider mg as a ray which will be reflected along eg since the mukefi compositio (=Prop. 1 of Alhazen’sD e speculis comburentibus and
Prop. IX.41 of Witelo’s Perspectiva). Once more the proof is obvious in
12 For the mention of an instrument invented in antiquity by Isidore of Miletus and called a the prior construction of the curve where the points of the curve were
διαβήτης (=diabeta), see my Archimedes in the Middle Ages, Vol. 2 (Philadelphia, 1976), shown to be those that produce the equality of the angles of incidence and
37vT. reflection to and from the tangents to those points (see Fig. 5.7). The point
 THE FOURTEENTH AND FIFTEENTH CENTURIES 173
172 ARCHIMEDES IN THE MIDDLE AGES

Fontana is that Fontana is also credited with a Libellus de speculo mukesi

of departure of Fusoris’ proof is that eb (the distance from the focus to any
(=mukefi), which, however, is not extant, if indeed it ever existed.18 But
point of tangency) is always equal to eh (the axial segment between the
what we are left with from Fontana’s hand on this subject are the above-
focus and the point of juncture of the tangent with the axis extended). This
mentioned notes to BN lat. 9335, notes which were composed to clarify the
too was explicit in the prior construction of the curve. Incidentally, in con­
basic terms and notions connected with the sectio mukefi and found in the
nection with this property, Fusoris relates his two-dimensional construction
copy of Alhazen’s De speculis comburentibus that was included in BN
of the curve with its conventional construction from a right-angled cone
lat. 9335. Indeed in the margin of folio 88r, opposite the expression “ secun­
when he states that Idg is the diameter of the base of the cone.
dum figuram corporis (or compositionis) mukefi” in the text of Alhazen’s
The fourth and last property given by Fusoris is strictly his own, namely,
work Fontana has written “ In hoc libro in carta 134 ego iohannes fontana
that a perpendicular erected on the midpoint [/] of the tangential segment
posui 3 formas extractas ex eadem pyramide ex sectione mukefi secundum
between the point of tangency and the juncture of the tangent with the axis
artem predictam.” I append the Latin text and English translation of these
extended must always pass through the focus e (see Fig. 5.7). Though it is
notes from folio 134r as Text B below. Observe that Fontana throughout
not proved by Fusoris, it is immediately evident from the equality of eb
and eh. uses the Venetian spelling baxis instead of the conventional basis. He some­
times writes orthogonius and orthogonaliter, while at others he writes
In the fifth and final chapter, Jean Fusoris describes a system of inter­
ortogonius and ortogonaliter. I have altered his occasional seccio to sectio.
connected rulers and a right angle that allows a steel point b by its resultant
motion to describe the parabolic curve directly (see Figs. 5.7 and 5.8). The The notes, in fact, demand little comment. They comprise seven paragraphs
(to which I have added paragraph numbers), from which it is clear that
instrument was designed to exploit the properties outlined in the fourth
Fontana had not only read Alhazen’s work, whose existence in the manu­
chapter. I think it evident from his account that, though Fusoris was an
script had stimulated him to compose the notes, but also Gerard of Cremona’s
important instrument maker, he had not, at the time of this tract, actually
fragmentary translation from the Arabic text of the Conics that preceded
constructed the instrument, for he several times uses rather tentative lan­
Alhazen’s tract in Fontana’s codex, Witelo’s Perspectiva, and the Speculi
guage like the clause “ que mihi pronunc occurrit.” Furthermore, he seems
almukefi compositio. In the second paragraph Fontana gives the Euclidian
not to have known for sure whether the inclusion of a rule that is perpen­
description of the generation of a right-angled cone by the rotation of an
dicular to hb at the latter’s midpoint would make unnecessary rule zb,
which he had suggested earlier. Surely, had he built such an instrument, he isosceles right triangle, as indeed had Witelo and the author of the Speculi
almukefi compositio. The third paragraph gives the formation of the axial
would have not been so hesitant. I suspect that the idea for such an instru­
triangle of the cone in a manner that had by this time become conventional
ment came from Witelo’s detailed description of a similar instrument for
from the circulation of the earlier works. Next (Paragraph 4) Fontana briefly
describing an “eccentric line.” 13
describes the passage of a cutting plane (not through the apex of the cone)
Not long after the composition of the Libellus of Jean Fusoris, some brief
that is perpendicular to the plane of the axial triangle so that the common line
notes on the parabola were appended by the Venetian physician Giovanni
of the two planes is parallel to one side of the axial triangle. Such a section
Fontana to MS Paris, BN lat. 9335, 134r, the celebrated codex of Gerard
produces a surface contained by a curved line (which is either called an “ec­
of Cremona’s translations, which Fontana owned.14 I remark in passing
centric line” — Fontana’s name for it19— or a linea mumiami [/]) and a
that it may have been Jean Fusoris to whom Fontana refers when he men­
straight line which is the diameter of the base circle of the cone. It is ob­
tions “a certain man of subtle genius who constructed an instrument in
vious that Fontana has misread the word munani that is quite clearly given
1416 called the Mirror of the Planets.” 15 Indeed Fusoris did build such an
in BN lat. 9335 in the fragmentary translation by Gerard of Cremona of
instrument, though he appears to have finished it by 1414 when he proposed
the first part of the Arabic text of the Conics (see above, Chap. 1, note 5,
its sale to Richard Courtenay, the Bishop of Norwich, with whom a friend­
passage 4). Fontana then goes on to say that the line that is the common
ship developed that led to charges of treason against Fusoris.16 Fusoris had
section of the cutting plane and the axial triangle is called the “ axis” or
also composed a treatise that described this instrument. Fontana may well
“ arrow” of the surface containing the eccentric or curved line. This axis
have had some knowledge of, or even met, Fusoris, when the latter visited
or arrow is necessarily intersected by the axis of the cone in a point which
Bologna in 1410 to meet with Pope John XXIII and give him an astrolabe
Fontana designates as the “ common point.” Any perpendicular to the sec­
and a sphere.17 Another rather vague connection between Fusoris and
tion’s axis (within the surface contained by the curved line) is called a “ line
1:! Clagett, Archimedes in the Middle Ages, Vol. 2, pp. 24-26. ls Clagett, Archimedes in the Middle Ages, Vol. 3, pp. 250-51.
u Ibid., Vol. 3, pp. 243, 250. 1S See Text B, Passage [4], where he calls the curved line a linea eccentrica ‘‘because it
15 Ibid., p. 241. does not have a center.” A similar term (linea eccentralis) was used by Witelo for an entirely
1β Poulle, Un constructeur, pp. 2-3. different curve (see Clagett, Archimedes in the Middle Ages, Vol. 2, p. 25n).
17 Wickersheimer, Dictionnaire biographique, p. 403.
174 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 175

of order.” This name was also taken by Fontana from Gerard’s frag­ lonius’ Conics (see my discussion below).22 At least, both of these copies
mentary translation (ibid., passage 7). Finally in Paragraph 4 Fontana of the Gerard of Cremona fragment were made without any interpretative
notes that the line of order passing through the midpoint of the [segment of comments by the copier. It also appears that Regiomontanus copied Al-
the] axis intercepted between the vertex of the section bounded by the hazen’sDe speculis comburentibus. This is extantin MS Vienna, National­
curved line and the axis of the right-angled cone is called the “latus rec­ bibliothek 5258, 39r-46v. Thus it just precedes one of the copies of the
tum.” The source of this definition was Conclusion 4 of the Speculi al- Gerard of Cremona fragment noted above and it appears in the manuscript
mukefi compositio (see above, Chap. 4, Text, Concl. 4). Paragraph 5, immediately after Regiomontanus’ notes to the Speculi almukefi com­
which has its origin in Proposition 5 of Alhazen’s De speculis comburenti­ positio (which notes we shall discuss below). Zinner declared that this
bus, describes three forms that can be made and used as cutting instruments copy of the De speculis comburentibus contained annotations by Regio­
whereby the burning mirrors can be fashioned. These forms are half-parab­ montanus.23 But collation of it with Alhazen’s text demonstrates that
olas of differing focal distances (from the vertexes). Presumably he meant this opinion was incorrect, since the few marginal notes (e.g., on folios
for these half-parabolas to be mounted by their axes so that they could be 40r and 41r) are merely phrases from the original text that were omitted in
rotated to hollow out the mirrors. Paragraph 6 observes that one can deter­ the copying. This copy of Alhazen’s text is not a particularly good one, for
mine the focal distances from the vertexes by measuring off the length of the copier has omitted quite a few lines beyond those added in the margins.
one quarter of the latus rectum of each form, the measure first given by Hence, though this work may have been important in stimulating Regio­
Alhazen in the first proposition of his De speculis comburentibus. Finally montanus’ interest in conics, it has no interest to us for any new contribu­
in Paragraph 7 Fontana gives the Arabic and Latin names of the parabolic tion to the general knowledge of the subject.
section. These he took from Proposition 1.98 of Witelo’s Perspectiva (see Regiomontanus’ importance in our history of the medieval traditions of
above Chap. 3, note 14). conic sections lies then not in these aforementioned copies but rather in
Fontana’s interest in the parabolic burning mirror as presented by Al­ his treatment of the Speculi almukefi compositio, of which he presented an
hazen was shared by his contemporary, the celebrated Leon Battista Al­ improved version and to which he added additional notes. The version
berti.20 However, Alberti’s concern with the parabolic mirror was non- (Text C below) occupies folios 27r-35r and the additional notes (Text D
mathematical and hence requires no discussion here. below) folios 35v-38v of Vienna manuscript 5258. Let us first consider
To encounter the next important mathematical discussion of the burn­ Regiomontanus’ version (hereafter cited as Text C) and compare it with
ing mirror, we must examine the contributions of Regiomontanus to the the original version presented in Chapter Four above (hereafter called
developing knowledge of conic sections. It should be noted at first that Orig.).
Regiomontanus twice copied (or had copied?) with little alteration (but Proem. This is greatly reduced in Text C. The remarks by the author of
with the omission of the concluding lines) the fragment translated by Orig. that stressed his personal difficulties in understanding the Apollonian
Gerard of Cremona from the Arabic version of the Conics of Apollonius propositions cited by Witelo without proof and the author’s fruitless search
(MSS Vienna, Nationalbibliothek 5203,141r-v, and 5258,47r-48v).21 This for a copy of Apollonius have been virtually eliminated in Text C. One
was the fragment I discussed in Chapter One above. Assuming that both might argue, I suppose, that Regiomontanus’ omission of the comment
copies were indeed made by Regiomontanus, I would suspect that they about the lack of the text of Apollonius was the result of Regiomontanus’
were made when Regiomontanus was in Vienna (that is, before 1461), and having seen or heard of a Greek manuscript of the Conics, and that this
thus before he had learned Greek and seen a Greek manuscript of Apol­ would mean that Regiomontanus prepared this version after he went to
Italy. But I doubt this argument since there is no persuasive evidence, in
20 L. B. Alberti, “ Della Prospettiva,” Opere volgari, ed. of A. Bonucci, Vol. 4 (Florence, his version of the Speculi almukefi compositio or in the added notes, that he
1847), pp. 138-39. “ Ma questo specchio concavo e fatto artificialmente per modo e forma che knew the Conics at first hand, other than in the fragment translated by
tutti quanti li razzi reflessi si uniscono in un punto solo, per la quale ragione infinito ed im­ Gerard of Cremona which appears in the same manuscript. Rather I think
menso caldo si genera. . . . Di questo specchio parla il Maestro delli specchi molto prolisso that the shortening of the proem was the result of a deliberate effort on
e lungo, e molti altri autori. E perche la materia e lunga e non fa al proposito la lascero per
brevita. . . .” Bonucci misidentifies the Maestro delli specchi as Euclid. Alberti probably Regiomontanus’ part to remove or limit the personal remarks made by the
intended Alhazen (i.e. the author of the De speculis comburentibus), but it could be that he original author and to eliminate superfluities.
was referring to Witelo.
21 E. Zinner, Leben und Wirken des Joh. Muller von Konigsberg genannt Regiomontanus, 22 The character of the various tracts in the appropriate sections of the two manuscripts cer­
2nd. ed. (Osnabruck, 1968), p. 308, notes that this fragment occupies folios 141r-v of MS tainly seems to be of the kind he was interested in during his pre-Greek period. Cf. my Archi­
Vienna, Nationalbibl. cod. 5203, but he leaves the erroneous impression that it is a work of medes in the Middle Ages, Vol. 3, pp. 342-54.
Regiomontanus (see below, note 30). 23 Zinner, Leben und Wirken, pp. 308-09.
176 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 177

Definitions. To the first definition concerning the formation of a cone Conclusion 5. This proposition has been presented with very little
Regiomontanus added (perhaps from a marginal reference like that in MS change by Regiomontanus. However, he added in the margin (Text C, lines
O — see variant readings to Orig., Descrip., var., line 2) a reference to the 21-24) that it is valid not only when the line of order is drawn between the
beginning of Book XI of the Elements of Euclid. In connection with the vertex and the position of the latus rectum (through the focus) but also
third definition, it will be remembered that because the additional remarks when the line of order is drawn after the latus rectum or even after the inter­
which gave the three names of each of the three conic sections were miss­ section of the axis of the cone with the axis of the section. He concludes
ing in the oldest manuscript (MS C), I relegated them to the variant read­ that from this it is inferred that the more the section is extended from the
ings (Descrip., var., line 37). Regiomontanus, however, included the sub­ vertex the more the lines of order grow in size.
stance of them in his text (Text C, Def., lines 34-37), and thus perpetuated Conclusion 6. Note Regiomontanus’ change of intrinseco (Orig., lines 4-5)
the mistaken inclusion of the word mukefi in the Latinized Arabic names to extrinseco (Text C, lines 3-4), a change that seems felicitous. Further­
for hyperbola and ellipse. However, Regiomontanus quite correctly used more, Regiomontanus has quite properly added the specification of the
the expression sectio addita without the offending mukefi at the end of Con­ enunciation: “ Dico quod latus erectum sectionis se habeat ad he sicut eg
clusion 10 (Text C, Concl. 10, line 93), while the printed edition of Regio­ se habet ad ea." (Text C, lines 22-23), which was missing in the original
montanus’ version substituted sectio hyperbole (ibid., var., line 93). version (or at least in the manuscripts of it which are now extant). Cor­
Finally, we should note that at the end of the last definition (not numbered by rect geometrical form certainly demands this specification. Again notice
Regiomontanus) the original author’s statement concerning his own that Regiomontanus’ proof is more compressed than the original, even in
discovery of the nature of the latus rectum (Orig., Descrip., lines 53-57) this short proof. To the end of the proof Regiomontanus has added in the
was omitted by Regiomontanus. margin (Text C, lines 29-32) a note to the effect that, since from the next
Conclusion 1. To the first conclusion Regiomontanus has added a mar­ conclusion ea = ag (and hence eg = 2 ea), accordingly from this proposi­
ginal note to the effect that it is true for every cone, whether it is right or tion eh will always be one half of the latus rectum (i.e., with latus rectum /
oblique (Text C, lines 8-9). eh = eg I ea and with eg / ea = 2/1, hence eh is one half the latus rectum).
Conclusion 2. Regiomontanus has made no significant change in this Conclusion 7. Regiomontanus has included the construction of per­
conclusion. pendicular^ (Text C, line 13) that properly belongs among the necessary
Conclusion 3. In his version Regiomontanus has added a marginal note specifications but which was missing in Orig., where it should have ap­
to the effect that the third proposition is valid for every cone, whether it is peared after lineata in line 14. Similarly, when the converse is specified
right or oblique (Text C, lines 6-7). Regiomontanus has made no significant (Text C, lines 14-16) Regiomontanus again added the construction of ed
change in the proof of this proposition, though he has altered the reference (but in the margin), though this construction was once more missing in the
to Euclid VI.2 (Orig., line 16) to VI.4 (Text C, line 17). In fact, Regio­ original version. Regiomontanus has presented the proof substantially as it
montanus has made a great many changes in the Euclidian numbers through­ was in the original version, again with some compression. However,
out the work (and particularly in the next conclusion), and I shall not detail Regiomontanus has added a few lines to clarify and complete the proof
these changes here. It is not the case that the original author cited wrong (Text C, lines 53-57). In this marginal addition Regiomontanus has used
propositions but merely that Regiomontanus preferred others that seemed the word implicatio (line 57) with the meaning of “ contradiction.” Im­
to him more directly applicable. The references to the Elements in both plicatio in this usage has become the abstract noun that abbreviates the
texts are to the Campanus version. common logical expression implicare contradictionem. In fact, in French
Conclusion 4. Regiomontanus has made no significant change in this con­ this became the principal logical meaning of implication. Incidentally,
clusion, though, as I have just said, he changed a number of the proposi­ we should note that Regiomontanus in the second half of the proof has sub­
tion numbers in the citations to Euclid. While I shall not detail these changes, stituted basis for the letter c used in the original version.24 Finally, it will
we ought to note that Regiomontanus quite properly omitted the reference be evident to the reader that Regiomontanus has omitted, or rather radi­
cally reduced the final lines of this conclusion (Orig., lines 61-67), lines
to Euclid VII. 16 and VII. 17 (Orig., line 31) since the proof is couched
in terms of continuous geometrical magnitudes rather than numbers. In the which remarked that this proposition was a conclusion of Apollonius cited
proof of the second part of the proposition, Regiomontanus has presented
a somewhat different manipulation of the magnitudes but without any sub­ 24 He made this change no doubt because in the original version two different c's seem to
have been implied. The parabola was designated abc in the text of the Speculi almukefi com­
stantial change. Regiomontanus’ proof is more direct. At the end of the positio, which hence implied the use of a letter c on the left side of the parabola, while the
proof Regiomontanus has added a marginal note indicating that the second second half of the proof indicated that c was beyond d on the right side of the parabola (see
part of the proposition can be proved in the same way when the line of order Fig. 4.8, where I have added the first c as [c']). Regiomontanus' solution here was simply to
(m V ) is placed beyond the axisfo (see Fig. 4.7 and Text C, lines 68-69). substitute the word basis for the second c .
178 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 179

by Witelo and the perspectivists without proof and which added a reference occasionally corrected minor errors, and added a few marginal notes that
to the other proposition of Apollonius demonstrated in the second half of served to generalize the propositions somewhat. Regiomontanus has thus
Conclusion 4. Again I suggest that Regiomontanus was motivated by his left us with a tighter and more reliable version of the Speculi almukefi com­
desire to eliminate personal and superfluous remarks. positio. It was fortunate that Regiomontanus’ version fell into the hands of
Conclusion 8. Again notice a certain compression in Regiomontanus’ A. Gogava, who published it in his Cl. Ptolemaei Pelusiensis mathematici
version of this conclusion, but he has made no substantial change. He has operis quadripartiti in Latinum sermonem traductio . . . Item, De sec­
altered Orig.’s citation of the Pythagorean theorem from per dulcarnon tione conica orthogona, quae parabola dicitur: Deque speculo ustorio,
(Orig., line 74) to its numerical designation per penultimam primi (Text C, libelli duo. . . . (Lovanii, 1548). Gogava changed the terminology some­
lines 60-61). what (witness the title De sectione conica orthogona, quae parabola
Conclusion 9. Regiomontanus has omitted the transitional paragraph dicitur), as the reader will readily see if he examines the variant read­
(Orig., lines 10-14), no doubt for the reason stated at the end of my com­ ings to Text C below. I have not included Gogava’s orthographic changes
ment on Conclusion 7. or transposed letters representing magnitudes in the variant readings. I
Conclusion 10. As I noted in the preceding chapter, the letters and the give no English translation of Regiomontanus’ version of the Speculi
diagrams for this proposition were corrupt in the extant manuscripts of almukefi compositio, since I have already added a translation to the earlier
the original version, and I had to correct that proposition on the basis of version in the preceding chapter.
Regiomontanus’ version. I particularly point to the possible interchange of More important than the few marginal comments that Regiomontanus
letters o and m in the original version (see Fig. 4.12 and the variant read­ added to his version of the earlier tract, and which we have reported on
ings to Orig., lines 62, 66 and 86). But had I adopted the interchange, this above, were those fairly extensive additions made by Regiomontanus fol­
would have left other inconsistencies in the text. Hence my appeal there to lowing the text, the additions I have edited below as Text D.
Regiomontanus’ version. The most important step that Regiomontanus Passage [7]. In the initial paragraph, Regiomontanus shows how to com­
took in connection with this proposition was his elimination of the long ap­ pute the length of a ray reflected to the focus from a point on the parabola
pendix “ On the Conditions of Good Steel.” This is not surprising in view of (where that parabola arises from the cutting of a right-angled cone). In the
his practice of omitting material that was mathematically superfluous. examples, he posits a latus rectum of 100. Then he declares immediately
Note in conclusion that at the end of his version of this proposition (lines that the burning distance from the vertex (i.e. the focal distance) is 25 (since
83-91) Regiomontanus has given a correct computation of the focal dis­ the focal distance is equal to the axial intercept of the latus rectum when
tance from the vertex of the parabolic face of the conic segment (to be used the latter is the line of direction through the focus and the axial intercept in
to make a cutting instrument for hollowing out a paraboloidal mirror) this case was shown to be one fourth the latus rectum in Proposition 4 of the
whose base circle has a radius of 40 and whose depth (the perpendicular Speculi almukefi compositio). In the second example where the point of
distance from the vertex of the parabolic face to the base circle of the seg­ reflection is assumed at the end of the latus rectum (drawn as a line of order
ment) was 1. He computed the focal distance as a little more than 27. He through the focus), the focal distance is correctly designated as 50 (since
added a similar computation when the radius was 5. These calculations that distance is obviously half of the latus rectum). In a third example,
were added by Regiomontanus to clarify and correct the statement made in Regiomontanus assumes an axial intercept of 5, indicates that the square of
the original version (lines 92-94): “ et habebis speculum comburens ad the line of order (here and throughout used not for the whole line of order
tantam distantiam quanta est medietas semicirculi cuius portio abc [Fig. from one side of the curve to the other but for its half [i.e. the ordinate])
4.12] lineatur in prima regula,” the specific value of the radius having been erected on the terminus of the axial intercept is 500, and hence he con­
assumed to be 40 feet to produce a burning distance of 20 feet (Orig., cludes that the ray to the focus from the point at which that line of order
lines 5-8). intersects the parabola is 30. He then notes the general procedure for any
Conclusion 11. Again Regiomontanus has compressed the proof while point on the parabola between the vertex and the point of intersection of the
making no substantial change in it. The only point worth noting is that once latus rectum through the focus. If you start with lines of order positioned at
more Regiomontanus has omitted general remarks at the end of a proposi­ will on the axis, then you will be able to determine the lengths of the lines of
tion (Orig., lines 84-91) in favor of the terse statement: “ These things order [i.e. ordinates] by the [second half of] Proposition 4 of the Speculi
briefly suffice.” almukefi compositio. Then “ add its square [i.e. of the line of order] to the
The conclusion that we may draw from the preceding considerations is square of the residual segment of the axis [intercepted] between the line
that Regiomontanus did not substantially change the form and intent of the of order and the point of intersection of the axis of the section with the
proofs of the original version, though he often compressed these proofs, latus rectum [i.e. the focal point]. And the root of that sum will be the dis-
180 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 181

tance of combustion." Regiomontanus then refers to the accompanying sion as had that in the earlier treatise. After the first proof, Regiomontanus
table given on folio 36r (and presented below with the Latin text). adds a further proof based on the converse of the conclusion. That is, he ap­
The table is highly interesting, for it permits the establishment of 50 pears to hold that if the proposition is assumed then its converse is true (and
points that lie on a parabolic curve whose latus rectum is 100. In addition to vice versa). But Regiomontanus shows by a simple demonstration that the
assuming the latus rectum of 100 (and accordingly the focus located on the converse is true; hence the proposition is true.
section’s axis at a distance of 25), the table assumes that the axis of the Passage [5]. This concerns Conclusion 4 of the Speculi almukefi com­
section between the vertex and the point of intersection of the axis of the positio. The latter had been developed on the basis of a parabola formed
section with the axis of the cone is divided into 50 parts (i.e., the table as­ from a right-angled cone. Regiomontanus here extends his consideration to
sumes axial intercepts of fifty lines of order, the intercepts proceeding from parabolas formed from acute-angled as well as obtuse-angled right cones,
a length of 1 to a length of 50). As Regiomontanus explains in his notes the section in either case being produced by a plane perpendicular to the
to the table, we must first erect indefinitely long lines of order (i.e. perpen­ axial triangle of the cone when the linear common section of the two planes
diculars) on the termini of the axial segments el (see Fig. 5.11). Then we can (i.e. the axis of the section formed) is parallel to a side of the axial tri­
lay off on these perpendiculars the lengths Im in any one of three ways. angle. He commences with an acute-angled cone whose vertical angle is 60°.
That is, we can determine the fifty values of m corresponding to the fifty He shows (see Fig. 5.13a) that, if the axial segment fe is one quarter the
assumed values of el in one of three ways. First we could simply take the axial segment intercepted between the vertex and the point of intersec­
values of Im directly from the table, extending a compass successively to tion of the section’s axis with the cone’s axis, then the whole line of order
each of these lengths; and with the fixed foot on points / we can directly through e, i.e. pg, is equal to Afe, and pg is hence the latus rectum. His
mark off the lengths Im on the indefinitely long perpendiculars with the second proof shows that the line of order equal to the latus rectum can be
mobile foot of the compass. Or we could take from the table the successive found for any parabolic section formed from a cone whose vertical angle is
values of hm that depend on the assumed values of el and then, with the acute, so long as we know the sides of the axial triangle. He remarks that
compass extended to the values of hm and with the fixed foot always on h , the latus rectum in such a case falls between [i.e., intersects the section’s
the successive points m can be marked on the indefinitely long perpen­ axis between] the vertex of the section and the middle point of the axial
diculars. Or we can proceed by a method that depends on Proposition 5 of segment between the vertex and the point of intersection of the axis of the
the Speculi almukefi compositio. We add one half el to one half the latus section with the axis of the cone. (The reader will recall that in the case of
rectum. Extending the compass to the magnitude represented by this sum, the parabola formed from a right-angled cone the author of the Speculi
we put the mobile foot on the vertex of the section and let the fixed foot fall almukefi compositio showed that the line of order equal to the latus rectum
on the axis. Then the mobile foot will cut off the value Im on the indefinitely passes through the above-mentioned middle point.) Finally Regiomon­
long perpendicular. This of course is done successively for all values of tanus determines, for the specific case of a parabola formed from an ob­
el. Once more we have thus established the points m that will lie on the tuse-angled cone the sides of whose axial triangle are each 72, thatfe is 25,
parabola. The inference that we might draw from these procedures and i.e., that the line of order equal to the latus rectum passes through e at a
perhaps from the rather unclear remark "that no interruption may be seen distance of 25 from the vertex. He then concludes generally that the point
in it (the curve?)" is that the parabola is the locus of all such points m. At of intersection of the line of order equal to the latus rectum in the case of a
any rate, the procedures here outlined establish that we can find as many parabola formed from an obtuse-angled cone must lie, on the axis, between
points as we desire through which to draw a parabola with a given latus the middle point of the section’s axis (i.e. of the axial segment intercepted
rectum. between the vertex and the common intersection of the two axes) and that
The notes concerning the table conclude by expressing two important common intersection. Finally, he observes that the closer to d (the com­
linear equations that relate the various magnitudes that appear in the table: mon intersection of the axes) that the latus rectum intersects the axis of the
(1) Im + hm = Vi latus rectum, when the line of order Im lies between the section the more obtuse is the vertical angle of the cone, and that the closer
vertex of the section and the focus, and (2) hg + hi - hm, when Im lies be­ to a (the vertex of the section) that it (the latus rectum) intersects the axis of
yond the focus (proceeding in the direction away from the vertex). Regio­ the section formed from an acute-angled cone the more acute is the vertical
montanus observes finally that the validity of these relationships will be angle of the axis.
established below [in Passage (4)]. The assumptions that underlie all of this discussion are that the latus
Passage [2]. Regiomontanus here gives alternate demonstrations of Con­ rectum is four times the axial segment it intercepts and that the axial intercepts
clusion 7 of the Speculi almukefi compositio. The principal demonstra­ of lines of order are proportional to the squares of those lines of order (i.e.
tion is an indirect proof that does not depend on the converse of the conclu- ordinates) intercepting the segments.
182 ARCHIMEDES IN THE MIDDLE AGES THE FOURTEENTH AND FIFTEENTH CENTURIES 183

Passage [4]. Here Regiomontanus gives clear and simple proofs of the sion 4. He notes further that one can then deduce (for this parabola] every­
two linear equations presented above at the end of Passage [1]. He first thing that was deduced for a parabola produced in a right cone. Finally,
proves (see Fig. 5.14) that if point / on which line of order Im is erected lies he shows how hg, and thus its halffh, may be determined from the known
between e and h (i.e. between the vertex and the focus), then Im + hm sides ab and be when the axial relationship dq2/ hg2 = df/fh is used, or, in
= hg. But if / is beyond the focus h (see Fig. 5.15), he then proves that gh another way, howpg, and hence its quarterf h , may be determined from the
+ hi = hm. The proofs need no further comment here. Regiomontanus same known quantities by using the mean proportional relationship rewrit­
ends this passage by remarking on the marvelous uniformity of the rela­ ten as pg ~ dq2 / df.
tionship between the lines of order and the axial segments they intercept. All of these passages which Regiomontanus has added to his version of
He compares it to the uniformity of the relationship that exists between the the Speculi almukefi compositio show how well Regiomontanus had mas­
right-sines and versines of similar arcs. He says that the curvature of sec­ tered the medieval traditions of conic sections. It shows, too, how he
tions having the same relationship are the same whether the section is large extended the medieval treatment of parabolas, which was primarily limited
or small. to parabolas formed from right-angled cones, to consider parabolas formed
Passage [5]. This short passage indicates how to find the length of the as well from cones with any other vertical angles, acute or obtuse. One
axial segment which the latus rectum intercepts and also the latus rectum might, at first glance, believe that he made this extension after reading
itself of an annular paraboloidal mirror when we know the radii of the bases the Greek text of the Conics. But I am sure that such a conclusion is
of the mirror. Though the procedure needs no comment here, we should not warranted, since he mentions no Apollonian proposition beyond those
note that Regiomontanus proposed such a problem in a letter to Christian treated in the Speculi almukefi compositio. Still, his extension of the medi­
Roder.25 eval treatment (particularly the extensions evidenced in marginal notes re­
Passage [6]. Here Regiomontanus proposes to elicit a parabolic sec­ garding oblique cones and in Passages [3] and [6] of his additional notes)
tion from any cone (i.e., not just from the right cones of Passage [3]) when was no doubt influenced by his reading of the fragment of the Arabic ver­
the sides of the cone’s axial triangle are known. The generation of the parab­ sion of Apollonius’ Conics translated by Gerard of Cremona. From this
ola given by Regiomontanus here is strictly the Apollonian one, namely he would have learned of the more general generation of the parabola
by passing through the cone a plane perpendicular to the axial triangle and from right cones with any vertical angle and indeed even from oblique
parallel to one side of the triangle. As I shall note in my general remarks be­ cones. In line with his generalizing of the medieval treatment of parabolas,
low, Regiomontanus no doubt took this generation from the fragment that we ought to remark in summary (1) that he emphasized the equation of
Gerard of Cremona translated from the Arabic version of the Conics rather the parabola in terms of abscissa and ordinate (i.e., to use modern terms,
than from the Greek text of the Conics. The author does not in this pas­ y 2 = px, wherep is the latus rectum), (2) that he ignored the Archimedean
sage simply prove that the curve is a parabola, for it is clear that he assumes definition of the latus rectum as “ twice the line to the axis’’ (since such a
this by supposing that in the section so generated there is some line of order definition is true only for a parabola formed from a right-angled cone),
{hg) that is twice its axial intercept ( fh , see Fig. 5.17). Rather Regiomonta­ emphasizing rather the concept of the latus rectum as the whole line of
nus is here attempting to show how to determine the latus rectum pg and order that passes through the focus, a conception of the latus rectum which
hence its intercept fh when we know the sides of the triangle. True, after is true for all parabolas, and (3) that he developed and stressed two key
assuming pg = 4f h , he does prove that in the parabola here described any complementary linear equations that hold for parabolas (i.e. the equations
line of order mn is the mean proportional between pg and f m , noting that Im + hm ~ hg and gh + hi = hm demonstrated in Passage [4]).
this can be proved in the same way as in Conclusion 4 of the Speculi al- A final word is in order concerning the various pieces that arose from
mukefi compositio. In the course of his own brief proof, he asserts the axial Regiomontanus’ consideration of the conic sections concerned with burn-
property hg2 / mn2 = hf / fm , which indeed he could have easily proved ning mirrors. In the list of works that he hoped to publish, he included
with the fundamental property of the circle, as it was so proved in Conclu- as the final item among his own works a tract De speculis ustoriis atque
aliis multorum generum ususque stupendi.261 have found no work among
25 M. Curtze, “ II. Der Briefwechsel Regiomontan’s mit Giovanni Bianchini, Jacob von
Speier und Christian Roder,” Urkunden zur Geschichte der Mathematik im Mittelalter und
his manuscripts with such a title. Perhaps he intended to expand his treat­
der Renaissance, Abhandlungen zur Geschichte der mathematischen Wissenschaften, XII. ment of the Speculi almukefi compositio, which I have published here as
Heft (Leipzig, 1902), p. 335: “ Habeo speculum Archimedis annulare ex portione parabolica Texts C and D, but was prevented from doing so by his premature death.
factum, cuius margo circularis maior quinque pedes continet, minor autem tres; profunditas It seems likely that all of the above-discussed activity of Regiomontanus
vero speculi est bipedalis; quero locum ustionis itemque sagittam cum latero erecto. Voco
autem profunditatem speculi partem axis, que inter centra duorum circulorum marginalium
conducitur." 26 Zinner, Leben und Wirken, Tafel 26, c. 2.
184 ARCHIMEDES IN THE MIDDLE AGES

took place before he saw the Greek text of Apollonius' Conics. There is A.
little doubt that after he went to Italy he obtained a copy of the Greek
text, for he lists it among the works he will publish.27 Indeed it seems Johannes Fusoris
certain that the extant manuscript Nuremberg, Stadtbibliothek, Cent. V, The Booklet on the Parabolic Section
App. 6, bearing his name is that manuscript.28 And no doubt it was this
manuscript that Bernhard Walther, his associate at Nuremberg from about
1471-75, used in making his lost Latin translation of the Conics.29 Zinner’s
suggestion that Regiomontanus himself made a translation of the beginning [Libellus Johannis Fusoris de seccione mukefi]
of the Conics is probably incorrect.30 At any rate, in the list of works to
be published, Regiomontanus mentions no Latin translation of the Conics, /Libellus iste est de seccione mukefi mediante qua componuntur specula
either of his own or Walther’s. From this we can probably conclude that ad magnam vel parvam distantiam comburentia prout placet.
if Walther did indeed make a translation, as seems probable from its
mention in the Catalogues of 1512 and 1522, he made the translation [Proemium]
after Regiomontanus' death in 1476. But one thing that does bear on the
5 Subtili indagacione reperierunt perspecti vi geometrice, inter omnes
extent of Regiomontanus’ knowledge of Apollonius’ text is the fact that
specularum figuras, seccionem mukefi in agregatione radiorum aptissimam
none of his works (at least, of those I have seen) seems to contain any
et in combustione propter radiorum unionem efficacissimam; ipsius tamen
discussions of the details of Apollonius’ work. One would certainly have
descriptionem prescisam esse difficilimam. Ea propter libellus iste tertius
expected some work reflecting his knowledge of the Conics had he trans­
prescisam ipsius seccionis mukefi docet descriptionem protractione con­
lated, or otherwise mastered, the Greek text.
io tinua, veluti describeretur circulorum periferia. Qui libellus quatuor vel 5
continebit capitula.

[Capitulum primum]
Capitulum primum declarat quid vocetur seccio mukefi.
Seccio mukefi, arabice sic dicta, que latino sermone seccio parabola
sive seccio rectangula nuncupatur, est communis seccio piramidis rotunde
5 rectangule et superficiei plane eam secantis eo modo qui postquam statim
dicetur. Sed ad hoc intelligendum prescire oportet quid sit pyramis ro­
tunda rectangula, sive orthogona, quod idem est. Protinus intellectum sit
triangulus orthogonus sive rectangulus abc [Fig. 5.1] cuius angulus a sit
rectus et quilibet angulorum b et c sit medietas recti ita quod due linee
10 ab et ac sintequales, qui triangulus revolvatur supra axem ad quiescentem,
tunc talis triangulus in revolutione sua piramidem rotundam rectangulam
27 See the same Tafel 26, c. 1. describit cuius cuspis sive vertex est a, due vero linee ab et ac describunt
28 Apollonii Pergaei quae Graece exstant, ed. J. L. Heiberg, Vol. 2 (Leipzig, 1893), p.
XIII, MS n. 20. Cf. Zinner, Leben und Wirken, p. 328.
longitudinem piramidis, scilicet superficiem curvam, linea vero bc de­
29 This translation was referred to in the list of 1512. See Zinner, Leben und Wirken, p. 328: scribit basim ipsius que est circulus. Unde ymaginemur quod talis triangu-
“ Traductio Apollonii per Bernardum.” This was no doubt the same work as that mentioned 15 lus revolvatur infra unam massam ceream; tunc illa porcio de cera quam
in the list of 1522: “ Apollonius de concicis. In Latino." (Zinner, ibid.) There is no trace of this abscindet erit pyramis rotunda rectangula. Et breviter est una consimilis
translation.
30 Zinner, ibid., mentions that in the 1522 list the following title is given: Inicium Apolonii
ex Greco in Latinum traducti," from which he jumps to the conclusion that Regiomontanus Tit., Proemium
began a translation of the Conics. Perhaps he made this inference from the fact that a frag­ 1 [Libellus . . . mukefi] addidi
ment of the first book is found on folio 141r-v of MS Vienna, Nationalbibl. 5203, as Zinner 4 [Proemium] addidi
notes on p. 308. But we have already seen above (in the text over note 21) that this fragment Cap. primum
is in fact that translated by Gerard of Cremona from the Arabic version of the Conics and is 1 [Capitulum primumi addidi
not a composition by Regiomontanus, as Zinner implied. Perhaps it was this fragment that was 7 orthogona fort e orthogonia
mistakenly characterized in the list of 1522 as being translated from the Greek. 8 orthogonus.forte orthogonius

185
186 ARCHIMEDES IN THE MIDDLE AGES FUSORIS DE SECCIONE MUKEFI 187

quasi figura sicut est trocus de quo dicit Cato troco ludo. Hac igitur etiam iuxta hoc quod si a puncto medio linee kb erigatur perpendicularis,
declaratione premissa sumatur una talis piramis rotunda rectangula, cuius tunc ipsa transibit per punctum e . Sic ergo patet quod concursus radiorum
vertex, id est, cuspis, sit a, quam secundum axis longitudinem secet fit ad punctum medium sagitte seccionis mukefi cuius sagitta est dya­
20 superficies plana et sit seccio communis triangulus abc ita quod cb est meter quadrati cuius quadrati latus esset 4a pars dyametri pyramidis rec-
dyameter circuli qui est basis et d centrum eius. Trahatur igitur in eadem 30 tangule illius seccionis. Et sic patet concathenatio et convexio earum et
basi alia dyameter edf orthogonaliter secans in puncto d dyametrum pri­ qualiter unum potest notificare aliud et econtra. Nota quod presupposita
mam bc. Deinde a puncto d trahatur linea dg perpendiculariter ad lineam figuratione hic posita potest probari et declarari propositio proposita per
ac que erit equedistans ab et tunc iste due linee edf et dg declarabunt 39[aml, 40[aml et 41am propositiones 9e (/ 9l?) perspective Witelonis; nec
25 nobis quid sit seccio mukefi, quia ipse sunt in eadem superficie plana oportet ponere aliam probacionem, saltem nisi pro maiori declaratione
secante predictam piramidem; quarum seccio communis, scilicet egf, est 35 media que pro certo prima facie est bene obscura. Notandum quod qui
seccio mukefi, ut apparet ex primo libro, puta ex commento illius pro­ per huiusmodi secciones vellet componere specula mukefi non esset
positionis “ Omnis superficiei plane secantis piramidem rotundam non per necesse ipsam facere secundum totale completum earum sed sufficeret
verticem et superficiei conice pyramidis communem seccionem figuram quod fierent secundum unam portionem earum, puta versus cuspides;
30 triangularem esse est impossible. ’’ Ipsam pono integram quia nescio quota verum est tamen quod quanto portio esset maior tanto esset maioris effi-
sit propositio primi libri. 40 cacie. Verum est etiam quod ipsa possent fieri secundum unam portionem
ipsarum in medio vel circa bases earum et tunc ipsa essent annularia,
Capitulum secundum hoc est, ad modum annuli.

Assignata pyramide rectangula distantiam notificare ad quam valeat Capitulum 3um

sua seccio mukefi radios congregare; et econverso etiam data distantia
Predictarum seccionum mukefi pro parva vel magna distantia deser­
ad quam valeat aliqua seccio radios congregare pyramidem eius assignare.
vientium modum descriptionis seu protractionis declarare.
5 Item etiam pro data distantia propriam seccionem mukefi designare ac
Ista igitur fuit et est huius presentis libelli principalis intentio, scilicet
etiam ipsius seccionis piramidem notificare.
5 investigare qualiter possimus predictas secciones mukefi describere ut
Ista omnia nunc dicta sunt concathenata ita quod ex uno habetur aliud.
scilicet ipsis mediantibus possimus predicta specula mukefi componere.
Pro quorum omni declaratione sit pyramis rectangula Igm [Fig. 5.2], vel
Quas secciones describere, saltem cum prescisione, reputaverunt geo­
mgl proprie loquendo triangulus rectangulus eam describens in revolutione
metres et perspectivi, multum difficile. Unde ego Johannes Fusoris unicum
10 sua supra axem d m , trahaturque linea da equedistans lateri Im, que est
modum eas describendi vidi qui scriptus est in libraria venerabilis coi­
seccio communis predicto triangulo et seccionis mukefi basi (/ super­
to legii cerbone (/ Sorbone?), qui talis est in sententia. Componamus pira­
ficiei) cuius seccionis volo investigare distanciam ad quam valeat facere
midem rotundam rectangulam cuius dyameter sue basis sit quadrupla ad
concursum radiorum. Et ut breviter me expediam dico generaliter quod
distantiam quam elegimus facere reflexionem radiorum, que sit de metallo,
quelibet seccio mukefi facit congregationem radiorum ad punctum me­
ligno vel plastro, vel quacumque alia materia; tamen quanto fiet de ma­
is dium sue sagitte; ita quod [si] sit linea da divisa in duo equalia in puncto
teria firmiori et densiori tanto melius, in qua capiatur seccio mukefi seu
e, tunc dico quod e est punctus concursus radiorum per talem seccionem.
15 parabola, quod idem est iuxta doctrinam 2e propositionis precedentis. Et
Unde trahatur linea da in continuum et directum ita quod ak sit equalis
patet quod ipsa seccio mukefi secat et dividit ipsam piramidem in duas
da\ et presupposita figuratione ut apparet quod linea kb sit contingens
partes. Limetur ergo vel scindatur vel abradatur vel sarretur vel quoque
predictam seccionem in puncto b et cadiat (/) radius tb equedistanter sagitte
alio modo removeatur una illarum partium, alia remanente integra et sana.
204v 20 da supra lineam kb in puncto b / tunc faciliter potest probari quod reflec­
In qua parte remanente integra et sana palam est esse seccionem mukefi,
tetur ad punctum e, qui est punctus medius linee da per ypothesim. Unde
20 scilicet in loco abscisionis. Applicetur igitur illa sectio alicui lamine plane
dico causa brevitatis quod reperientur due linee eb et ek equales et sic
anguli supra basim equales, per 5am primi Euclidis. Modo angulus inci­
dende tbh equalis est angulo k, per 20am primi Euclidis, quare etiam 32 supra figuratione scr. MS scilicet lateris------(?)
25 angulo reflexionis ebk. Et sic patet quod fiet reflexio ad punctum e. Dico Cap. tertium
1 Capitulum 3um mg. MS
Cap. secandum 5 predictas correxi ex predictam(?)
1 Capitulum secundum mg. MS 12 post reflexionem scr. et del. MS ref
11 predicto triangulo correxi ex predict! trianguli quia basi apparet in eadem linea 15 post precedentis scr. et del. MS in qua piramide capiatur seccio mukefi
188 ARCHIMEDES IN THE MIDDLE AGES FUSORIS DE SECCIONE MUKEFI 189

calibee, ferree vel enee, et cum stilo calibeo supra ipsam laminam de­ seccionem pro maiori distantia. Sed tu petes quomodo sciam pro quali
scribatur predicta seccio mukefi trahendo scilicet ipsum stilum supra 65 vel quanta distantia valebit triangulus unus datus. Respondeo quod valebit
illam laminam, sic tamen quod ipse stilus continue tangat predictam partem pro tanta distancia sicut esset medietas linee ac si esset protracta usque
pyramidis in qua est seccio mukefi. Et illud est quod ego Iohannes Fusoris ad planum Ip.
25 vidi de descriptione seccionis mukefi scriptum. Et quia apparuit mihi
aliqualiter tediosum et difficile per istum modum predictas secciones [Capitulum quartum]
describere, saltem magnas, ideo alium modum ymaginatus sum, qui talis
Attento quod Witelo in probando suam predictam principalem 41am
est. Componatur unus triangulus orthogonus ligneus vel ferreus vel quod
propositionem sui 9' libri perspective sue, a qua tamen omnia predicta
melius est, qui sit adc [Fig. 5.3], cuius angulus d rectus et duo latera
habent vigorem et efficaciam, presupponit duas propositiones tanquam
30 da et dc ad invicem equalia, fiatque latus ac quasi duple quantitatis quam
5 demonstratas et probatas ab Appollonio (/) pergeo in libro de conicis
sit distantia ad quam volumus radios congregare, qui triangulus sic aptetur
elementis, quem tamen librum ego Iohannes Fusoris non vidi. Ideo causa
quod cum prescisione et firmiter possit revolvi circa axem ad supra duas
exercitii ut excitetur et pulsetur intellectus et ymaginatio ad reperiendum
5r cuspides in punctis a et d intrantes in re quiescente stabili / et firma.
descriptiones linearum curvarum pro compositione predictorum speculo­
Deinde sub predicto triangulo aptetur superficies plana, firma et quiescens,
rum, etiam supposito quod nunquam fuisset reperta seccio mukefi et etiam
35 que sit, verbi gratia, pl·, aptetur, dico, taliter quod linea ac usque ad ipsam
io nihil supponendo ab Appollonio (/) pergeo, hic premitto unum modum
protracta incidetur perpendiculariter supra ipsam undiquamque relin­
per quem grosso modo potuerunt tales secciones seu linee curve describi,
quendo angulos rectos. Qua superficie plana sic ordinata et situata ymag-
qui modus potuit forte esse occasio et via ad reperiendum adhuc alios
inemur lineam ac trahi usque ad punctum b et ymaginemur totam lineam
modos meliores quia ipse procedit cum demonstratione geometrica, qui
ab revolvi cum triangulo adc revoluto supra suum axem ad, et tunc clare
40 patet quod predicta linea acb describet in predicta superficie seccionem modus talis est.
15 Sit, verbi gratia, linea ae data distantia pro qua volumus speculum
mukefi. Ad propositum igitur descendendo sumatur una lamina plana cali-
componere ad talem distantiam reflectens radios ita quod e erit punctus
bea, ferrea vel enea in qua volumus seccionem mukefi describere et ipsam
concursus radiorum et a erit primus punctus seu initium seccionis seu
aptemus et firmemus firmiter supra predictam superficiem planam quies­
linee curve investigande [Fig. 5.4]. Tracta igitur linea ae indefinite versus
centem. Quo facto fabricemus unum stilum ferreum, puta smhk, tante
d [et versus x] et persumo ah equalem ae. Deinde erigo a puncto e per-
45 longitudinis vel ymmo maioris quam latus trianguli ac, cuius cuspis k
20 pendicularem eb et equalem eh et traho lineam hbk. Deinde item traho
sit calibeus sitque eius superficies inferior hk plana et taliter aptetur supra
tb equedistanter da et sit velut radius incidens supra punctum b linee
latus ac dicti trianguli quod firmiter et cum prescisione possit colare supra
hbk, quem radium dico reflecti ad punctum e, quod patet quia angulus
ipsum accedendo, scilicet, versus a, vel descendendo versus c quando
incidentie tbk equalis est angulo reflexionis ebh, quod probatur quia angu­
oportuit ad hoc, scilicet quod in revolutione predicti trianguli circa suum
lus tbk equalis est angulo h, per 20fam] primi Euclidis, quia linea hb cadit
50 axem ad, cuspis predicti stili continue tangat predictam laminam sibi
205v 25 supra duas lineas equedistantes / hd, bt. Modo angulus h est equalis angulo
describendo seccionem mukefi. Quod ut apparet ex precedentibus valebit
ebh, per quintam primi Euclidis, cum duo latera eh et eb sint equalia per
pro reflexione radiorum ad tantam distantiam sicut est linea dc. Faciemus
ypothesim, et sic patet preassumptum. Sic igitur patet quod seccio quam
autem quod predictus stilus ferreus colabit firmiter et cum prescisione
querimus transibit per a ad punctum b . Sed ut etiam habeamus alia puncta
super latus ac trianguli adc per hunc modum quia ipse fiet uniformis
que sunt inter a et b per que transire debet seccio seu linea curva quam
55 secundum eius longitudinem tam in longitudine quam latitudine; potest
30 querimus modus est iste. Quia signetur quicumque punctus qui placuit
tamen fieri, si placet, magis latus quam altus. Deinde fient duo cramponi
in linea ae, qui, verbi gratia, sit z, facio ai equalem az, erectaque a puncto
ferrei vel enei adherentes lateri ac per quorum medium colabit predictus
z perpendiculari zq indefinite quantitatis, et posito pede circini immobili
stilus cum prescisione. Deinde est etiam quod qui vellet posset bene fieri
in puncto e, alium pedem extendo usque ad punctum i. Deinde duco
negocium sine huiusmodi cramponis quia sufficiunt duo cuspides vel 40r
ipsum usque ad predictam lineam zq quam secet, verbi gratia, in puncto g.
60 elevati perpendiculariter supra latus ac, residuum vero fiet conducendo
cum manibus elevando et deprimendo, cumprimendo modicum fortiter Cap. quartum
predictum stilum supra latus ac predictum et contra predictos cuspides. 1 [Capitulum quartum] addidi
Iuxta quedam notandum quod quanto fiet maior triangulus tanto describet 2 Attento: Actento MS
10 premitto: premicto MS
28 orthogonus forte orthogonius 19 [et versus x] addidi
39 ad correxi ex ab 20 traho2: traho* MS
190 ARCHIMEDES IN THE MIDDLE AGES FUSORIS DE SECCIO NE MUKEFI 191

35 Tunc dico quod seccio quam querimus transibit per punctum g. Unde 75 cionis, tunc dico quod semper linea eb erit equalis linee eh, et sic per
tracta linea ab ipso i per g et cadente radio mg equedistanter ad ea con­ quintam primi Euclidis semper angulus h erit equalis angulo ebh. Et per
similiter et eodem modo probabitur quod reflectetur ad punctum e sicut 20[aml primi Euclidis angulus tbk erit equalis cuilibet ipsorum, qui tamen
probabatur quod radius tb reflectebatur ad ipsum e et per easdem proposi­ angulus tbk est angulus incidentie et angulus ebh est angulus reflexionis.
tiones. Quemadmodum autem inter b et a repertus est punctus g per quem [4.] Quarta conditio sive proprietas est: presupposita figuratione nunc
40 debet transire seccio vel linea curva quam querimus, consimiliter et per 80 posita pro 3a proprietate [Fig. 5.7 iterum], scilicet quod linea recta hbk
easdem propositiones reperientur alia puncta intermedia quotquot placuit contingat seccionem ag etiam ubicumque fuerit, sive prope conum a sive
— voluerimus centum vel mille— quousque possimus pertransire a puncto r in puncto g / sive in medio, verbi gratia, in puncto h, tunc dico, si linea
ad punctum lineam curvam quam querimus. Quemadmodum autem locuti hb dividatur per medium et ab illo puncto medio erigatur perpendicu­
sumus de latere bga consimiliter intelligendum est de alio latere versus/. laris, quod ipsa necessario transibit per punctum e, qui est punctus con-
45 Ex predictis omnibus et premissis pono quedam correlaria sive quasdam 85 cursus radiorum illius seccionis.
propositiones laterales que sunt velut conditiones seu proprietates pre-
dictarum seccionum seu predictarum linearum curvarum, quod idem est. [Capitulum quintum]
[1.] Prima igitur conditio sive prima proprietas seccionis mukefi est: Hic autem volui recitare predictas proprietates seccionis mukefi quia
quod, verbi gratia, sit seccio mukefi lag sitque e concursus radiorum per eas poterit intellectio et ymaginatio excitari ad inquirendum modum
50 [Fig. 5.51. Tunc tracta linea leg dico quod quadratum linee eg est equale seu modos descriptionis seu protractionis cum cuspide calibeo tales sec-
ei quod fit ex tota leg in lineam ea ita quod linea eg est medio loco pro­ 5 ciones mukefi. Verbi gratia, volo unum modum aperire qui mihi pronunc
portionalis inter lineam Ig et lineam e a , et sic per consequens iuxta octa­ occurrit et est talis [cf. Fig. 5.7 iterum]. Queratur modus qualiter linea
vam 61 Euclidis, si sumatur ed equalis Ig et dividatur ad per equalia, perpendicularis zb moveatur uniformiter a puncto e usque ad a et punctus
verbi gratia, in puncto o tunc posito pede circini immobili in puncto o h etiam continue proportionaliter moveatur veniendo usque ad idem punc­
55 extendatur alius pes usque ad punctum a et describatur circulus, tunc tum a, sic tamen quod continue linea za sit equalis linee ha et quod etiam
ipse transibit per punctum g . 10 continue linea eb sit equalis linee eh. Unus autem modus hoc faciendi
Consimiliter dico quod, si ab aliquo puncto linee ea erigatur perpendicu­ qui mihi pronunc occurrit est iste, quem nunc tangam in universali. Primo
laris usque ad predictam seccionem, que, verbi gratia, sit zb, tunc con­ habeatur angulus rectus ferreus cuius unum latus sit linea z b , aliud vero
similiter dico quod quadratum ipsius zb est equale ei quod fit ex linea latus colabit supra lineam dax. Brachium autem predictum zb habebit
60 Ig in lineam z a , ita quod linea zb est medio loco proportionalis inter lineam fissuram per quam colabit cuspis calibeus b continue appropinquando
Ig et lineam z a , et sic per consequens iuxta 8[am] 6’, si sumatur zc equalis 15 versus z. Deinde erunt due alie linee ferree, scilicitcb et hb. Tuncymaginor
Ig et dividatur ac per equalia in puncto i, tunc posito pede circini immobili quod in puncto a sit velud quoddam foramen vel gracilis cuspis per quod
in puncto / extendatur alter usque ad punctum a et describatur circulus; uniformiter trahantur duo puncta z et h versus punctum a. Cuspis autem
tunc dico quod ipse transibit per punctum b. calibeus predictus affixus erit extremitati linee eb, que linea eb habebit
65 [2.] Secunda conditio seu proprietas est: quod si aliqua linea recta, fissuram prope e et colabit in gracili cuspide affixo puncto e secundum
verbi gratia, hbk [Fig. 5.6], contingat seccionem mukefi, verbi gratia, in 20 quod linea eb abbreviabitur per motum ipsius b versus a. Consimiliter
puncto b , etiam ubicumque fuerit, sive prope conum a sive versus g, et etiam linea hb habebit fissuram prope b in qua colabit cuspis calibeus b
a puncto contactus b trahatur perpendicularis bz supra lineam a d , tunc secundum quod linea hb abbreviabitur per approximationem ipsius b ad
dico quod semper linea za erit equalis linee ah. a. Supra vero lineam xa erit una linea ferrea recta cui affixus erit gracilis
70 [3.] Tertia conditio sive proprietas est: sit, verbi gratia, ae distantia cuspis quem cuspidem gracilem amplectetur undiquamque regula hb in
pro qua valet seccio data abg, et supposita figuratione sicut hic est facta 25 puncto h. Deinde adhuc pro maiori securitate posset fieri quod a medio
[Fig. 5.7], scilicet quod linea hbk contingat seccionem in puncto b etiam puncto linee hb erigeretur perpendiculariter regula ferrea que necessario
ubicumque fuerit, sive prope conum a sive versus g sive etiam in puncto transiret per punctum e et continue colaret per quandam fissuram per
g, etiam supposito quod linea Idg sit dyameter basis piramidis date sec-36 punctum e ita quod ibi esset una regula que contineret duos gnomones.
Et pro nunc apparet mihi quod ipsa est valde bona, ymmo quasi neces-
36 ante ab seri, et del. MS ab / post mg scr. et del. MS probabitur quod
reflectetur ad punctum e
40 post seccio scr. et del. MS curva Cap. quintum
41 placuit’ MS 1 [Capitulum quintum] addidi
59 dico . . . ipsius supra scr. MS 23 post lineam scr. et del. MS et
192 ARCHIMEDES IN THE MIDDLE AGES FUSORIS’ ON THE PARABOLIC SECTION 193

30 saria. Unde ipsam faciet quod angulus h continue erit equalis angulo ebb. two lines ab and ac describe the length of the cone, that is [its] curved
Ymmo etiam apparet mihi pro nunc quod ipsa excusat nos a regula zb. surface, while line be describes its base, which last is a circle. Whence
Nota tamen pro nunc quod unus modus approximandi vel elongandi puncta let us imagine that such a triangle be rotated in a waxy mass; then that
z et h ad punctum a sine corda vel cylia esset per unum talem triangulum waxy portion which it cuts out will be a right-angled cone. And, briefly,
de metallo sicut est a 'b 'd ' [Fig. 5.8] et quod punctus a colaret per fis- it is a figure like a top such as that of which Cato speaks as a play-top.
35 suram eius mediam cd' perpendiculariter ad lineam dx ita quod puncta With this explanation premised, let one such right-angled cone be taken,
z (/ h) et h (/ z) continue tangent latera a 'd' et b ' d ' . one whose apex, that is, cusp, is a. Let a plane surface cut it according
Et sic est finis libelli de compositione et protractione vera seccionis to the length of the axis and let the common section be triangle abc,
mukefi. Jo. Fusoris. so that cb is the diameter of the base circle and d is the latter’s center.
Therefore, let there be drawn in the same base another diameter e d f
that orthogonally cuts the first diameter be in point d. Then from point d
let line dg be drawn perpendicular to line ac and thus parallel to ab.
[The Booklet of Jean Fusoris on the Parabolic Section] And then these two lines e d f and dg will make clear to us the nature of
a parabolic section, because they are in the same plane surface that cuts
This booklet concerns the parabolic section by means of which are the aforesaid cone. The common section of these [i.e. the plane and the
composed mirrors that will burn [something] at a great or a small distance, cone], namely egf, is a parabolic section, as appears from the first book
as one pleases. [of Witelo’s Perspectiva], i.e. from the comment of that proposition [which
reads]: “ The common section of any plane surface that cuts a cone not
[Proem] through the apex and of the surface of the cone cannot be a triangular
figure.” I quote this whole [enunciation] because I do not know [its] prop­
By subtle investigation the perspectivists have discovered geometrically
that, among all the figures of mirrors, the parabolic section is the most osition number in the first book.
apt in concentrating rays and [thus] in [producing] combustion because
Chapter Two
of [its] most efficacious union of the rays, and yet that its precise descrip­
tion is most difficult. Accordingly this third booklet (/ the third chapter?) With a right-angled cone assigned, to make known the distance at which
teaches the precise description of the parabolic section by continuous its parabolic section is effective in concentrating the rays; and also con­
protraction in the manner that the circumference of circles would be versely, with the distance at which some section is effective in concen­
described. This booklet will contain four or five chapters. trating the rays given, to assign its cone. Also, to designate the particular
parabolic section [effective] for a given distance and, further, to make
[Chapter One] known the cone of this section.
All of these statements are interconnected, so that from one, another
The first chapter makes clear what is called a parabolic section.
is had. For their complete clarification let there be a right-angled cone
The section mukefi, so-called in Arabic, which in Latin speech is named
Igm [see Fig. 5.2], or mgl, which properly speaking is the right-angled
a seccio parabola or seccio rectangula , is the common section of a right-
triangle describing it (the cone) by its rotation about axis dm. And let
angled cone and a plane surface that cuts it in the manner presently stated.
line da be drawn parallel to side Im , this line da being the common section
But to understand this it is necessary to know beforehand what is a right-
of the aforesaid triangle and the surface of the parabolic section whose
angled cone (or orthogonal cone, which is the same thing). First let an
distance that is effective to produce a concourse of rays I wish to investi­
orthogonal or right-angled triangle abc be conceived, whose angle a is a
gate. And in order that I might briefly get myself ready, I say generally
right angle and each of whose other angles b and c is half of a right angle
that any parabolic section produces a union of its rays at the middle point
so that the two lines ab and ac are equal [see Fig. 5.1]. Let this triangle
of its axis, so that if line da be bisected in point e, then I say that e is
be rotated on [its] axis ad, which axis is at rest. Then such a triangle in
the point of concourse of the rays for such a section. Whence let line
its rotation describes a right-angled cone whose cusp or apex is a, and the3456
da be extended in a straight line so that ak — de\ and with the figure as
it is and it being supposed that line kb is tangent to the aforesaid section
33 post a scr. et del. MS sive sue
34 a'b'd': abd MS
in point b and that ray tb (drawn parallel to axis da) falls on line kb in
35 cd': cd MS point b, then it can be proved easily that it (the ray) is reflected to point
36 a'd': ad MS / b'd': bd MS e, which is the middle point of line da by hypothesis. Whence I say,
194 ARCHIMEDES IN THE MIDDLE AGES FUSORIS’ ON THE PARABOLIC SECTION 195

for sake of brevity, that the two lines eb and ek will be found to be equal, have chosen to produce the reflection of rays.1 We let this cone be made
and so the angles on the base will be equal, by 1.5 of Euclid. Now the of metal, wood, plaster, or any other material (however the more that it
angle of incidence tbh is equal to angle k, by 1.20 of Euclid, and hence is made of a harder and denser material, the better it is). In this material
also equal to the angle of reflection ebk, and so it is evident that the let there be taken a sectio mukefi or parabola, which [terms] are identical
reflection will take place to point e. I also say accordingly that, if from according to the doctrine of the second proposition that precedes [this
the middle point of line kb there be erected a perpendicular, then it will one, that is, the doctrine of the first chapter]. And it is evident that this
pass through point e. Therefore, it is evident that the concourse of the parabolic section cuts and divides the cone into two parts. Therefore,
rays takes place at the midpoint of the axis of the parabolic section, which let one of these parts be filed, cut, scraped, sawed, or removed in some other
axis is the diagonal of a square whose side would be !4 the diameter way, leaving the other part whole and sound. It is clear that a parabolic
of the [base of the] right-angled cone of that section.1 And thus is evident section [now] exists in the part that remains whole and sound, evidently
the interconnection of these [statements] and how they come together in the place [i.e. plane] of cutting [where it constitutes the upper surface
and how one can make another known and contrariwise. Note that with of the part]. Then let this section be applied to some steel, iron, or bronze
the presupposed figuration here posited, the proposed proposition can be plate; and with a steel style let the aforesaid parabolic section be de­
proved and made clear by means of Propositions 39, 40, and 41 of [Book] scribed on the plate by drawing the style over the plate, but in such a
IX of the Perspective of Witelo. Nor is it necessary to pose any other way that the style continually stays in touch with the aforesaid part of
proof except for the greater clarification of what, [though presented] as the cone in which the parabolic section lies. And that is what I, Johannes
prima facie evident [by Witelo], is thoroughly obscure. Fusoris, saw [in the tract] written on the description of a parabolic section.
It should be noted that he who wishes to fashion a parabolic mirror But because it seemed to me a little tedious and difficult to describe
by means of sections of this sort need not necessarily fashion it according the aforesaid sections by this method, or at least [so to describe] large
to their whole section. But it would suffice that they (i.e. the mirrors) ones, therefore I have imagined another method as follows. Let a right-
be made according to one segment of them, as for example toward the angled triangle be made of wood, or of iron, or of anything that is better,
vertexes. It is true, however, that by the amount that the segment [of and let it be adc [see Fig. 5.3], whose angle d is a right angle and whose
the parabolic surface] would be greater, by that amount would it be of two sides da and dc equal one another; and let the side ac be made twice
greater efficacy [for burning, since more rays would be reflected to the as long as the distance at which we wish the rays to focus [the latter dis­
focal point]. It is also true that these [mirrors] could be made according tance measured from the vertex to the focus of the desired section]. Let
to a segment of them in the middle or at the bases, and then they would this triangle be so fastened that it can be rotated with precision and firmly
be annular, that is, in the manner of a ring. about axis ad on its two cusps in points a and d , the latter being inserted
into something at rest that is stable and firm. Then under the aforesaid
Chapter Three triangle is fastened a plane surface that is firm and at rest, which, for
example, we let be p i — fashioned, I say, so that line ac, if protracted to
To clarify the method of the description or protraction of the aforesaid
it, will impinge on it perpendicularly and everywhere remain at right
parabolic sections that serve for [burning at] a small or great distance.
angles to it. With the plane surface thus arranged and situated, let us
Therefore, this was and is the principal intention of this present booklet,
imagine that line ac be extended to point b, and let us [also] imagine
namely, to investigate how we could describe the aforesaid parabolic
that the whole line ab be revolved along with the rotation of triangle adc
sections, evidently so that we could by means of them fashion the afore­
on its axis ad; and then it is clearly evident that the aforesaid line acb
said parabolic mirrors. To describe these sections, at least with precision,
will describe a parabolic section in the aforesaid surface. Then proceeding
geometers and perspectivists have thought to be very difficult. Whence I,
to what has been proposed, let us take a plane plate of steel, iron or
Johannes Fusoris, have seen a unique method of describing them written
bronze in which we wish to draw a parabolic section, and we fasten it
down in the library of the venerable college of the Sorbonne, which method
and place it firmly over the aforesaid plane surface which is at rest. With
is [given] in the [following] description. Let us make a right-angled cone
this done, let us fashion a style of iron, e.g., sm hk, of such length as, nay
the diameter of whose base we let be quadruple the distance which we

Chapter Two
1 This is incorrect. It is rather the slant-height or hypotenuse of the right-angled cone used
1 This is correct for a parabolic section formed from a right-angled right cone where the to generate the parabola that is quadruple the focal distance. I have discussed this error
axis of the parabola is the line intercepted between the vertex of the parabola and the axis and its relationship to a somewhat similar error found in the Speculi almukefi compositio
of the right-angled cone. in the body of my chapter above.
196 ARCHIMEDES IN THE MIDDLE AGES
FUSORIS’ ON THE PARABOLIC SECTION 197
in fact longer than, side ac of the triangle. Let its cusp k be of steel
and let its lower surface hk be flat and so fastened on side ac of the said of Perga, I shall set forth here a method by which, in gross fashion, such
triangle that it could slide firmly and with precision along the side by sections or curved lines could be described. This method could perhaps
approaching [i.e. ascending] toward a or descending toward c as need be the occasion and way for finding still other better methods because
be so that in the rotation of the aforesaid triangle about its axis the cusp it proceeds by geometric demonstration. This method follows.
of the aforesaid style continually touches the aforesaid plate while de­ For example, let line ae be the given distance for which we wish to
scribing on it the parabolic section. [Hence] it appears from the preceding fashion a mirror to reflect rays at such a distance, so that e will be the
[statements] that it will be efficacious for the reflection to as much distance point of concourse [i.e. the focal point] of the rays and a will be the first
as is line d c 2 Now we shall assure that the aforesaid iron style will ride point or beginning of the section or curved line to be investigated [see
firmly and with precision on side ac of triangle adc by means of this Fig. 5.4]. Then with line ae indefinitely extended toward d [and toward
method because it is made uniform throughout its length both in its length x], I take ah to be equal to a e . Then I erect from point e a perpendicular
and in its width. However it can be made, if one pleases, more wide eb equal to eh and I draw line hbk. Then I also draw tb parallel to db
than high. Then let two iron or bronze cramps be made to adhere to side and let it be as a ray falling on point b of line hbk, which ray, I assert,
ac, by means of which cramps the aforesaid style will ride with precision. is reflected to point e. This is evident because the angle of incidence tbk
Then it is also [possible] that he who wishes could accomplish the under­ is equal to the angle of reflexion ebh. This is proved, for angle tbk is
taking without cramps of this sort because two or four cusps elevated equal to angle h, by 1.20 of Euclid, line hb falling on two parallel lines
perpendicularly on side ac [might] suffice [to keep the style aligned] while hd and bt. Now angle h is equal to angle ebh, by 1.5 of Euclid, since
the rest is accomplished by manually guiding [the style] up and down the two sides eh and eb are equal by hypothesis, and thus the earlier
[in its sliding motion], all the while pressing with some strength the afore­ assumption is evident. So, it is therefore evident that the section we seek
said style on the aforesaid side ac and against the aforesaid cusps. Accord­ will pass through a to point b. But in order that we may also have other
ing to certain things it should be noted that the larger the triangle is made points between a and b through which the section or curved line we seek
the greater is the parabolic section described [and thus is efficacious] ought to pass, we use the following method. For with any point z designated
for a greater distance. But you ask how I know for which, and how great in line ae, I make ai equal to az. And having erected from point z a
a, distance a given triangle will be valid. I respond that it will be valid perpendicular zq of indefinite length and having placed the immobile foot
for the distance that would be half the line ac if it were protracted to of a compass in point e, I extend the other foot to point i. Then I draw
plane Ip.3 it (a circular arc) as far as the aforesaid line zq, which it cuts, for example,
in point g. Then I say that the section we seek will pass through point g.
[Chapter Four] Whence, with a line drawn from i through g and with a ray mg falling
[on g] and being parallel to e a , it will be proved that it (the ray) is reflected
It has been noted that Witelo, in proving his principal proposition, to point e in the same way and by the same propositions that it was
no. 41, of Book IX of his Perspective (from which everything said above
proved that ray tb was reflected to e. Now just as there has been found
has force and efficacy), presupposes two propositions as demonstrated between b and a a point g through which the section or curved line we
and proved by Apollonius of Perga in the book On Conic Elements — seek ought to pass, so similarly and by the same propositions any inter­
which book, however, I, Johannes Fusoris, have not seen. Therefore, mediary points of any desired number— a hundred or a thousand— will be
as an exercise to stir and stimulate the intellect and the imagination toward found until we could traverse from point to point the [whole] curved line
finding the description of curved lines [to be used] for fashioning the
which we seek. Now the same thing we have said concerning side bga
aforesaid mirrors, and it having been supposed that a parabolic section [of the curved line] ought to be understood for the other side [of the
had not yet been found, and further supposing nothing from Apollonius
curve] towards/.
2 This is essentially a repetition of the error given earlier in this chapter of the translation
From all of the aforesaid, premised things, I pose certain corollaries
and singled out in note 1 since line dc is equal to Va the diameter of the base circle of or certain collateral propositions which are like conditions or properties
the right-angled cone, the base circle having been formed by passing a plane perpendicular of the aforesaid sections or aforesaid curved lines (these being the same).
to the axial triangle through the point of intersection of the axis of the cone with the axis [1] Therefore [this] is the first condition or first property of a parabolic
of the parabola.
section: For example, let there be a parabolic section lag and let e be its
3 Notice that, unlike the statement made earlier in this chapter of the translation and
discussed above in notes 1 and 2, this statement is quite correct and is equivalent to that
concourse of rays [see Fig. 5.5]. Then, with line leg drawn, I say that the
given in the second chapter of the translation and singled out in footnote 1 of that chapter. square of line eg is equal to the product of the whole line Ig and line ea
[i.e., eg2 = lg-ea], so that line eg is the mean proportional between line
198 FUSORIS' ON THE PARABOLIC SECTION 199
ARCHIMEDES IN THE MIDDLE AGES I
Ig and line ea [i.e., Ig / eg = eg / ea]. And consequently, by VI.8 of continually and proportionally moved and approach the same point a,
Euclid, if ed is assumed equal to Ig and ad is bisected in point o, then, but so that line za be continually equal to line ha and also line eb be
with the immobile foot of a compass placed in o, let the other foot be continually equal to line eh. Now one method of doing this which occurs
extended to point a and let a circle be described; then it will pass through to me at this time is the one which I now describe completely. In the
point g. first place let there be had an iron right angle one of whose sides is line
Similarly I say that, if from some point of line ea a perpendicular zb zb while the other side will ride above line dax. The aforesaid arm zb
is erected [and extended] to the aforesaid section, then zb2 = lg-za, so will have a slot in which a steel cusp b will continually ride as it approaches
that line zb is the mean proportional between line Ig and za [i.e., Ig / zb z. Then there will be two other iron lines (i.e. rules), namely eb and hb.
= zb / za]. And so consequently, according to VI.8 [of Euclid], if zc is Then I imagine that in point a there is a certain hole or slender cusp by
assumed equal to Ig and ac is bisected in point i, then, with the immobile means of which the two points z and h are uniformly drawn toward point
foot of a compass placed in point /, let the other [foot] be extended to a. But the aforesaid steel cusp will be affixed to the extremity of line eb,
point a and let a circle be described; then I say that it will pass through which line eb will have a slot near e and will ride on the slender cusp
point b . affixed to point e according as line eb will be shortened by the motion
[2] The second condition or property : If some straight line, for example of its b toward a. Similarly also line hb will have a slot near b in which
hbk, is tangent to the parabolic section, say in point b [see Fig. 5.6], the steel cusp b will ride according as line hb will be shortened by the
wherever that is— either near the vertex a or toward g — and if from the nearness of its b to a. Now on line xa there will be an iron straight line
point of tangency b a perpendicular bz is drawn to line ad, then I say (i.e. rule) to which will be affixed a slender cusp, which slender cusp the
that line za will always be equal to line ah. rule hb embraces on both sides in point h . Then for still better assurance
[3] The third condition or property: For example, let ae be the [focal] [the same thing] could be accomplished by erecting perpendicularly from
distance for which the given section abg is valid; and with the figuration the middle point of line hb an iron rule which necessarily would pass
supposed as here made [see Fig. 5.7], that is, where line hbk is tangent through point e and would continually ride, by means of a certain slot,
to the section in point b , wherever it is— whether near vertex a or toward past point e, so that there would be a rule which would contain two
g or even in point g — and it having been supposed that Idg is the diameter gnomons. And it appears to me at this time that this [last rule] is very
of the base of the cone of the given section, then I say that line eb will good, nay almost necessary. Whence this will insure that angle h will
always be equal to line eh, and so, by 1.5 of Euclid, angle h will always be continually equal to angle ebh. Nay, it also appears to me at this time
be equal to angle ebh. And, by 1.20 of Euclid, angle tbk will be equal that this [last rule] exempts us from [needing] rule z b .
to each of these [angles]. But angle tbk is the angle of incidence and angle Note, however, at this time that one method of bringing points z and h
ebh is the angle of reflection. closer to or farther from point a without a cord or a thread would be to
[4] The fourth condition or property: With the figuration just posited use a metal triangle like a'b'd' [see Fig. 5.8] in which point a would ride
for the third property [again] proposed [see Fig. 5.7 again], evidently through its middle slot cd', which slot is perpendicular to dx, so that
[so] that straight line hbk is tangent to section ag [at some point], wherever points h and z will continually touch sides a'd' and b'd' [respectively].
that point is— whether near vertex a or in point g or in between, e.g. in And so ends the Booklet on the Composition and True Protraction
point b — then I say that, if line hb is bisected and from the midpoint of a Parabolic Section by Johannes Fusoris.
a perpendicular is erected, it will necessarily pass through point e,
which is the point of concourse of the rays of that section.

[Chapter Five]
Now I have wished to recite the aforesaid properties of a parabolic
section because by means of them a conception and imagery can be
devised for seeking out a method or methods of describing or protracting
such parabolic sections with a steel point. For example, I wish to lay out
one method which occurs to me at this time [see Fig. 5.7 again]. This
method follows. Let a method be sought by which perpendicular line zb
may be moved uniformly from point e up to a and point h may also be
 NOTES OF GIOVANNI FONTANA 201

B. duabus rectis et una curva linea. Et quelibet earum est facta vere secundum
35 artem et extracta ex pyramide ortogonia, et sic potest esse forma pro
Notes for Alhazen’s De speculis comburentibus formando speculo de calibe, ut docetur in tractatu, per revolutionem
as prepared by Giovanni Fontana forme circa sagitam, que ideo axis dicitur.
[6] Si vis scire ad quantam distantiam potest comburere speculum
ex aliqua istarum formarum extractum, accipe longitudinem 4° partis sui
40 lateris recti infra notati et computa in sagita a puncto g versus b [Fig.
i34r / Dicta atque notata per me Johannem
5.9(c)] et ubi terminatur illa longitudo ibi est punctus concursus radiorum
fontana physicum venetum et et generationis ignis; et est in 2a maior quam in prima et in 3[a) quam
probata in 2a longitudo.
[7] Et illa sectio pyramidis que facta est per superficiem planam que
[1] Pro declaratione speculi ardentis quedam sunt principia declaranda. 45 continet lineam eccentricam et sagitam modo predicto est sectio mukefi
5 [2] Pyramis orthogonia rotunda est que fit ex ductu trianguli ortogonii vel parabola vel ortogonia, quod idem est. Et hec sunt que declarare
duorum equalium laterum, uno fixo manente alioque circumducto. Et latus voluimus; reliqua habentur in tractatu.
illud fixum eficitur axis pyramidis tendens a puncto pyramidis super baxem
perpendiculariter in centrum circuli qui per secundum latus describitur.
[3] Superficies plana dividens pyramidem per medium et transiens per Statements and Notations Made and Proved by me
io axem est triangulus rectangulus cuius baxis est dupla ad latus et est tri­ Giovanni Fontana, Venetian Physician
angulus duplus ad illum qui descripsit pyramidem.
[4] Superficies vero plana dividens pyramidem non transiens per conum [1] Certain principles should be made clear for the explanation of a
pyramidis et secans superficiem triangularem immediate antedictam burning mirror.
ortogonaliter ita ut linea recta communis sectionis ipsarum sit equidistans [2] A right-angled cone is that which arises from the movement of a
15 lateri superficiei triangularis antedicte est superficies duabus lineis com­ right triangle when one of [its] two equal sides remains fixed and the other
prehensa, scilicet curva que dicitur linea eccentrica (quia non habet cen­ is rotated. And that fixed side forms the axis of the cone and proceeds
trum unum) vel linea mumiami (/) et alia recta que est in baxe pyramidis, from the apex of the cone perpendicularly upon the base to the center
et illa linea que est communis in sectione harum duarum superficierum of the circle which is described by the second side.
planarum dicitur axis huius superficiei que continet lineam eccentricam, [3] The plane surface that bisects the cone and passes through the axis
20 vel ipsius sagita, que linea vel sagita necessario secatur ab axe pyramidis is a right triangle whose base is double the [base] side [of the initial
si quelibet earum in continuum et directum protrahatur, et punctus in quo triangle], and the [axial] triangle is double that which describes the cone.
se secant est punctus communis; linea recta quecunque in superficie plana [4] But the plane surface that divides the cone and does not pass through
que continet lineam eccentricam dictam orthogonaliter secans sagitam the apex of the cone but cuts the immediately aforesaid triangular surface
dicitur linea ordinis et illa que transit per punctum medium sagite cum orthogonally so that the straight line which is the common section of these
25 fuerit a sagita a capite sectionis usque ad axem pyramidis vocatur latus [two planes] is parallel to one side of the aforesaid triangular surface is
rectum, de quo fit mentio pro habenda distantia ad quam speculum comburit. a surface contained by two lines: one a curved line which is called an
[5] Formas pro speculo fiendo sic accipies. Invenias in pyramide rec- “ eccentric line” because it does not have one center, or [it is called]
tangula, ut supra, superficiem planam secundam, quia scilicet illam que a “linea mumiani (/ munani),” and the other a straight line which is in
continet lineam eccentricam et sagitam; et latus rectum etiam invenias the base of the cone; and that line which is common in the section of
30 ad ipsam. Et accipe longitudinem 4e partis lateris recti et serva; post these two plane surfaces is called the “ axis” of this surface which con­
reseca de ista superficie medietatem et serva aliam medietatem in qua tains the eccentric line, or [it is called] its “arrow.” This line or arrow
sit tota sagita, medietas baxis et medietas linee eccentrice, sicut ap[p]aret is necessarily cut by the axis of the cone if each of these [lines] is extended
in istis tribus formis triangularibus [Fig. 5.9] quarum quelibet continetur continually in a straight line; and the point in which they intersect each
other is called the “common point.” Any straight line at all in the plane
4 [1] et sequentes numeros addidi surface containing the said eccentric line that orthogonally cuts the arrow
7 puncto correxi ex puncta / baxem corr. MS ex baxi is called a “line of order” , and that one which passes through the middle
15-16 comprehensa corr. mg. MS ex comprensa
25 usque . . . pyramidis mg. MS 34 quelibet earum mg. MS

200
202 ARCHIMEDES IN THE MIDDLE AGES
c.
point of the arrow drawn from the vertex of the section to the axis of the De speculis comburentibus (= Speculi
cone is called the “ latus rectum,” of which mention is made in determining
the distance at which the mirror burns. almukefi compositio) as prepared
[5] You will obtain forms for making the mirror as follows. You will by Regiomontanus
find in a right-angled cone (as above) the second plane surface because it
is the one which contains the eccentric line and the arrow; and you will
also find the latus rectum for it [i.e. the section]. And take the length 27 r / De speculis comburentibus
of one quarter part of the latus rectum and retain it. Afterward cut away
[Praefatio auctoris]
one half from this surface and retain the other half, in which latter is
contained the whole arrow, half the base, and half the eccentric line, as [Q]uoniam omnes quos de speculis comburentibus ad certam distantiam
it appears in the three triangular forms [in Fig. 5.9], each of which is tractare perpendi duas conclusiones ab Apolonio Pergeo allegatas sup-
contained by two straight lines and one curved line. And each of these 5 ponunt super quibus tamen tota eorum intentio fundatur— ipse etiam
[forms] is accurately made by following the art and produced from a right- Vitelo auctor perspective unum librum premittit in quo demonstrare niti­
angled cone. And in this way can exist a form for making a steel mirror tur omnes conclusiones sibi necessarias atque preambulas ad suum pro­
(as is taught in the treatise [On Burning Mirrors]) by the rotation of the positum ut totus suus liber sufficeret sibiipsi, predictas tamen conclusiones
form around the arrow, which latter is therefore called the “ axis.” Apolonii supponit tamquam notas sed non probat— et licet veritatis earum
[6] If you wish to know at how great a distance a mirror that is pro­ io notitia me diu latuerit non obstante quod diligentiam magnam adhiberi-
duced from one of these forms can burn, take the length of one quarter mus pro demonstratione earundem obtinenda, finaliter tamen ad verum
part of its latus rectum noted below [in the figure] and measure it in the intellectum unius earum cum labore perveni. Deinde circa probationem
axis from point g toward b [see Fig. 5.9(c)], and where that length is alterius donec divino influxu sapiam insistere non tardavi. Probationibus
terminated there is the point of concourse of the rays and of the gen­ itaque earum scitis me tam sollicitus labor meus perderetur cogitavi eas
eration of fire. And in the second [form] the length is greater than in the is in scriptu redigere et mediantibus eis et aliis infra dicendis speculum
first, and in the third [greater] than in the second. comburens ad quantamlibet libueret distantiam fabricare docebo. Premit­
[7] And that section of the cone which has been made by means of the tende tamen sunt alique descriptiones et conclusiones.
plane surface that contains the eccentric line and the axis in the aforesaid
manner is called a “sectio mukefi” or “parabolic section” or “ right- [Definitiones]
angled section,” all of which are the same. 1. [P]iramis rotunda est transitus trianguli rectanguli, altero suorum
And these are the things we wished to explain. Everything else is in laterum rectum angulum continentium fixo, et triangulo ipso donec ad
the treatise [On Burning Mirrors].
Tit. 1 De. . . . comburentibus: De sectione conica, Orthogona, quae parabola dicitur ed
2 Praefatio auctoris ed, om.R
3-17 [Q]uoniam. . . . conclusiones: Cum viderem omnes qui in Perspectiva aliquid
scripserunt, ubi de Speculis ad certam distantiam comburentibus disputant,
duas conclusiones ab Apollonio Pergaeo allegatas praesupponere, quibus
reliqua omnia innituntur, adeo ut etiam ipse Vitellio, cum peculiarem
librum praemittat, ubi omnes propositiones illi necessarias, omniaque
lemmatia (/) quibus post utitur demonstrat, ut opus universum seipso
contentum esset, praedictas tamen conclusiones supponat tanquam necessarias,
minime vero demonstrat: nulla inquirendi defatigatione reiectus fui, quin ad
veram intelligentiam unius earum, ac tandem ad utriusque cognitionem per
demonstrationem pervenirem. Ne igitur hoc labore meo mihi soli profuisse
videar, decrevi cogitationes has meas mandare literis, earumque interventu ac
aliorum quae subnectam Speculum ustorium ad quantamlibet distantiam qua
ratione fabricandum sit demonstrare. Primum vero a definitionibus exordiamur
ed.
13 divina R
Def.
1 Definitiones ed, om. R
2 1. R, om. ed
203
 REGIOMONTANUS’ VERSION 205
204 ARCHIMEDES IN THE MIDDLE AGES
Licet in sectione mukefi quecunque linea possit vocari basis recta seu
locum unde moveri cepet redeat circumducto. Si tunc latus fixum lateri latus rectum sectionis mukefi intercepte inter dictam lineam ordinis et
5 circumducto fuerit equale, erit piramis rectangula. Si vero longius, acuti- 40 caput sectionis, non tamen ab Apolonio sic vocatur. Ymmo hec est vera
angula. Et si brevius, obtusiangula. Hoc satis patet in principio undecimi 28r eius descriptio: Sectionis mukephi (/) cuius tantum / descenderit sagitta a
elementorum Euclidis. capite sectionis quousque ex parte basis cum axe piramidis concurrerit
2. In omni rotunda piramide orthogonia secta per superficiem planam linea ordinis transiens per punctum medium dicte sagitte vocatur latus
non transeuntem per caput piramidis nec equedistantem basi, differentia erectum illius sectionis mukephi (/) quantumcunque illa sectio ex parte
io communis dicte superficiei et dicte piramidi vocatur sectio quam continet 45 basis longa fuerit sive curta. Unde patet ex hoc de omni sectione mukefi
linea muniani. Hec descriptio ab Apolonio allegatur. quod illa sectio quecunque eius pars intercepta inter aliquam lineam ordi­
3. Quando secatur piramis rotunda orthogonia et rectangula cum duabus nis et caput sectionis ac inter partes linee muniani singule tales partes
superficiebus planis, quarum una per caput et per centrum basis transit, sunt sectiones mukefi et omnes tales idem obtinent latus erectum. Et erit
27v secans piramidem secundum triangulum, / ut infra patebit, et altera non illud latus erectum necessario duplum ad sagittam sectionis mukefi pro­
15 transit per caput— ymo secat eam cum superficie quam continet linea so tensam a capite illius sectionis quousque concurrat cum axe piramidis,
muniani iuxta descriptionem premissam sic quod una dictarum super- ut in 4ta conclusione deo favente probabitur. Istis quatuor descriptionibus
ficierum planarum secantium stet super alteram secundum rectos angulos, premissis secuntur conclusiones.
tunc linea recta que est differentia communis dictis duabus superficiebus
secantibus necessario vel secabit unum duorum laterum trianguli et eque-
20 distabit alteri, et tunc vocatur sectio mukefi; vel secabit unum latus tri­ [Conclusiones]
anguli et concurret cum altero latere in infinitum protracto, quod con­ l a. IN OMNI ROTUNDA PIRAMIDE ORTHOGONIA SECTA IN
tingere potest dupliciter, vel ex parte coni, et tunc sectio quam continet DUO EQUALIA CUM SUPERFICIE PLANA TRANSEUNTE PER
linea muniani vocatur sectio addita; vel secabit ex parte basis, intra vel CAPUT PIRAMIDIS ET CENTRUM SUE BASIS DIFFERENTIA
extra, et tunc (si non sit circulus) vocatur sectio diminuta. Iterum dicta 5 COMMUNIS PIRAMIDI ET DICTE SUPERFICIEI SECANTI ERIT
25 linea que est differentia communis dictis superficiebus quarum altera est TRIANGULUS RECTILINEARIS CUIUS BASIS ERIT DYAMETER
sectio mukefi vocatur axis sive sagitta mukefi cuius extremitas versus BASIS PIRAMIDIS ITA SECTE.
conum piramidis vocatur caput sectionis mukefi et caput linee muniani. [mg.\ Vera etiam est de omni pyramide rotunda sive sit recta sive
Et omnes linee protracte orthogonaliter super dictam sagittam ab uno obliqua.]
latere linee muniani usque ad alterum eius latus erunt equedistantes et 10 Ista conclusio statim se offert intellectui, ex prima scilicet descriptione
30 dividuntur per dictam sagittam in duo equalia et vocantur singule linee piramidis per triangulum circumductum. Et sic patet veritas illius quod
sic protracte linee ordinis illius sagitte. Omnes isti termini ab Apolonio descriptio tertia supponit, et vocantur latera dicti trianguli ypotenuse
summuntur. Et licet triplex sectio inter sectiones non transeuntes per caput piramidis, ut per primam conclusionem de curvis superficiebus patet.
piramidis describatur non ex eis ad propositum nostrum indigebimus nisi
sectione mukefi, ut patebit. Sectio rectangula, parabola et mukefi idem 2a. OMNIUM DUORUM PUNCTORUM IN SUPERFICIE PIRA­
35 sunt; item sectio ambligonia, yperbola et mukefi addita idem sunt; item MIDIS ROTUNDE EXISTENTIUM LINEA RECTA INTER DUO
sectio oxigonia, eclipsis (/ ellipsis) et mukefi diminuta idem sunt— secun­
dum auctores.
39 mukefi: parabolae ed
8 2. R, om. ed 40 Apollonio ed hic et ubique
11 post linea add. ed Arabice dicta 40-41 Ymmo . . . descriptio: sed quae sic describitur ed
12 orthogonia om. ed / cum om. ed 41,44,45
20 mukefi: parabola ed 48,49 mukephi {or mukefi): parabolae ed
23 post addita add. ed sive hyperbola
Conci. 1
24 post diminuta add. ed sive Ellipsis
1 [Conclusiones] addidi
26 mukefi1: parabola ed / mukefi2: parabolae ed
2 la: PROPOSITIO PRIMA ed
27 mukefi: parabolae ed
4 sue om. ed
29 erunt: erunt mutuo ed
6 rectilineus ed
32 sumuntur ed
8 -9 Vera . . . obliqua mg. R ,o m . ed
33 non ex: non tamen ed / ad . . . nostrum: hoc loco ed
34\38 mukefi: parabola ed Conci. 2
34-37 Sectio . . . auctores om. ed 1 2a: PROPOSITIO SECUNDA ed
37 post auctores scr. et del. R Non tamen ab Apolonio sic vocatur
206 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ VERSION 207

3“ Euclidis e f est linea continens circulum et centrum eius est punctus g,

PUNCTA PROTRACTA, SI IN INFINITUM PROTENSA NON TAN-
GAT CONUM PIRAMIDIS, TUNC DICTA LINEA VADET NECES- quod erat probandum.
5 SARIO INTRA PIRAMIDEM. 4a. IN OMNI SECTIONE PARABOLA LATUS ERECTUM DUPLUM
Sint in piramide abc duo puncta d, e inter que procedatur linea de EST NECESSARIO AD SAGITTAM, IDQUE LATUS ERECTUM AD
que ulterius quantumcunque protensa non transeat per conum piramidis OMNEM PERPENDICULAREM EXEUNTEM A SAGITTA USQUE
[Fig. 4.5]. Dico lineam de ingredi piramidem. AD CURVUM LATUS SECTIONIS SE HABEBIT SICUT ILLA PER-
Protraham a cono a per duo puncta d et e duas lineas rectas usque 5 PENDICULARIS SE HABET AD PARTEM SAGITTE INTERCEPTAM
io ad circumferentiam basis, que sint adb, aec, et coniungam c et b per INTER DICTAM PERPENDICULAREM ET CAPUT PROPOSITE
lineam rectam, que necessario secabit circulum per 2am 3HEuclidis. Tri­ SECTIONIS.
angulus ergo abc secat piramidem. Sed in eo est linea de. Igitur ipsa est In piramide adb [Fig. 4.7] secta sectione mukefi, sectio mukefi sit p c f
28v intra piramidem, cum totus / iste triangulus sit intra piramidem. Hoc (/ p f c ), cuius sagitta f o concurrat in puncto o cum axe piramidis ao.
etiam patet, signato / i n linea de et protracta linea a f usque quo occurrat io Per h medium sagitte transeat linea ordinis p h g orthogonalis super sagit­
15 cb in puncto g, hec enim linea tota est intra piramidem, quare et linea tam, que quidem linea ordinis vocatur latus erectum dicte sectionis per
de est intra piramidem. quartam descriptionem. Dico quod linea p g est dupla ad sagittam f o et
29r quod pg / se habet ad perpendicularem mn exeuntem a sagitta ad latus
3a. IN OMNI ORTHOGONIA ROTUNDA PIRAMIDE SECTA CUM
curvum sectionis sicut eadem perpendicularis mn se habet ad m f partem
SUPERFICIE PLANA EQUEDISTANTE BASI SECTIO COMMUNIS
15 sagitte interceptam inter dictam perpendicularem et caput sectionis.
CURVE SUPERFICIEI PIRAMIDIS ET SUPERFICIEI SECANTI
Secet triangulus abd propositam piramidem in duo media transiens per
ERIT LINEA CONTINENS CIRCULUM CUIUS CENTRUM ERIT
5 IN AXE PIRAMIDIS. caput eius et per eius centrum sic quod iste triangulus stet super sectionem
mukefi secundum angulos rectos, prout vult 3a descriptio. Cum igitur in
[mg.: Hec propositio 3a etiam vera est de omni piramide rotunda cuius
proposito linea p g necessario stabit perpendiculariter super dictum tri-
basis est circulus, sive sit recta sive obliqua.]
20 angulum, ymaginor circulum transire per puncta g et p dividentem pirami-
Sit piramis rotunda abc [Fig. 4.6], cuius basis centrum sit d et axis
dem equedistanter basi piramidis cuius dyameter sit khl. Ergo quadratum
piramidis ad\ seceturque cum superficie equedistante basi que faciet in
linee hg erit equale quadrangulo quod fit ex kh in hi, per 30mam 3U et
io superficie piramidis lineam curvam per precedentem, que sit linea ef.
8vam sexti Euclidis. Cum ergo quadratum oc linee, que est semidiameter
Dico quod linea e f erit circumferentia circuli et eius centrum est super
basis piramidis erecta orthogonaliter super dyametrum bd, resultet eadem
axem piramidis.
25 ratione ex ductu bo in od, et linee hl et od sint equales per 34tam primi.
Signentur enim in linea e f duo puncta e e t / ubicumque, per que ab
Ergo quadratum oc fit ex ductu hl in b o . Et quia quadratum gh resultat
a cono ad circumferentiam basis ducantur linee aeb, afc. Aut axis ad
ex hl in kh, productorum autem et producentium eadem est proportio,
15 ibit per centrum superficiei plane secantis [et] sic habetur propositum
ut elici potest ex prima 6U in quantitatibus continuis, erit proportio bo
aut non. Tunc signetur g punctus; et ductis ge et g f et db et dc lineis
ad kh sicut quadrati oc ad quadratum hg. Sed bo ad kh est sicut f o ad
erit per 4tam 6U proportio bd ad eg sicut cd ad f g et permutatim sicut bd
30 f h , ex 4a 6li. Ergo proportio quadrati oc ad quadratum hg est sicut pro­
ad cd ita cg ad f g . Sed bd et cd sunt equales quia semidiametri basis.
portio/? ad f h . Et consimiliter probari potest de omnibus duabus per­
Igitur eg et f g erunt equales. Et consimiliter probabitur de quibuslibet
pendicularibus exeuntibus a duobus punctis sagitte usque ad latus curvum
20 aliis lineis protractis a puncto g ad lineam curvam ef. Ergo per novam

6 procedatur: ducatur ed Conci. 4

10 sint: sunt ed 1 4a: PROPOSITIO IIII. ed
14 signata ed 2 necessario om. ed
Conci. 3 8 mukefi1,2: parabolae ed
1 3a: PROPOSITIO TERTIA ed / cum om. ed. 10 pghg ed
6 -7 Hec . . . obliqua mg. R.om. ed 15 post et dei R et
9 cum om. ed 18 mukefi: parabolae ed
15 et ed, om. R 20 imaginabor ed / transiisse (/) ed I p: d ed
17 bd: db ed 22 ex: per ed
17- 18 et . . . fg om. ed 27 proportio: propositio ed
19 consimiliter: similiter ed 31 probari potest tr. ed
208 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ VERSION 209

figure mukefi quod linea sagitte intercepta inter perpendicularem remotio­ 70 Patet ergo conclusio secundum utramque sui partem, cuius secunda
rem et verticem figure mukefi se habet ad lineam sagitte interceptam pars fuit conclusio Apolonii una. Pro probatione vero alterius sunt alique
35 inter perpendicularem propinquiorem et verticem sicut quadratum per­ preambule conclusiones premittende.
pendicularis remotioris ad quadratum perpendicularis propinquioris. Cum
igitur ex premissis o f ad h f sit sicut quadratum oc ad quadratum hg, o f 5a. PROTRACTA IN SECTIONE MUKEFI QUECUNQUE LINEA
autem est dupla ad h f ex ypotesi, ergo quadratum linee oc erit duplum ORDINIS INTER LATUS EIUS ERECTUM ET CAPUT SECTIONIS
ad quadratum linee hg. Igitur per correlarium 17me 6U oc se habet ad hg DESCRIPTOQU E CIRCULO TRANSEUNTE PER CAPUT SEC-
40 in medietate duple proportionis, oc autem et od sunt equales cum sint 30r TIONIS / ET PER EXTREMITATES DICTE LINEE ORDINIS, CEN-
semidiametri eiusdem circuli. Igitur od ad hg se habet in medietate duple. 5 TRUM CIRCULI SIC DESCRIPTI ERIT NECESSARIO IN SAGITTA,
Similiter ad se habet ad od in medietate duple, per 46tam primi et cor- INSUPER ET DIAMETRI EIUS TRANSEUNTIS PER CAPUT SEC­
29v relarium 17mi 6U. Ergo ad / est dupla ad hg. Sed et eadem ad est dupla TIONIS ET PER CENTRUM DICTI CIRCULI SIC DESCRIPTI PARS
ad f o eo quod ad ad /o sicut bd ad bo, per 4tam 6li, et bd dupla ad bo. INTERCEPTA INTER DICTAM LINEAM ORDINIS ET CIRCUM­
45 Quare per 7mam 5U hg et f o sunt equales; pg autem est dupla ad hg. FERENTIAM CIRCULI EX PARTE BASIS PIRAMIDIS ERIT NECES-
Ergo pg est dupla ad f o . Patet igitur prima pars conclusionis proposite 10 SARIO EQUALIS ERECTO LATERI SECTIONIS.
cum pg sit latus erectum sectionis et f o eius sagitta. Et ex consequenti Quod enim circulus possit per tria puncta protrahi patet per doctrinam
sequitur quod latus erectum sectionis necessario sit quadruplum ad medi­ 5te 4li. Et quod centrum necessario erit in sagitta liquere potest ex cor-
etatem sagitte, scilicet pg quadrupla sit ad f h . relario prime 3U. Et quod pars dyametri circuli sic protracti intercepta
50 Nunc ad secundam partem conclusionis probandam, scilicet quod pg inter verticem et perpendicularem, que est medietas protracte linee ordi-
latus erectum sectionis se habet ad perpendicularem mn sicut mn se habet 15 nis, se habet ad illam perpendicularem sicut ipsa perpendicularis ad residu­
ad mf, describo circulum transeuntem per puncta m et n dividentem pira- um dyametri patet per 8vam 6U. Sed per precedentem pars dyametri inter
midem equedistanter basi cuius dyameter sit linea rms. Cum igitur ms verticem et perpendicularem se habet ad perpendicularem sicut ipsa per­
et hi, per 33mprimi Euclidis, sint equales et ex ductu ms in rm fiat quadra- pendicularis se habet ad latus erectum sectionis mukefi. Ergo latus erec­
55 tum linee mn atque ex hi (equali ms) in kh fiat quadratum linee hg, ergo, tum et pars dyametri antedicta, scilicet intercepta inter perpendicularem
ut supra, eadem est proportio kh ad rm que quadrati linee hg ad quadratum 20 et circumferentiam a parte basis piramidis, necessario sunt equales.
linee m n, kh autem ad rm est sicut f h ad f m . Quare f h ad f m sicut quadra­ [mg.: Non solum veritatem tenet de lineis ordinis inter caput et latus
tum linee hg ad quadratum linee mn. Et si sic, ergo per 15am 5li eadem erectum, verum etiam de eis que sunt post latus erectum et post concursum
est proportio p g, quadruple ad f h ut ostensum est, ad f m sicut quadrati axis piramidis cum sagitta sectionis. Ex hoc infertur quod quanto plus
60 p g , quadrupli ad quadratum hg per correlarium 17me 6U, ad quadratum sectio protenditur a capite tanto plus linee ordinis crescunt.]
linee mn. Cum autem per idem correlarium proportio quadrati pg ad
6a. IN SECTIONE MUKEFI PROTRACTA EIUS SAGITTA EX
quadratum mn sit dupla ad proportionem linee pg ad lineam m n, per
PARTE UTRIUSQUE EXTREMI QUANTUMLIBET SIGNATA[QUE]
suppositionem 10am 5U ergo per eandem proportio linee pg ad lineam f m
LINEA RECTA DICTAM SECTIONEM IN ALIQUO PUNCTO EX-
est dupla ad proportionem linee pg ad lineam mn. Quare mn est medio
TRINSECO LATERIS SUI CURVI CONTINGENTE, PROTRACTA
65 loco proportionalis inter pg et f m . Eadem est igitur proportio linee p g , que
5 QUOUSQUE EX PARTE CAPITIS CONCURRAT CUM SAGITTA,
est latus erectum sectionis, ad lineam mn perpendicularem que est eiusdem
SI A PUNCTO CONTINGENTIE DESCENDERINT DUE LINEE
mn ad f m , quod erat secunda pars conclusionis probanda.
USQUE AD SAGITTAM QUARUM ALTERA SIT PERPENDICU­
[mg.: Simili via idem ostendes quando linea ordinis mn etiam ponatur
LARIS SUPER SAGITTAM ET ALTERA PERPENDICULARITER
post sagittam f o , ut rubee linee < m 'n ' et x s> ostendunt.]
EXEAT A LINEA CONTINGENTIE, TUNC EADEM ERIT PRO-
33,34 mukefi: parabolae ed io PORTIO LATERIS ERECTI SECTIONIS AD PARTEM SAGITTE
35 et verticem: vertici ed
39 hg1: gh ed Conci. 5
49 ante ad scr. et del. R f 1 5a: PROPOSITIO QUINTA ed I mukefi: parabolae ed
55 ex hl: phl ed 17 verticem: f ed / et: et inter ed I sicut: sic ed
58 post 5“ scr. et dei. R pg ad mn 18 mukefi: parabolae ed
59 ante fh del. R h (?) 20 necessario: mutuo ed
63 suppositionem 10am: propositionem 11. ed / per eandem om. ed (et debet dei.) 21-24 Non. . . . crescunt mg. R. om. ed
67 quod: quae ed Conci. 6
68-69 Simili . . . ostendunt mg. R, om. ed 1 6a: PROPOSITIO VI. ed / mukefi: parabola ed
69 m'n' et xs addidi 2 signata(que): signaat ed
210 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ VERSION 211

INTERCEPTAM INTER EXTREMA DICTARUM PERPENDICU­ et perpendicularis exiens a d super f a sit de. Dico ea et ag lineas esse
LARIUM SICUT PARTIS SAGITTE INTERCEPTE INTER PER­ equales, est enim vera eius conversa: si ag et ea sunt equales [et ed est
PENDICULAREM PROPINQUIOREM CAPITI ET PUNCTUM CON­ 15 perpendicularis super ag], linea gd transiens per punctum d contingit
CURSUS LINEE CONTINGENTIE CUM SAGITTA AD PARTEM sectionem in d puncto.
15 SAGITTE INTERCEPTAM INTER EANDEM PERPENDICULAREM Quia si non, tunc secabit sectionem per conclusionem secundam vel
ET VERTICEM SECTIONIS. ergo inter d et a vel inter d et basim sectionis. Non autem inter d et a,
Sit sectio mukefi abc' [Fig. 4.8], super cuius sagittam a f descendat quia sit k per adversarium talis punctus, ducatur etiam kl perpendiculariter
perpendicularis de et in puncto d sectionem contingat linea recta dg, 20 super sagittam; quia igitur ga et ae ex ypotesi sunt equales, quare ga maior
que concurrat cum sagitta ulterius protracta in g, et a puncto d exeat quam al. Sit excessus ergo gm. Unde gm et le erunt equales atque al
20 perpendicularis super sagittam, que sit de, et ex eodem puncto elevetur et am equales, et per consequens ml dupla ad al. Sed, ut patuit ex proba­
perpendicularis super contingentem d g , que concurrat cum sagitta in h . tione 4te conclusionis, ea ad al proportio est dupla ad proportionem ed
Dico quod latus erectum sectionis se habeat ad he sicut eg se habet ad ad Ik, et sicut ed ad Ik ita ex 4a 61 eg ad Ig. Igitur ea ad la proportio
ea. Nam angulus gdh est rectus a quo descendit perpendicularis de super 25 est dupla ad proportionem eg ad Ig . Sed sicut ea ad la ita duplum unius
30v basim / trianguli, ergo per 8vam 6li he ad ed sicut ed ad eg. Sed per ante- ad duplum alterius per 15mam 5li. Ergo eg ad Im proportio est dupla ad
25 premissam latus erectum sectionis ad ed sicut ed ad ea. Igitur quadratum proportionem eg ad Ig. [Sic Ig erit medio loco proportionalis inter eg
de tam resultat ex ductu lateris erecti in ea quam ex ductu he in eg. et Im]. Quare, per I0mam diffinitionem 5U, sicut eg ad Ig ita Ig ad Im. Cum
Igitur per 15mam 6U latus erectum se habet ad he sicut eg ad ea, quod igitur eg totum se habet ad Ig totum sicut idem Ig pars totius eg ad Im
fuit propositum. 30 partem totius Ig, ergo per 19vam 5U el residuum totius eg se habet ad mg
[rag.(30r): Ex hoc et sequenti manifestum est quod pars sagitte inter 3ir residuum totius Ig sicut eg / totum ad Ig totum. Sed eg est maior quam
30 extremitates perpendicularium dictarum intercepta semper sit medietas Ig; ergo el est maior mg; et per consequens, si eis addantur equalia,
lateris erecti; propterea quod sequens probat ea et ag equales esse sic scilicet al et am , erit ag minor quam ae. Sed posite fuerunt equales.
ea erit medietas eg.] Ergo idem est minus et equale, quod est impossibile, et hoc sequitur si
35 gd secet figuram mukefi inter d et a. Ergo hoc est falsum.
7a. QUAMCUNQUE SECTIONEM MUKEFI LINEA RECTA CON­
Preterea si dicatur quod g d secabit sectionem inter d et basim, esto
TIGERIT QUE LINEA SIC CONTINGENS PROTRAHATUR USQUE
quod in o. Descendat ergo linea on perpendiculariter super sagittam, et
QUO EX PARTE CAPITIS CONCURRAT CUM SAGITTA ULTERIUS
ponantur ag et ae equales ut prius, et sit gp equalis en. Unde na et ap
QUANTUMLIBET PROTENSA ET A PUNCTO CONTINGENTIE
erunt equales et np dupla ad na et eg dupla ad ea. Sicut igitur liquet ex
5 DESCENDAT PERPENDICULARIS SUPER SAGITTAM, TUNC
40 quarta huius et eius probatione, proportio na ad ea est dupla ad propor­
PARS SAGITTE INTERCEPTA INTER DICTAM PERPENDICU­
tionem no ad ed, ergo et dupla ad proportionem ng ad eg. Sicut autem
LAREM ET CAPUT SECTIONIS ATQUE LINEA INTERCEPTA
na ad ea ita np ad eg per 15am 5li. Ergo proportio np ad eg est dupla ad
INTER IDEM CAPUT ET PUNCTUM CONCURSUS LINEE CON­
proportionem ng ad eg. Quare per diffinitionem 10mam 5U np ad ng sicut
TINGENTIE CUM SAGITTA SIC ULTRA PROTRACTA NECESSARIO
ng ad eg. Cum igitur np totum se habet ad ng totum sicut ng pars totius
io ERUNT EQUALES.
45 np se habet ad eg partem totius ng. Ergo per 19vam 5u pg residuum totius
Sit ut prius sectio mukefi abc' [Fig. 4.8], quam in puncto d contingat linea
np se habet ad ne residuum totius ng sicut np ad ng. Sed cum np sit
gd protensa usque quo concurrat cum sagitta f a continuata in puncto g

12 intercepte: quae intercipitur ed 13 ea . . . lineas: ag et ea ed / esse: sunt ed

15 intercepta ed 14-15 et2 . . . ag mg. R ,o m . ed
17 mukefi: parabola ed / abc': abc R 17 ante si scr. et del. R non / ante vel add. ed huius
20 de corr. R ex dhe 19 ante ducatur add. ed et et om. etiam ante kl / ante kl scr. et del. R d
27 15mam: 16. ed 25-26 duplum . . . alterius: dupla eorum ed
29-32 Ex . . . eg mg. R, om. ed 26 15mam correxi ex 15tam
Conci. 7 27-28 Sic . . . Im1 mg. R, om. ed
1 7a: PROPOSITIO VII. ed / mukefi: parabolae ed 29 totum1: tota ed / totum2: totam ed / idem: eadem ed
4 ,8 -9 contingentie: contactus ed 31 sic ed
8 cursus ed 33 posite correxi ex per te (c/. Spec, almukefi) et ed hab. partes
11 mukefi: parabolae ed / abc': abc R 35 mukefi: parabolae ed
212 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS' VERSION 213

maior quam ng, erit gp maior quam ne. Sed ex ypotesi sunt equales. erit medius lateris erecti sectionis, ut patet ex quarta descriptione huius.
Igitur erunt sibi equales et inequales, quod est impossibile. Hoc autem Sit punctus d in curvo latere sectionis, a quo exeat linea di equedistans
sequitur si linea gd secet figuram mukefi inter d et basim. Nullibi igitur 20 sagitte ak, et in eodem puncto d contingat linea gdh sectionem, protra-
50 secabit. Quare necessario continget in puncto d si ae et ag sunt equales haturque linea de. Dico quod angulus edg sit equalis angulo Idh.
[et ae et ed continent angulum rectum]. Et si sic, ergo et econtra: si gd Vel igitur angulus deg acutus est vel rectus vel obtusus. Sit primo acutus.
contingat sectionem in puncto d , tunc ae et ag erunt equales, quod erat Tunc d f perpendicularis super sagittam per 32Laml primi cadet inter a et
probandum. [Quoniam si essent inequales resecta ae aut continuata ad e. Linea vero contingens concurret cum sagitta continuata in g, eritque
equalitatem cum ag in / vel n, ductaque linea Ik vel no orthogonaliter 25 angulus f g d , per 32laml primi, acutus, et per premissam ag est equalis af.
55 super sagittam, et ducta gk vel go, erit (?) gk vel go contingens (?) sec­ Itaque linea ae divisa est in puncto / , et illi additur ag equalis alteri
tionem. Sed Ik est minor ed. Sic gk continuata secabit ed. Item gd con­ dividentium. Erit per 8vam 21 quadratum linee eg equale ei quod fit ex
tinuata secabit no. Sic non erunt contingentes (?) et fiet implicatio.] Patet ductu linee ea in lineam ag vel a f sibi similem (/ equalem) quater et
ergo conclusio Apolonii ante allegata. quadrato linee f e . Sed linea ea est quarta pars erecti lateris sectionis
30 parabole ex prima parte 4te huius. Ergo, per primam 6li et primam 2\ quod
8a. IN OMNI SECTIONE PARABOLA SIGNATO PUNCTO IN fit ex ductu a f in ae quater est equale ei quod fit ex ductu a f in latus
MEDIO SAGITTE IPSIUS, QUI QUIDEM PUNCTUS ERIT ETIAM erectum sectionis semel. Ergo quod fit ex a f in latus erectum cum quad­
IN MEDIO LATERIS ERECTI EIUS, TUNC A QUOCUNQUE rato f e est equale quadrato linee eg. Sed per secundam partem 4te huius
PUNCTO LATERIS SUI CURVI DESCENDAT RECTA LINEA VER- quod fit ex a f in latus erectum valet quadratum f d , duo vero quadrata
5 SUS BASIM EQUEDISTANTER SAGITTE, SI AB EODEM PUNCTO 35 f d et f e valent quadratum linee ed, per penultimam primi. Ergo quadratum
IN CURVO LATERE EXISTENTE EXIERINT ALIE DUE LINEE linee eg est equale quadrato linee ed. Quare linee eg et ed sunt equales.
RECTE QUARUM ALTERA EX UTRAQUE PARTE PROTENSA Ergo per 5tam primi angulus egd est equalis angulo edg. Sed per 29nam
CONTINGAT SECTIONEM ET RELIQUA PROCEDAT QUOUSQUE primi angulus Idh est equalis angulo egd. Igitur angulus edg est equalis
IN PUNCTO PRIUS SIGNATO CONCURRERIT CUM SAGITTA, angulo Idh, quod fuit ostendendum. Et consimiliter probatur de qualibet
3iv io / CAUSABUNTUR DUO ANGULI EQUALES QUORUM ALTER 40 variatione puncti d interim quod angulus e est acutus.
CAUSATUR EX LINEA CONTINGENTIE A PARTE CAPITIS SEC­ Deinde sit rectus [Fig. 4.10]. Erit itaque de perpendicularis super sagit-
TIONIS ET LINEA PROTENSA USQUE AD PUNCTUM SIGNATUM 32r tam. /Ergo per premissam ag equalis est ae. Sed ae est quarta pars lateris
ET RELIQUUS CAUSATUR EX LINEA PROTRACTA EQUEDIS­ erecti sectionis. Igitur ge est medietas lateris erecti. Sed et ed est medietas
TANTER SAGITTE ET LINEA CONTINGENTIE A PARTE ALTERA lateris erecti. Igitur de et ge sunt equales. Ergo per 5tam primi angulus
15 VERSUS BASIM. 45 edg equalis est angulo egd. Sed per 29nam primi angulus Idh equalis est
In sectione parabola bac [Fig. 4.9] sit sagitta ak protensa quousque angulo egd. Igitur angulus Idh equalis est angulo e d g , quod fuit probandum.
concurrerit cum axe piramidis in k, cuius punctus medius sit e, qui etiam47 Preterea sit angulus deg obtusus [Fig. 4.11]. Ducatur perpendicularis
df. Erit ergo punctus/ inter e et k. Quare linea a f est maior quam linea
47 ante ne scr. et del R ng
49 sequetur ed I mukefi: parabolae ed
51 et1 . . . rectum mg. R om. ed 18 medium ed
53-57 Quoniam. . . . implicatio mg. R. om. ed 19 equedistans: parallela ed
22 primum ed
Conci. 8 23 Tunc: et ed
1 8a: PROPOSITIO VIII. ed 23-25 per. . . . et om. ed et scr. At
2 quidem: idem ed 26 divisa: secta ed I et illi: illique ed / additur: adiuncta ed
6 existente om. ed 27 dividentium: segmentorum ed
8 quousque: donec ed 27-28 ex . . . quater: sub lineis ea et ag vel af ed
9 concurrerit: coeat ed 30 parabole om. ed / ex . . . 4U': per quartam ed
10 causabunter: fient ed 31 ex . . . ei: sub af et ae aequale est illi ed
10-11 quorum . . . a: unus a linea contactus ex ed 36 ed1: cd ed
12 protensa: producta ed 39 Et consimiliter: Similiter ed
13 et . . . ex: alter a ed 40 interim quod: dum ed
13-14 equedistanter: parallela ed 41-42 super sagittam; sagittae ed
14 contigentie a: contactus ex ed 43 sectionis om. ed I medietas'2: dimidium ed
16 protensa: producta ed 44 Igitur: quare ed
17 concurrerit: coeat ed / punctus medius sit: medium si ed / qui: quod ed 47 Preterea: denique ed / obtususi (!) ed
214 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ VERSION 215

ae. Sed per premissam a f et ag sunt equales. Ergo ag est maior quam portio lineari. Habeas etiam duas regulas planas quadrangulares eneas
50 ae. Fiat igitur per 3am primi am equalis ae. Ergo gm equalis erit f e , et vel ferreas quarum utraque sit tante longitudinis quantam vis esse speculi
per consequens addita utrobique linea em erit f m equalis ge. Ergo quadra­ latitudinem, et sit utriusque regule latitudo tanta ut in ea ab uno extremo
tum//?? erit equale quadrato linee ge. Cum autem linea f a sit divisa in ad aliud lineari possit portio circuli secundum quantitatem semidyametri
puncto e et sibi addita am equalis uni dividentium, per 8vam 2' Euclidis io supradicte. In altera earum lineetur portio abc [Fig. 4.12] secundum quanti­
quod fit ex ductu a f in am , vel ae, quater cum quadrato linee f e erit tatem dicte semidyametri; in alia portio d ef secundum quantitatem semi­
55 equale quadrato linee f m vel linee eg. Sed quod fit ex ductu f a in ae dyametri brevioris parum, ut verbi gratia per quantitatem pedalem. Deinde
quater, ut patet ex premissis, est equale ei quod fit ex ductu f a in latus ab utraque regula auferatur cum lima quidquid fuerit extra gibbositatem
erectum sectionis. Igitur quod fit ex ductu f a in latus erectum sectionis portionum lineatarum. Deinde iaceat super superficiem orizontis corpus
cum quadrato f e equale est quadrato f m . Per 2am autem 4te huius partem, 15 oblongum quadrangulare et basium equedistantium et equedistantium
quod fit ex f a in latus erectum sectionis valet quadratumf d . Quare quadra- laterum atque rectangulum cuius longitudo sit ut regule abc, ut gratia
60 tum f d cum quadrato f e valent quadratum f m . Sed cum per penultimam exempli 8 pedum. Protrahendo ergo super latus eius superius unam lineam
primi quadratum de valeat duo quadrata d f et f e , quadratum de valebit equedistantem linee gh et distantem ab ea per pedale, scilicet secundum
quadratum f m et per consequens quadratum eg. Ergo linee de et eg sunt quantitatem qua semidyameter circuli def est minor semidyametro circuli
equales et anguli supra basim erunt equales. Erunt igitur anguli egd et 20 abc, et sit punctus medius illius linee e. Deinde ponatur regula def super
Idh equales, quod erat ostendendum. latus superius corporis oblongi ita quod punctus medius curvature def
contingat in e lineam equedistantem linee gh, et secundum eam protrahatur
9a. IN SPECULO CONCAVATO CONCAVITATE SECTIONIS
linea curva super latus superius corporis oblongi et vocetur illa curvatura
MUKEFI OMNES RADII SOLARES SUPER DICTUM SPECULUM
def. Deinde ex transverso lateris superioris dicti corporis protraho lineam
INCIDENTES EQUEDISTANTER SAGITTE AD UNUM ET EU N­
25 rectam transeuntem per e equedistantem utrique extremorum illius lateris,
DEM PUNCTUM RELUCENT, QUI QUIDEM PUNCTUS TAM ERIT
et ab extremis illius linee medie transeant alie due linee ex transverso
5 MEDIUS SAGITTE QUAM ETIAM ERECTI LATERIS SECTIONIS. lateralium costarum equedistantes a suis extremis quousque concurrant
Hec ex premissa et communi principio perspective, scilicet / quod anguli cum basi super qua iacet illud corpus. Et inter extrema inferiora harum
incidentie et reflexionis sunt equales, ad oculum est manifesta. 33r duarum linearum ducatur linea recta ex transverso J basis; verbi gratia
10a. INSTRUMENTUM DE CALIBE VEL DE ALIA MATERIA 30 ubi dicta linea in basi tangit lineam ik signetur b et vocetur linea trans­
DURA PERFICERE QUO CONCAVITATE PARABOLA CAVARI versalis basis bm. Elevetur igitur basis et iaceat super latus ghfd oppositum
POTERIT MATERIA SOLIDA SPECULI COMBUSTIVI. basi et regula abc applicetur super latus ikac ita quod punctus medius
Si velis ut comburat ad distantiam 20 pedum, oportet ut habeas perti- curvature abc contingat lineam ik in puncto b et secundum eius curva­
5 cam vel virgam 40 pedum qua tamquam cum semidyametro possit circuli turam protrahatur linea curva, que in latere ikac vocetur abc. Sunt ergo

51 per consequens: ideo ed 6 liniari ed / eneas: aereas ed

52 divisa: secta ed 8 ante ea del. R ab
53 et sibi: eique ed / dividentium: segmentorum ed 9 deliniari ed
54 ductu om. ed I quater om. ed 12 parum: paulo ed
55 ductu om. ed 13 cum om. ed
56 quater om. ed 17 protrahenda ed
57 sectionis2 om. ed 18 per pedale: pedali intercapedine ed
62 per consequens: proinde ed / Ergo linee: Lineae igitur ed 20 regula om. ed
63 erunt1 om. ed 21 quod: ut ed
64 erat: fuit ed 23 curvatura: arcuatio ed
24 dicti: oblongi ed / protrahe ed
Conci. 9 28 super . . . corpus: cui corpus illud innititur ed
1 Ψ\ PROPOSITIO IX. ed / concavato: excavato ed 30 ik: ak ed / signetur: fiat ed
2 mukefi: parabolae ed 3 0 - 31 transversalis: transqersa (/) ed
3 1 - 32 Elevetur . . . basi: Vertatur igitur corpus ut incumbat basi oppositae ed
Conci. 10 32 super latus: lateri ed / quod: ut ed
1 103: PROPOSITIO X. ed 33 ik: ak ed
3 comburentis ed 33-34 curvaturam protrahatur linea: ductum ducatur ed
5 cum om. ed 34 in . . . ikac om. ed
216 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ VERSION 217

35 due curvature in lateribus oppositis, que sunt abc et def portiones duorum / Ergo superficies curva abcdef est portio superficiei curve piramidis
circulorum, quorum centrum superioris directo supponitur centro in­ rectangule, quod fuit probandum.
ferioris ita quod linea protracta inter centra est perpendicularis super 60 His sic habitis super lineam bo protractam ulterius signetur punctus
utrumque circulum. Per instrumentum ergo cuius acuties sit linea recta n tantum distans ab o quantum o distat ab ipso b ita quod bo sit equalis
abradatur ex costa laterali que est ghki quidquic (/) fuerit extra curva- om. Protrahatur igitur super basim ikac una linea transiens per n eque-
40 turas abc, def. Relinquetur ergo in corpore isto oblongo una curva super­ distans ibk, per cuius longitudinem et punctum e transeat una super­
ficies que est def abc, que est portio curve superficiei piramidis rectangule ficies plana abscindens quidquid est extra illam superficiem ex parte b.
inter duos circulos equedistantes. 65 Differentia ergo communis illi superficiei plane et superficiei curve est
Quod sit portio curve superficiei piramidis satis patet quia centra duorum sectio mukefi. Nam si ymaginemur superficiem planam triangulam tran­
circulorum quorum portiones constitunt latera curva superficiei proposite seuntem per conum et per centrum basis piramidis atque per lineam bom ,
45 sunt in linea perpendiculari super utrumque circulum. Et quod talis piramis cuius latera sunt in superficie curve piramidis, dividentem piramidem
sit rectangula probatur. Iaceat enim corpus super basim eius priorem ita in duo equalia, tunc linea en erit equedistans alteri lateri dicti trianguli,
quod curvatura abc sit in latere inferiori. Tunc a puncto e ymaginemur 70 linea enim bn protracta usque ad latus trianguli ex parte n causabit cum
lineam perpendicularem super basim que necessario cadet in linea bm dicto latere ex parte coni angulum medietatem recti eo quod ipsa est
transversali. Cadat igitur exempli gratia in puncto o. Cum igitur eo precise equedistans dyametro basis piramidis, angulus autem eno est medietas
50 sit pedale eo quod altitudo corporis posita fuit pedalis et cum etiam bo recti. Ergo per 28[am] primi en et dictum latus equedistant. Sic igitur habetur
sit pedalis eo quod e distabat a linea gh per pedale ex ypotesi, igitur superficies plana disposita in lateribus secundum sectionem mukefi. Illi
in triangulo eob duo latera eo et ob sunt equalia et angulus o est rectus. 75 igitur applicetur lamina de calibe et formetur secundum formam dicte
Ergo quilibet aliorum duorum angulorum est medietas recti. Cum autem sectionis et postquam secundum eandem formam acuatur ita quod acuties
linea eo equedistet axi piramidis, si transiret bc usquequo concurreret sit linea curva curvitate mukefi. Deinde cum illo instrumento cavetur
55 cum axe esset angulus beo equalis angulo medietatis coni, scilicet medietati alia lamina calibea vel ferrea et postea poliatur et habebis speculum com­
recti. Et per consequens totus angulus coni piramidalis esset angulus burens ad minus ad tantam distantiam quanta est medietas semidyametri
rectus. Quare piramis est rectangula per descriptionem primam huius.35*78942601
58 Ergo: Quare ed i superficies curva: arcuatio ed / superficiei curve om. ed
59 quod . . . probandum om. ed
35 curvature: curvae ed / portiones: partes ed 60 sic habitis: ita constructis ed I protractam: productam ed / ulterius signetur:
36- 37 centrum . . . inferioris: centra directe opponuntur ed cadat ed
37 quod: ut ed / protracta . . . est: coniungens illa sit ed 61 distat ab ipso: a ed I quod: ut ed
37- 38 super utrumque circulum: utrique circulo ed 62 om: on ed / Protrahatur: Ducatur ed
38 acuties sit: acies sii ed 6 2 - 63 equedistans: parallela ipsi ed
39 ex: a ed I que . . . ghki om. ed I quidquic(7) fuerit: quidquid est ed 63 per . . . longitudinem om. ed / punctum: per punctum ed
39- 40 curvaturas abc, def: curvas abcdf ed 6 3 - 64 una . . . plana: planum ed
40- 41 in . . . est1: fornix ed 66 mukefi: parabola ed
41,43 portio: pars ed 66-67 Nam . . . transeuntem: Transeat enim planum triangulum ed
42 equedistantes: parallelos ed 67 piramidis om. ed / atque per: perque ed
43 Quod: Nam quod ed / curve superficiei om. ed / satis om. ed i quia: cum ed 68-69 sunt . . . equalia: secent pyramidem bifariam ed
44-45 quorum . . . linea: sint in ed 69 tunc . . . equedistans: erit linea en parallela ed
45 super: ad ed / talis om. ed. 7 0 - 71 usque . . . medietatem: constituit in dicto latere versus conum angulum
46 Iaceat: Incumbat ed / super . . . priorem: basi priori ed dimidium ed
47 quod: ut ed / curvatura: curva ed / Tunc a: et ex ed / ymaginemur: cogitemus ed 7 1 - 72 est equedistans: parallela ed
48 super basim: basi ed 7 2 - 73 medietas recti: recti dimidius ed
49 transversa ed I igitur exempli: ergo verbi ed / punctum ed 73 28[aml: 29 ed I et . . . equedistant: lateri dicto aeque distat ed
49-50 precise . . . pedalis: sit pedalis ob positam corporis altitudinem ed 74 superficies . . . secundum: planum larerum (/) sitorum iuxta ed / mukefi:
50 et . . . etiam: cumque ed parabolam ed
51 per pedale: pedali intervallo ed 74-75 Illi igitur: Cui ed
52 est om. ed 75 de . . . formam: chalybea eiusque ductu ed
53 quilibet . . . medietas: uterque reliquorum dimidius ed 76 quod acuties: ut acies ed
54 equedistet: invicem sit ed / usquequo: donec ed 11 curva . . . mukefi: incurva curvitate parabolae ed I Deinde . . . illo: Quo ed /
55 medietatis: dimidio ed / medietati: dimidius ed excavetur ed
56 per consequens totus: proinde ed / piramidalis om. ed / angulus2 om. ed 78 et . . . poliatur: poliaturque ed
57 Quare: et ed / est: ipsa ed 19 ad minus: ut minimum ed / medietas: semis ed
218 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ VERSION 219

80 circuli cuius portio lineatur in prima regula superius designata abc. unam aliam superficiem planam secantem piramidem predictam eque-
Quod enim ad minus ad tantam comburat patet ex nona conclusione distanter triangulo predicto ab c . Et sit communis sectio curve superficiei
huius. piramidis et illi superficiei plane sic equedistanti triangulo predicto linea
[Verum si velis precisius cognoscere distantiam ad quam comburat, 15 eghf ita quod e sit punctus eius propinquior capiti piramidis. Erit igitur
sic invenis (Fig. 5.10). Si ponis bx 40 et eo unum, similiter erit ob unum linea eghf necessario curva per secundam huius cum non transeat per
85 et nx 38, cui est equalis xq, quare qn erit radix quadrata de 2888, que caput piramidis. Que si in infinitum protrahatur ex parte/super piramidem
34r est modico / plus quam 53, et sic tota eq erit aliquo plus quam 54, cuius in infinitum elongatam ex parte basis secundum superficiem secantem
medietas est modico plus quam 27. Quare si semicirculus factus est cum in infinitum ymaginatam, et ac latus trianguli cum elongatione piramidis
40, speculum comburet ad distantiam 27 pedum et modico plus. Item si 34v 20 in infinitum elongetur, tunc dico quod linea curva / ef et linea recta ac in
bx esset 5, nx erit 3, eo 1, similiter ob 1, nq erit radix quadrata de 18, infinitum protracte continue plus et plus appropinquant et tamen nun­
90 ne radix quadrata de 2. Iungendo autem eos proveniet eq radix quadrata quam concurrent. Quod nunquam concurrant patet quia sunt in super-
de 32, cuius medietas est radix de 8; ad tantam fieret tunc combustio.] ficiebus equedistantibus planis ex ypotesi. Ergo semper distabunt ad
Sequitur conclusio mirabilis prima facie, que licet proposito non per­ minus per tantum quanta est perpendicularis protracta ab una superficie
tineat, ipsa tamen ex sectione addita sequi videtur. 25 ad aliam; et quod continue plus appropinquabunt probatur. In linea eghf
curva signa duo puncta g et h, per que transeant duo circuli secantes
l l a. DUAS LINEAS QUARUM ALTERA RECTA ET ALTERA
piramidem equedistanter basi. Et ubi dicti circuli secant lineam ac sint
CURVA VEL QUARUM UTRAQUE CURVA CONSIMILI CURVI­ puncta k et /. Dico lineam rectam inter h et l breviorem esse linea recta
TATE POSSE PROTRAHI QUE QUANTO PLUS PROTRAHUNTUR inter g et k.
AMPLIUS APPROPINQUANT ET TAMEN NUNQUAM CONCUR- 30 Corda enim dupli arcus portionis circularis intercepte inter h et l est
5 RENT ETIAM SI IN INFINITUM FUERINT PROTRACTE.
equalis corde dupli arcus portionis circularis intercepte inter g et k eo
Capiatur piramis rotunda orthogonalis, quam piramidem secet in duo quod utraque corda sit dupla ad perpendicularem interceptam inter utram­
equalia triangulus abc [Fig. 4.13] transiens per caput piramidis a et centrum que superficiem. Sed illi circuli sunt inequales quia transiens per hl est
quod sit d. Iaceat autem dicta piramis super orisonte sic quod caput eius maior. Unde minor curvior est. Ideo sinus versus portionis gk est maior
sit versus boream, basis versus austrum, ita trianguli abc secantis ipsam 35 sinu verso portionis hl. Et quia sinus recti eorum (!) sunt equales, ergo
io in duo media latus ab iaceat supra orizontem. Unde dictus triangulus per penultimam primi linea recta inter g et k est longior quam linea recta
secundum se totum erit in superficie circuli meridiani. Ymaginor ergo inter h et /, quod fuit probandum.
80 delineatur e d I superius designata om. ed
Sed diceres: non est hoc mirabile quia linee ille non sunt in eadem
81 Quod . . . conclusione: quod patet ex nona ed plana superficie sed curva. Ideo probabo etiam conclusionem de duabus
83-91 Verum. . . . combustio add. R., non in Spec, almukefi comp.
83-84 Verum . . . invenis: Exactius vero intervallum sic inveneris ed
86 modico: semissae unitatis ed I et sic: atque ita ed / aliquo: aliquantulo ed
12 unam . . . secantem: aliud planum secans ed / predictam om. ed
87 medietas: S. ed I factus est cum: est ed 14 superficiei plane: plano ed
88 speculum . . . distantiam: erit intervallum ed / modico: paulo ed 15 ita quod: ut ed
88-89 si bx esset: sit bx ed 18 elongatam: productam ed
89 nx erit tr. ed / erit2 om. ed 19 ymaginatam: conceptam ed / elongatione piramidis: pyramide ed
90 de om. ed I Iungendo . . . eos: Quibus abditis ed 20 elongetur ed elongentur R / tunc om. ed
91 medietas: S. ed / fieret tunc: videlicet intercapedinem istis positis fiet ed 21 plus et plus: magis ac magis ed
92-93 pertineat: inserviat ed 24 per tantum: tanto ed / post protracta del. R ambo . . . (?)
93 ex: e ed / addita: hyperbole ed 25 et quod: quodque ed
25-26 In . . . curva: In curva eghf ed
Conci. II 26 signa: sint ed
1 1la: PROPOSITIO XI. ed 27 piramidem equedistanter: conum invicem ed / dicti circuli om. ed
5 fuerint protracte: protrahantur ed 28 lineam om. ed / linea om. ed
6 ante Capiatur add. ed Prima pars propositionis / orthogona ed / piramidem 30 Corda: Subtensa ed / portionis . . . intercepte: intercepti ed
om. ed 31 corde: subten. ed / portionis . . . intercepte: intercepti ed
7 triangula ed 32 corda: subtensa ed
8 dicta om. ed / ante orisonte del. R centrum? / orisonte . . . eius: centrum 34,35 portionis: arcus ed
horizontis ut conus ed 35 ergo: erit ed
9 ipsum ed 36 linea1,2 om. ed / est om. ed
10 in duo media om. ed / iaceat supra orizontem: incumbat horizonti ed 38 dices ed / quia: cum ed / linee ille tr. ed
10-11 dictus . . . erit: triangulus totus est ed 39 Ideo . . . conclusionem: Idem ergo demonstretur ed
 REGIOMONTANUS’ VERSION 221
220 ARCHIMEDES IN THE MIDDLE AGES

40 lineis in eadem plana superficie existentibus. Capio enim eandem super­ sinus versus portionis pr minor sinu verso portionis oq. Sed duo puncta
ficiem quam prius, scilicet euntem per e , g, h, f equedistanter triangulo o , p distant a linea mn solummodo quantitate sinuum versorum predic-
abc, et extendatur ista superficies supra piramidem quousque orthogonali- tarum portionum. Ergo punctus p minus distat a linea mn quam punctus
ter secet unam aliam superficiem planam iacentem superae orthogonaliter 65 o, quod fuit probandum. Patet ergo de duabus lineis quarum alia est
ad superficiem trianguli abc, et sit communis sectio istarum duarum super- recta, alia curva.
45 ficierum linea mn recta, que erit equedistans linee ac et in eadem super­ Quodsi ymaginemur aliam piramidem similem et equalem priori iacere
ficie plana cum curva linea eghf. Dico eas lineas mn et e g h f continue plus super piramidem priorem sic quod linea ac sit communis utrique et ymag­
35r et plus versus ba/sim sibi appropinquare et tamen nunquam concurrere. inemur superficiem transeuntem per eghf equedistanter triangulo abc
Quod nunquam concurrent patet quia superficiem planam superpositam 70 extendi ulterius quousque secuerit superiorem piramidem eodem modo
ac lateri nullus punctus piramidis nisi qui fuerit in linea ac tanget. Sed quo secat inferiorem, tunc sectio communis dicte superficiei et curve
50 in ea est linea mn cuius nullus punctus tanget lineam ac. Igitur nullus superficiei superioris piramidis erit linea curva consimilis curvature cum
punctus linee mn tanget aliquem punctum piramidis, nec econtra. Linea linea curva eghf, que due linee curve continue plus et plus appropin­
autem e g h f curva est in superficie piramidis curva. Ergo linea e g h f nun­ quabunt eo quod utraque continue plus et plus appropinquat ad lineam
quam concurret cum linea m n, quod fuit ostendendum. 75 mn et tamen nunquam concurrent cum mn linea media. Sic ergo patet
veritas conclusionis ad utramque partem. Hec breviter sufficiunt.
Quod autem continue plus et plus appropinquabunt ex parte basis patet.
55 Nam quibuscunque duobus punctis in linea e g h f signatis, ut o, p , ita quod
o sit propinquior cono piramidis, si per ista duo puncta transeant duo
circuli dividentes piramidem equedistanter basi, qui quidem circuli tangant
lineam ac in punctis [q et r], sinus rectus portionis oq erit equalis sinui
recto portionis rp eo quod o et p sunt in superficie equedistante triangulo
60 abc in cuius latere ac sunt duo puncta q et r. Cum igitur pr sit portio
maioris circuli quam portio oq et per consequens minoris curvitatis, erit40123*67895

40 eadem . . . existentibus: eodem plano ed I enim om. ed

41 transeuntem ed / equedistanter: parallelam ed
42 et . . . supra: quae extendatur super ed
42-43 quousque orthogonaliter: donec ad rectos angulos ed
43 unam om. ed / iacentem super: innitentem ed / orthogonali ed
45 linea om. ed / equedistans: invicem ed
45- 46 eadem . . . plana: eodem plano ed
46 linea om. ed I eas om. ed et correxi ex ea in R
46- 47 continue . . . plus: magis magisque ed
47 et . . . nunquam: nusquam tamen ed
48 superficiem . . . superpositam: planum subpositum ed
49 fuerit: fit ed
50,51 tangit ed
51 aliquem: aliquod ed I contra ed 62 portionis1,2: arcus ed
51-52 Linea . . . curva1: curva autem eghf ed 63 solummodo: non nisi ed
53 fuit ostendendum: erat propositum ed 65 probandum: demonstrandum ed / alia: una ed
54 continue . . . basis: versus basin nunquam desinent propiores fieri ed 67 ymaginemur: cogitemus ed
67-68 similem . . . quod: priori aequalem illi incumbere ita ut ed
55 quibuscunque: quibuslibet ed
55-56 ita . . . sit: sitque o ed 70 ulterius om. ed / secuerit: secet ed
56 ista duo: illa ed 72 similis incurvationis ed
57 dividentes: secantes ed I equedistanter: invicem ed / quidem circuli om. ed 73 linee om. ed
58 [q et r] addidi 13-1A continue . . . appropinquabunt: non desinent continue propiores fieri ed
58,59 portionis: arcus ed 74 continue . . . plus2: magis magisque ed I ad lineam: lineae ed
59 rp corr. R ex rqp / equedistante: parallela ed 75 et . . . nunquam: nunquam tamen ed
60,61 portio: arcus ed 75-76 ergo. . . . sufficiunt: igitur utraque pars propositionis demonstrata esto.
61 oq corr. R ex opq / per consequens: proinde ed FINIS.e d
 REGIOMONTANUS’ ADDITIONS 223
Text D
[Tabula, 36r] Continued
Additions by Regiomontanus to the Speculi [Consult Fig. 5.11 and the Text in Passage 1]

almukefi compositio quad­ quad­ quad­ distantia

ratum ratum ratum cumbustionis
el Im Im Ih Ih hm hm
[Tabula, 36r]
[Consult Fig. 5.11 and the Text in Passage 1] partes partes partes minuta partes partes partes partes minuta
quad­
quad- quad­ distantia 39 3900 14 196 4096 64 0
ratum
ratum ratum combustionis 40 4000 15 225 4225 65 0
el Im Im ih lh hm hm 41 4100 16 256 4356 66 0
42 4200 17 289 4489 67 0
partes partes partes minuta partes partes partes
partes minuta 43 4300 18 324 4624 68 0
1 100 44 4400 19 361 4761 69 0
10 0 24 576 676 26 0 45 4500 20 400 4900 70 0
2 200 14 9 23 529 729 27
3 300 0 46 4600 21 441 5041 71 0
17 19 22 484 784
4 28 0 47 4700 22 484 5184 72 0
400 20 0 21 441 841 29 0 48 4800 23 529 5329 73 0
5 500 22 22 20 400 900 30 0 49 4900 70 0 24 576 5476 74 0
6 600 19 361 961 31 0
7 700 0 50 5000 25 625 5625 75
18 324 1024 32 0
8 800 17 289 1089 33 0
9 900 30 0 16 256 1156 34 0
10 1000 [Regiomontani additiones]
15 225 1225 35 0
11 1100 14 196 1296 36
12 1200 0
13 169 1369 37
13 1300 0 Π]
12 144 1444 38 0
14 1400 11 121 1521 39 0 35v / Si ponitur totum latus erectum in sectione mukefi 100, erit quarta pars
15 1500 10 100 1600 40 0 25; ad tantum combureret speculum ex capite sectionis factum. Sed si fiet
16 1600 40 0 9 81 1681 41 0
17 1700 8
ex medio sectionis scilicet partis eius quam attingit latus erectum com­
64 1764 42 0
18 1800 buret ad 50. Si fiet ex eo quod sequitur caput secundum sagittam quam
7 49 1849 43 0
19 1900 5 separet linea ordinis sit 5 fiet quadratum linee ordinis 500. Unde combustio
6 36 1936 44 0
20 2000 5 25 2025 45 ad 30. Et sic de aliis. Unde invenias singulas lineas ordinis et videbis
21 2100 0
4 16 2116 46 ad quantum fiat combustio. Sed quantitates linearum ordinis reperies ex
22 2200 0
3 9 2209 47 0
23 2300 2
5[a] (/ 4a) propositione, scilicet multiplicando sagittam in latus erectum
4 2304 48 0
24 2400 et producti radix est linea ordinis, cuius quadratum iunge quadrato residue
1 1 2401 49 0
25 2500 50 0 io partis sagitte inter lineam ordinis et punctum sectionis sagitte et lateris
0 0 2500 50 0
26 2600 1 1 2601 51 0 erecti, productique radix erit distantia combustionis. Ecce tabula.
27 2700 2 4 2704 52
28 2800 0
3 9 2809 53
29 2900 0 [Notae pro tabula]
4 16 2916 54
30 3000 0
5 25 3025 55
31 3100 6
0 36r / Lineabis itaque primo secundum mensuram libitam lineas ordinis super
36 3136 56 0
32 3200 7 49
sagittam protrahendo quantumlibet, et ex hac tabula poteris eas abscindere
3249 57 0
33 3300 8 is secundum quantitates Im [Fig. 5.11] que debentur quantitatibus el statutis,
64 3364 58 0
34 3400 9 81 3481 59 0 vel sic accipiendo quantitates hm et ponendo pedem unum circini im­
35 3500 10 100 3600 60 0 mobilem semper in h punctum combustionis et mobili reseca lineam
36 3600 60 ) 11 121 3721 61 0 ordinis Im, vel si adhuc aliter libuit ex propositione quinta iungendo medie­
37 3700 12 144 3844 62 0
38 3800 13 169
tatem el cum medietate lateris erecti et secundum quantitatem provenien-
3969 63 0
20 tem extende circinum et posito suo pede immobili in linea ex et mobili
222 in e et resecabis mobili lineam ordinis Im. Sic tribus diversis viis idem
224 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ ADDITIONS 225

reperies. Secunda tamen magis amplectenda est quod in ea nulla fractura ta ad ae, sed ta ad ae ut suorum duplorum tg ad er, ergo quadrati tn
videatur (?). ad quadratum ed est ut tg ad er. Sed cum tx linea sit maior tn, erit tx ad
Hic vides quod semper agregatum ex Ih et hm est equale medietati 40 ed proportio maior quam tn ad ed. Sed tx ad ed est ut tg ad ge. Quare
25 lateris erecti, hoc quidem dum linea ordinis Im est inter caput sectionis tg ad ge proportio est maior quam tn ad ed. Sed proportio tn ad ed est
et punctum combustionis; sed post hoc agregatum ex hg et hl equatur medietas proportionis quadrati tn ad quadratum ed seu medietas propor­
hm. Huius causam (?) post scriptis invenies. tionis tg ad er. Quare proportio tg ad ge est maior medietate propor­
tionis tg ad er. Posita itaque ge media inter tg et er, oportet ut proportio
[2] 45 tg ad ge sit maior proportione ge ad er. Ergo residui te ad residuum eg
proportio est maior quam proportio totius tg ad totum ge. Sed cum tg
36v / Septima propositio que videlicet est una ex conclusionibus Appollonii (/) sit maius quam ge erit a fortiori et maius quam eg. Inde et arg minor
sic potest ostendi. Sit sectio mukefi akdo [Fig. 5.12a] quam in puncto quam aet. Sed ipse iam erant equales. Non potest igitur esse ut ea sit
d contingat linea gd protensa usquequo concurrat cum sagitta/α continuata minor ag. Ex his aufertur necessitas propositi: ea et ag equales esse.
in puncto g et perpendicularis exiens a d super f a sit de. Dico ea et ag 50 Hanc propositionem pones per se. Et huius conversam pones quoque
5 lineas esse equales. de per se. Et potest conversa facile etiam sic ostendi. Sint ae et ag equales,
Si enim non sunt equales, erit altera maior. Sit primo ea maior. Resece- et gd sit ducta ad extremitatem linee ordinis ed. Dico gd contingere sec­
tur ergo in / ut sit al equalis ag et fiat agq equalis ale [Fig. 5.12a]. Unde tionem in puncto d. Quoniam si non contingit in eo, possibile est educere
gq erit equalis le. Et ducatur Ik linea ordinis que eque distabit linee ed a g puncto contingentem aliam que aut continget in puncto aliquo inter
et sit k in sectione ipsa mukefi ado. Et erit necesse ut Ik non attingat 55 a et d aut inter d et o; quocumque posito sequitur ex 7a partem esse
io lineam gd quoniam si Ik attingeret eam tunc essent duo puncta in linea equalem toti. Nam pars sagitte inter lineam ordinis a puncto contingende
g d que essent communia superficiei piramidis rotunde et non in linea eunte ductam et caput sectionis erit equalis ag. Quare erit et equalis ae pars
per conum piramidis. Ergo per secundam propositionem linea gd secaret toti aut totum parti.
piramidem; igitur non esset contingens, quod est contra positum. Ergo
Ik continuetur usque ad concursum eius cum gd in m. Quia (?) itaque ex [3]
15 4a propositione proportio quadrati ed ad quadratum Ik est sicut ea ad al,
sed ea ad al est ut suorum duplorum, scilicet qe ad gl, ergo proportio 37v / Nota circa quartam propositionem que dicit de sectione parabola que
quadrati ed ad quadratum Ik est sicut qe ad gl. Sed cum Im sit maior fit ex piramide orthogonia rectangula, si fiet sectio mukefi in piramide
Ik, erit, per I9[aml 5li, ed ad Im minor proportio quam ed ad Ik. Sed ed ad acutiangula aut obtusiangula inveniresque latus erectum in ea, tale quidem
Im est sicut eg ad gl. Quare eg ad gl proportio est minor quam ed ad Ik. inter lineas ordinis quod totum latus erectum quadruplum esset parti sagitte
20 Sed proportio ed ad Ik est medietas proportionis quadrati ed ad quadratum 5 inter caput linee muniani et ipsum latus erectum, tunc procederet demon­
Ik. Quare proportio eg ad gl est minor medietate proportionis quadrati stratio omnis quemadmodum fit in piramide rectangula. Quod vero possi­
ed ad quadratum Ik. Sed qe ad gl proportio est sicut tota proportio quad­ bile sit id fieri in acutiangula sic accipe.
rati ed ad quadratum Ik. Quare eg ad gl proportio est minor medietate Sit sectio triangularis per conum piramidis procedens abc [Fig. 5.13a];
proportionis qe ad gl. Posita itaque eg inter qe et gl sequitur cum qe axis ad; sectio mukefi, id est equedistans, orthogonalis super triangulum,
25 ad gl sit agregata ex duobus, scilicet qe ad eg et eg ad gl, et eg ad gl sit 10 pfgq, in qua sit sagitta f d equedistans lateri trigoni ac. Volo autem ut
minor medietate proportionis qe ad gl, erit necessario proportio qe ad ac sit equalis bc ita ut trigonus abc sit isocheles. Tunc ac fiet duplum dc;
eg maior medietate proportionis qe ad gl. Ergo et maior proportione itaque dc et d f equales. Denique f e sit quarta pars sagitte f d . Facta autem
eg ad gl. Ex hoc et penultima additarum Campani in 5t0 Euclidis erit linea ordinis peg, dico pg quadruplum esse f e . Nam quadratum dq seu
proportio qg residui ad el residuum maior quam proportio qe totius ad dc quadruplum fiet quadrato eg. Ergo dc dupla est linea eg. Quare et
30 eg totum. Sed qe est maior eg. Igitur a fortiori qg est maior quam el 15 d f dwpXa. est e g . Ideo d f et p g equales. Quare etiampg quadrupla est ad/<?.
et aq maior quam ae; sed ipse iam erant equales. Non potest igitur esse Sic vero invenies in omni piramide acutiangula latus erectum cadere
37r ut ea / sit maior ag. (? dei. ?) quod totum quadruplum est sagitte quam ipsum secat. Cadet (?)
Sit deinceps ea minor ag [Fig. 5 .12b]. Resecatur ag in r ut ar sit equalis inter punctum quod caput sectionis dicitur et punctum medium inter caput
ae, et fiat aet equalis arg. Unde et et rg equales erunt. Et ducatur tn et concursum sagitte cum axe, puta inter / e t h. Cognitis lateribus trigoni
35 linea ordinis. Et erit simili ratione ut antea necesse ut non attingat ad 20 abc note erunt dq et d f in quocunque trigono sive acuto sive amplo quod
lineam gd continuatam. Ergo continuetur ad eius concursum in x. Quia dq sit equalis dc et ac sit dupla ad df. Penes f e unam rem, d f igitur ad
itaque ex 4a propositione proportio quadrati tn ad quadratum ed est ut unam rem sicut quadratum dq ad quadratum eg. Ergo quadratum eg notum
226 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ ADDITIONS 227

erit in rebus, cuius quarta pars equale (/ equalis) erit uni censui. Sic census [5]
equatur rem. Notum igitur fietfe, quod querebatur. (Facilius autem sic:
25 quia dq est medio loco proportionalis inter d f et latus erectum, unde latus In speculo mukefi annulari datis semidiametris circulorum qui bases
erectum notum fiet.) sunt speculi cum profunditate speculi, locum concursus radiorum re­
(Idem facies in omni pyramide cuius basis est rotunda etiam si declivis flexorum, sagittam quoque cum latere erecto comperire.
aut obliqua fuerit ipsa piramis dum saltem latera trigoni data sunt.) Sit sagitta talis sectionis ag [Fig. 5.16]; linee ordinis bd et ce, cum
Ut in sectione ampligonii piramidis sint in trigono abc quodlibet laterum 5 parte sagitte quam intercipiunt scilicet bc, sint date. Propositum est in­
30 ab, ac 72, medietas vero lateris bc sit 60; erunt itaque dq 60 et d f 36. venire sagittam ab et inde latus erectum. Quia ex ductu lateris erecti
Sitfe una res; 36 ad rem sicut 3600, scilicet quadratum dq, ad quadratum in ab consurgit quadratum bd, item ex ductu eiusdem lateris erecti in ac
eg. Quare quadratum eg fiet 100 res. Huius quarta pars, scilicet 25 res, consurgit quadratum ce, ideo quadratum ce ad quadratum bd sunt ca ad
sunt quadratum unius rei, scilicet fe. Ergo fe est 25, quod querebatur. ab. Ideo disiunctim differentia quadratorum ce et bd ad quadratum bd
Sit in piramide ampligonia, invenies quod latum erectum, quod quadrup- io sicut cb ad b a . Sed tria prima sunt nota, igitur ab nota. Hinc latus erectum
35 lum est ad sagittam quam abscindit, cadet inter concursum axis cum sagitta et eius quarta, que querebantur.
et punctum medium inter caput sectionis et concursum sagitte cum axi;
tantoque magis accedit ad d quanto amplior sit angulus a . In acuto vero
angulo a, tanto magis accedit ad/ quanto acutior est angulus a. [6]

, / Ex quacunque pyramide rotunda (?) cuius basis est circulus sectionem

[4] mukefi elicere dum latera trianguli ipsam per equa secantis nota sunt.
Intellige pyramidem huiusmodi sectam super axe superficie plana ortho-
38r / Sit eh quarta lateris erecti [Fig. 5.14], hg orthogonaliter ei insistens gonali super basim. Sectio communis sit triangulus abc [Fig. 5.17], cuius
sit medietas lateris erecti, l sit punctus in sagitta super quo stet linea 5 basis dyameter erit bc, in qua centrum basis sit d. Alia superficies plana
ordinis que sit Im. Ducta linea hm dico, si / est inter e et h, quod Ih et orthogonalis super trigonum abc secet pyramidem ita quod differentia
hm simul iuncte equentur/zg. Producam he inp ut ep sit equalis le. Iam (?), communis huius et trigoni sit linea d f equedistans lateri ac. Huius secunde
5 per Svam2l Euclidis, quod fit ex pe in eh quater cum quadrato Ih equum superficiei plane differentia communis cum superficie curva pyramidis
est quadratop h . Sed ex/?e, seu el, in eh quater fit quadratum Im. Sic duo est linea muniani continens sectionem mukefi. Signate sint enim (?) in ea
quadrata Im et Ih equalia sunt quadrato p h . Sed et ea equalia sunt quadrato io linee ordinis mn, hg, dq orthogonales super sagittam fd. Contingit aliquam
hm. Ideo hm equalis est hp. Duplum autem eh equale est hg, etiam ph lineam ordinis reperire in sagitta que ad sagittam quam ipsa secat dupla
et hl simul equales sunt duplo eh. Ideo ph et hl iuncte simul equales sunt sit. Sit ea hg ita ut tota pg latus erectum sit et quadruplum ad sagit­
io hg. Quare et mh adiucta (/) hl fiet equalis hg, quod est intentum. tam f h . Dico etiam hic mn medio loco proportionalem esse inter pg et
Si vero linea ordinis Im sequatur h , scilicet punctum combustionis ut fm , quod eadem via probabis que superius in 4a propositione ostensa
in secunda figura hic posita [Fig. 5.15], dico quod gh et hl coniuncte sint 15 est. Est enim quadratum pg ad quadratum hg sicut pg linea ad fh lineam
equales hm. Continuatur enim le in 5 ut es sit equalis eh. Iam iterum quia utrobique quadrupla proportio. Item quadratum hg ad quadratum
quadratum Is, totius composite scilicet, equale erit ei quod fit ex le in es mn sicut hf linea ad fm lineam. Ideo quadratum pg ad quadratum mn
15 aut eh quater cum quadrato hl. Sed quod fit ex le in eh quater equale sicut pg linea ad fm lineam. Sed linee pg ad lineam mn proportio est
est quadrato Im, ex quarta huius. Ideo ut antea Is equalis erit hm. Sed medietas proportionis quadratic# ad quadratum mn, ideoque etiam medie-
hs est equalis hg quia utraque sit medietas lateris recti. Ideo gh et hl 20 tas proportionis linee pg ad lineam fm. Igitur proportio mn ad fm erit
simul equales sunt sl. Quare etiam equales hm, quod est propositum. altera medietas proportionis pg linee ad fm lineam. Sic mn est medium
Ex his concludes mirabilem uniformitatem habitudinis linearum ordinis proportionale inter fm et latus erectum sectionis. Inde trahes omnia (?)
20 et suarum sagittarum quas secant, in omni sectione mukefi, sive ipsa fit deinceps fieri que in sectione mukefi facta in pyramide orthogonia. In­
(forte sit) ex piramide rectangula sive ampligonia sive oxigonia, quemad­ venies autem quantitatem fh ex notis lateribus ac et bd (/ bc). Nam sic
modum in omni circulo uniformitas est habitudinis sinuum rectorum et 25 note fient dq et df, nam dq est equalis bd seu dc et d f est medietas a c .
versorum arcuum similium. Unde quod mirum est curvitas huius sectionis Sed quadratum dq ad quadratum hg sicut d f ad fh. Si pones itaque hg
eiusdem habitudinis in omni sectione sive parva sit sive magna neces- rem unam invenies propositum. Aut facilius sic: quia dq linea ordinis in
25 sario manebit. medio loco proportionalis inter df sagittam quam secat et latus erectum,
228 ARCHIMEDES IN THE MIDDLE AGES
REGIOMONTANUS’ ADDITIONS 229
ergo latus erectum consequenter (?) fiet dividendo quadratum dq per sag-
30 ittam dfi, hinc eius quarta fh. different ways. But the second matter to be better grasped is that no inter­
ruption may be seen in it (the curve?).
Here you see that always Ih + hm = Vi latus rectum, so long as the line
[Additions of Regiomontanus] of order Im is between the vertex of the section and the point of com­
bustion; but if [it is] beyond this [point of combustion], then hg + hi = hm.
[i] You will find the proof of this in the things written later [in Passage 4].
If the whole latus rectum of a parabolic section is posited as 100, [its]
quarter part will be 25. The mirror would then burn at such [a distance] [ 2]
taken [on the axis and starting] from the vertex of the section. But if it The seventh proposition [of the Speculi almukefi compositio], which
(the distance of combustion) is taken [as starting] in the midst of the sec­ is evidently one of the conclusions of Apollonius [in fact, Prop. 1.35],
tion [i.e. on the curve], evidently from the point at which the latus rectum can be demonstrated as follows: Let there be a parabolic section akdo
touches the section, it would burn at [a distance of] 50. If the distance [see Fig. 5.12a], which line gd (extended until it meets the axis fa , itself
which a line of order cuts off on the axis when starting from the vertex continued to point g) touches in point d. And let de proceeding from d
is 5, then the square of the line of order will become 500. Whence the be a perpendicular upon fa. I say that lines ea and ag are equal.
combustion [is] at [a distance of] 30. And similarly for other cases. For if they are not equal, one of them will be greater. First let ea be
Whence, if you find individual lines of order, you will [be able to] see greater. Therefore, let it (ea) be cut in / so that al = ag, and let agq be
at what distances combustion would take place [with respect to rays made equal to ale [see Fig. 5.12a]. Whence, gq —le. And let line of order
reflected from the extremities of the lines of order]. But you will find the Ik be drawn, which line will be parallel to line ed, and let k be on the
magnitudes of the lines of order from the fourth [proposition of the parabolic section ado. And it will be necessary that Ik does not reach to
Speculi almukefi compositio], evidently by multiplying the [segment of line gd, since, if Ik were to reach it, then there would be two points in
the] axis [intercepted by the line of order] into the latus rectum— and the line gd which would be common to the surface of the cone but not in a
root of the product is the line of order. Add its square [i.e. of the line line that goes through the apex of the cone. Therefore, by the second
of order] to the square of the residual segment of the axis [intercepted] proposition [of the Speculi almukefi compositio], line gd would cut the
between the line of order and the point of intersection of the axis of the cone; therefore it would not be tangent, which is against what was posited.
section with the latus rectum. And the root of that sum will be the distance Therefore let Ik be continued until it meets gd in m. And so, because
of combustion. Here is a table [given with the Latin text]. ed2 / Ik2 = ea I al (by the fourth proposition) and ea / al = 2 ea / 2 al
= qe / gl, therefore ed2 / Ik2 = qe / gl. But since Im > lk , by V.19 [of
[Notes for the Table] Euclid] ed I Im < ed / Ik. But ed / Im = eg / gl. Therefore, eg / gl < ed / Ik.
And so you will first draw (according to whatever measure you please) But ed / Ik = (ed2 / lk2)il2. Therefore, eg / gl < (ed2 / Ik2)112. But qe / gl
lines of order upon the axis by extending [them indefinitely] as far as you = ed2 / Ik2. Therefore, eg I gl < (qe / gl)'12. And with eg posited [as a
please, and by means of this table you will be able to cut them off according mean] between qe and gl, it follows, since qe / gl = (qe / eg)·(eg / gl)
to the quantities Im [see Fig. 5.11], the lines Im being dependent on the and eg I gl < (qe / g l)112, that necessarily qe / eg >(qe / gl)'12, and hence
assumed quantities el, or [you will proceed] as follows by taking [from that qe I eg > eg I gl. From this and the penultimate of those things added
the table] the [various] quantities hm [dependent on the corresponding by Campanus to the fifth [book] of Euclid, qg (the remainder) / el (the
assumed quantities el] and placing the immobile foot of a compass always remainder) > qe (the whole) / eg (the whole). But qe > eg. Therefore,
on h, the point of combustion; then, with the mobile foot of the compass, a fortiori qg > el and aq > a e . But these [aq and ae] were already [posited
cut off [on the previously erected indefinitely long perpendiculars on points as] equal. Therefore, it cannot be that ea > ag.
/] the line[s] of order Im. Or still another way is permissible from Prop­ Next let ea < a g [see Fig. 5.12b]. Let ag be cut in r so that ar —a e ,
osition 5 [of the Speculi almukefi compositio]: Add Vi el to Vi the latus and let aet - arg. Whence et and rg will be equal. And let line of order
rectum and extend a compass according to the resulting quantity. Then tn be drawn. By reasoning similar to that before, it will be necessary that
with its immobile foot placed in line ex and the mobile foot on [point] e, it does not reach to line gd continued. Therefore let it (tn) be continued
you will cut off the line[s] of order Im [on the previously erected indefinitely to meet it in x. And so, because tn2 / ed2 = ta / ae (by the fourth proposi­
long perpendiculars on points /]. So you will find the same thing in three tion) and ta / ae = 2 ta / 2 ae = tg / er, therefore tn2 f ed2 = tg / er. But
since line tx > tn, tx / ed > tn / ed. But tx / ed = tg / ge. Therefore, tg / ge
230 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ ADDITIONS 231

> tn / ed. But tn / ed = (tn2 / ed2)112 = (tg / er)112. Therefore, tg / ge because dq is equal to dc and ac = Idf. With/? one unknown, hence d f /
> (tg / er)1/2. And so with ge posited as a mean between tg and er, it is unknown = dq2 / eg2. Therefore eg2 will be known in terms of the un­
necessary that tg I ge > ge I er. Therefore, te (the remainder) / eg (the known; and a quarter part of eg2is equal to the square of the unknown [i.e.,
remainder) > tg (the whole) / ge (the whole). But since tg > ge, a fortiori f e 2 = eg2 / 4 from the property of the parabola]. Thus the square of the
et > e g . Thence arg < a e t. But these [arg and aet] were already equal. unknown equates [i.e. determines] the unknown [i.e.,/? / df = 4fe 2 / dc2,
Therefore, it cannot be that ea < a g . From these [arguments] is inferred or fe = dc2 / 4df, and, since d f = ac / 2 and dc and ac are known,] there­
the necessity of the proposal, [namely] that ea and ag are equal. fore f e will become known, which was sought. (But [it is found] more
[Or] you will posit this proposition per se. And you will also posit its con­ easily as follows: Because dq is the mean proportional between d f and the
verse per se. And the converse can be easily demonstrated as follows. Let latus rectum, hence the latus rectum will be known.)1 (You may do the
ae and ag be equal, and let gd be drawn to the extremity of the line of order same thing in every cone whose base is round, even if the cone is inclined or
ed. I say that gd is tangent to the section in point d. For if it is not tangent in oblique, so long as at least two sides of the triangle are known.)2
that point, it is possible to draw from point g another tangent which will be For example, in the case of a [parabolic] section of an obtuse-angled
tangent in some point either between a and d or between d and o. With cone, let either of the sides ab and ac in triangle abc be 72, and let half the
either [possibility] posited, it follows from the seventh [proposition] that a base be be 60, and so dq will be 60 and d f 36. Let fe be one unknown. [So]
part is equal to [its] whole. For the part of the axis between the line of order 36 / unknown = 3600 / eg2, 3600 being the square of dq. Wherefore, eg2
drawn from the point of tangency and the vertex of the section will be equal = 100 unknowns. [And thus eg2! 4 = 25 unknowns = 15fe , and eg2! 4 also
to ag. Therefore, ae, the part, will be equal to the whole or the whole to equalsf e 2since eg = Ife from the property of a parabola. Hence] the fourth
the part. part of 100 unknowns, i.e. 25 unknowns (25/?), equals the square of one un­
known (fe2). Therefore, fe ~ 25, which was sought. If it (the parabolic sec­
[3] tion) is in an obtuse-angled cone, you will find that the latus rectum, which
is quadruple the axial segment which it intercepts, will fall between the
Note concerning the fourth proposition [of the Speculi almukefi com­
intersection of the axis [of the cone] with the axis [of the section] and the
positio], which speaks in regard to a parabolic section that is made in a
middle point between the vertex of the section and the intersection of the
right-angled right cone, if the parabolic section will be made in an acute-
axis of [the section] with the axis [of the cone]; and the closer it (the latus
angled or obtuse-angled cone and you would find the latus rectum in it, the
rectum) approaches d the more obtuse is the angle at a . But in the acute
latus rectum being among the lines of order because the latus rectum would
angle a , the closer it (the latus rectum) approaches/ the more acute is angle a .
be quadruple the segment of the axis [intercepted] between the vertex of
the parabola and this latus rectum, then the whole demonstration would
proceed just as it does in connection with a right-angled cone. Now, that it [4]
is possible that this be done in an acute-angled [cone] accept as follows.
Let there be a triangular section abc proceeding through the apex of a Let eh be one quarter of the latus rectum [see Fig. 5.14] ; let hg, standing
cone [Fig. 5 .13a]; let the axis [of the cone] be ad. Let the parabolic section, perpendicularly on it, be one half of the latus rectum, and let / be the point
that is the equidistant one perpendicular to the triangle, be pfgq, in which in the axis [of the section] on which the line of order Im stands. Then with
the axis f d is parallel to side ac of the triangle. Moreover, I wish that ac be line Im drawn, I say that, if / is between e and h, Ih + hm - hg. I shall ex­
equal to be so that Aabc is isosceles [in fact, equilateral]. Then ac will be tend he to p so that ep = le. By II.8 of Euclid, p e -4eh + lh2 = ph 2. But
made double d c . And so dc and d f will be equal. Next, letf e be one-quarter pe-4eh = Im2, or, because pe = el, el-4eh = Im2. Thus Im2 + lh2 = ph2.
part of axis fd. With peg having been made a line of order, I say that pg
= 4fe. Fordq2 = dc2 = 4eg2. Therefore, dc = leg. Wherefore, d f = leg. 1 This remark which I have included in parentheses is written in the margin, with an ar­
Therefore, d f = pg, and so pg is also quadruple fe. row pointing to its location in the text.
2 This parenthetical note is also written in the margin, but above the preceding remark.
Now, you will find a latus rectum to fall [in a parabolic section arising] However, it belongs in the text below the former since it is a general note that applies to the
in every acute-angled cone because the whole [latus rectum] is quadruple whole procedure. The fact that Regiomontanus concluded that the procedure may be applied
the [segment of the] axis which it intercepts. It will fall [in such sections] be­ to oblique cones does not warrant the inference that he had read the Greek Text of the Conics.
tween the point which is called the vertex of the section and the middle It seems only to indicate that he had understood the more general generation of parabolas
from oblique cones as given in the fragment translated from the Arabic text of the Conics by
point between the vertex and the intersection of the axis [of the section] and Gerard of Cremona, and that he had seen, on his own, that the procedure could be extended to
the axis [of the cone], e.g., between/and h. With the sides of the triangle such cones. In fact, in Passage [6] below, Regiomontanus does extend these procedures of
abc known, dq and d f will be known in any triangle whether acute or obtuse calculation to parabolas generated from any cone.
232 ARCHIMEDES IN THE MIDDLE AGES REGIOMONTANUS’ ADDITIONS 233

But lm2 + lh2 = hm2. Therefore, hm = hp. But 2eh = hg and ph + hi double the axial segment which it intercepts. Let it be% , so that the whole
— 2eh. Therefore, ph + hi = hg. Hence mh + hi = hg, which was pg is the latus rectum and is quadruple the axial segment flt. I say also that
intended. here mn is the mean proportional between pg andfm , which you will prove
But if the line of order lm follows h, the point of combustion, as in the in the same way as it was demonstrated above in the fourth proposition
second figure placed here [see Fig. 5.15], I say thatg/ι + hi = hm. For let [of the Speculi almukefi compositio]. For pg2 / hg2 = pg / fh, since each is
le be continued to 5 so that es = eh. Once more, Is2 = le-Aes + hi2, where a quadruple ratio. Also hg21mn2 = hflfm . Therefore, pg 2/ mn2 = pg Ifm.
Is is the whole composite [magnitude], or Is2 = le -Aeh + lh2. But le -4eh But pg / mn = (pg2 / mn2)112. Hence pg / mn = (pg / fm )112. Therefore, mn /
= lm2, by the fourth [proposition] of this [Speculi almukefi compositio]. fm = (pg /fm )112. Therefore [pg / mn = mn /f m , and so] mn is the mean pro­
Therefore, as before, Is = hm. But hs = hg since each is one half the latus portional between fm and the latus rectum (pg) of the section. Thence you
rectum. Therefore, gh + hi = si = hm, which was proposed. then deduce all the things that can be done in a parabola produced in a right
From these [considerations] you will conclude the marvelous uniformity cone. Moreover you will find the quantity fh from the known sides ac and
of the relationship between the lines of order and the axial segments they b e , for dq and d f thus become known, since dq is equal to bd or dc and df is
intercept in every parabolic section, whether it is made from a right-angled ac / 2. But dq2 / hg2 — d f /fh. And so if you pose hg as one unknown, you
or an obtuse-angled or an acute-angled cone, just as in every circle there is will find what was proposed. Or [proceed] more easily as follows: Since
uniformity of relationship between the right-sines and the versines of line of order dq is the mean proportional between axis d f and the latus rec­
similar arcs. Whence it is marvelous that the curvity of the section having tum, therefore the latus rectum will consequently be produced by dividing
the same relationship in every section whether small or large will neces­ the square of dq by the axis df. Hence its quarter is fh.
sarily remain the same.

[5]
In an annular parabolic mirror, with the radii of the circles which are
the bases of the mirror as well as the depth of the mirror given, to find the
place of the concourse of the reflected rays, the axial segment [cut off by
the latus rectum] and also the latus rectum [itself].
Let the [full] axis of such a section be ag [see Fig. 5.16]; let there be given
the lines of order bd and ce, as well as the axial segment they intercept,
namely, be. It has been proposed to find the axial segment ab and thence
the latus rectum. Because bd2 = ab -latus rectum, and ce2 = ac ■latus
rectum, therefore ce2 / bd2 = ca / ab. Therefore, (ce2 - bd2) / bd2 = cb /
ba. But the three first terms are known; therefore ab is known. Thence
[known are] the latus rectum and its quarter, which were sought.

[ 6]

To elicit a parabolic section from any cone whose base is a circle so long
as the sides of the triangle that bisects it (the cone) are known.
Understand a cone of this sort cut through the axis by a plane surface
perpendicular to the base. Let the common section be Aabc [see Fig.
5.17], whose base diameter will be be, in which the center of the base is d.
Let another plane surface perpendicular to Aabc cut the cone so that the
common section of it and the triangle is line d f parallel to side ac. The com­
mon section of the second plane surface and the curved surface of the cone
is a curved line containing the parabolic section. For let there be desig­
nated in it lines of order mn, hg and dq [all] perpendicular to axis fd. It is
possible to find some line of order on the axis [of the section] which is
 CHAPTER 6

The M edieval Traditions o f Conic

Sections in the Early Sixteenth
Century: Giorgio Valla, Johann
Werner, and Albrecht Diirer

At the end of Chapter 5, I mentioned Regiomontanus’ ownership of

a copy of the Greek text of the Conics of Apollonius and his unfulfilled
hope of publishing it. Furthermore, it appears that at least four other
copies of the Greek text were made in the fifteenth century.1 I also men-

1 Of the five pre-fifteenth-century manuscripts of the Conics listed by Heiberg in the

prolegomena to his edition (Gr 1, pp. IV -V ; 2, pp. XI-XII) only two played a historical
role in the fifteenth century, namely Vat. gr. 206 (Heiberg’s V) and Vat. gr. 203 (Heiberg’s
v). Heiberg believed (Gr 2, p. LXXX) that the codex that Francesco Filelfo found in the
East and mentioned in a letter of 1427 was one of three manuscripts (the aforementioned
V and v and Paris, BN grec 2342 [Heiberg's MS p]). If I am correct in my belief (see above,
Chap. 3, n. 32) that v had been part of the papal collection at the time of Moerbeke’s
translating activity (ca. 1270), then we would have to conclude that Filelfo’s manuscript
was either V or p. And since, as we shall see, there is evidence for the existence of V at
the papal library (from at least 1475) and no evidence of the whereabouts of p at this time,
we may perhaps conclude that Filelfo’s manuscript was V. I confess, however, that I have no
idea as to how V came into the papal library. There is one further difficulty in tracing
the fortunes of V and v. If Heiberg is correct in his opinion that v was copied from V,
and if v passed to the papal library in the thirteenth century, and if V was always in the East
until the fifteenth century, then v would have been copied from V in the East before it (v)
came to the papal library in 1266 as part of the Angevin gift. While this chain of reasoning
is highly speculative, it is not entirely fanciful, for I believe that I reasoned correctly
(though tentatively) in Chap. 3 above when I concluded that some access to a Greek text
of the Conics on the part of Moerbeke and/or Witelo was required by Witelo’s treatment
of conic sections.
The fate of V and v after 1450 is not much clearer. In an inventory of the papal books made
for Calixtus III (1455-58), among the books noted as loaned to Cardinal Bessarion, is the fol­
lowing: “ Quidam (s/c) geometrica Theodosii et Apollonii in pergameno (R. Devreesse, Le
Fonds grec de la Bibliotheque Vaticane des origines a Paul V [Citta dei Vaticano, 1965],
p. 41).” Presumably Bessarion had his copy of the text made from this manuscript, the copy
which is still extant as Venice, Bibl. Naz. Marc. Z. graecus 518, the first of the fifteenth-
century manuscripts I mentioned above in the text. Now Heiberg concluded (Gr 2, p. XVIII)

235
 EARLY SIXTEENTH CENTURY 237
236 ARCHIMEDES IN THE MIDDLE AGES
igitur prisci illi in singulis triangulorum speciebus inspexerunt duos rectos, prius in aequi-
tioned the lost Latin translation of the Greek text of the Conics apparently latero, et rursus in isoscele, postremoque in scaleno, ita posteri generalius theorema, quod
executed by Bernhard Walther. Hence it seemed as if the text of Apol­ praeceptum nostri vocant, id demonstrarunt, omnis trianguli internos tres angulos binis rectis
aequales esse, ita et in coni sectionibus. Orthogonii coni sectionem, orthogonio soli cono
lonius was about to be the focus of an expanded study of conic sections.
inspexerunt secto plano recto ad latus unum coni inesse. Amblygonii coni sectionem in ambly­
But, alas, this was not immediately the case. Only a very limited use of gonio factam cono demonstrarunt, ut oxygonii in oxygonio, similiter in omnibus conis plana
Apollonius’ text can be detected before the appearance of Giambattista agendo recta ad latus coni unum; quod prisca linearum indicant nomina. Verum postea
Memmo’s Latin translation of it in 1537, while the medieval traditions Apolonius (/) pergaeus in totum inspexit, in omni cono tam recto quam non recto omnes
of Alhazen’s De speculis comburentibus, the anonymous De duabus lineis, sectiones esse iuxta plani at (/ ad) conum differentem ductum, quamobrem illius temporis
admirati homines mirificam demonstrationem suscipiendo magnum geometren ipsum
and the anonymous Speculi almukefi compositio continued to flourish
vocitarunt. Quod igitur Geminus refert ita esse paucis ostendamus quo ad alia geometrica
in the first half of the century, as we shall see in this and the succeeding considerando nobis sit proclivior accessus. Sit ex axe coni triangulum abe, ducaturque
chapter. ipsi ab a puncto e ad angulos rectos de, et per de planum iniectum rectum ad ab secemus
But let me turn first to the modest use of Apollonius’ Conics and Euto- conum. Rectus igitur quilibet qui sub aed, aef angulis. Rectangulo manente cono rectoque
cius’ commentary thereon by Giorgio Valla in his De expetendis et fugien­ inquam qui sub bac angulo, ut in prima descriptione, duobus rectis erunt aequales qui sub
abe (! bac), def (! a ef) anguli. Itaque parallelus erit de\f] ipsi ac. Sitque in coni superficie sectio
dis rebus opus (Venice, 1501). It is clear that Valla’s limited treatment
appellata parabole, nempe nomen adepta quod parallelus sit def, quae communis est sectio
of conic sections was based entirely on reading his own copy of the Greek planum (/ plani) secantis et ex axe trianguli, ipsi ac lateri trianguli. Quod si amblygonius
texts of Apollonius and Eutocius.2 Valla’s principal section, entitled De fuerit conus in secunda descriptione existente sub bac, recto autem sub def (! aef) duobus
comica (!) sectione, consists of the following twelve passages, all trans­ rectis maiores erunt qui sub bac, [ac/] anguli. Proinde non coincidet aef (! def) ipsi ac lateri in
lated from Eutocius or Apollonius (with an occasional side remark by partes quae sunt ad/ , [c], sed in eas quae ad a, e producta ca ad d. Faciet igitur planum
secando in coni superficie sectionem appellatam hyperbolen, ita nimirum dictam quod dictos
Valla):3
excedat angulos nempe qui sub aef, bac duos rectos, vel quod excedat e/, def verticem coni
et coincidat ipsa ca extra. Sin oxygonius sit conus acuto existente [qui sub bac], qui sub bac,
that Bessarion's manuscript was copied from V. But V hardly fits the description of the aef duobus rectis erit (/ erunt) minores; c/, ac productae coincident ubivis, nam augere conum
manuscript loaned to Bessarion, for Theodosius’ work is not included in V and, further­ possum. Erit igitur in superficie sectio quae nuncupatur ellipsis, ita nominata aut quod predicti
more, V is written on paper while the borrowed manuscript was described as being on parch­ anguli a duobus deficiant rectis aut quod ellipsis sit circulus imminutus. Ita nimirum prisci
ment. Manuscript v seems a somewhat better candidate for the borrowed manuscript since mathematici planum secans per def ad rectos angulos ipsi ab lateri per axem ipsius coni
it does include Theodosius’ Sphaerica as the first item. But it too is a paper manuscript, and trianguli et insuper conos differentes inspexerunt.
so unless the description of the loaned book is in error (which is of course possible) MS v “ [2] Caeterum quo sint quae dicuntur evidentiora: (1) Si a puncto aliquo ad circuli circum­
would also have to be rejected as the volume loaned to Bessarion. ferentiam qui in eodem plano non est puncto, recta linea coniuncta utrinque porrecta, et
Whatever was the source of Bessarion’s manuscript, I suspect that, despite the doubts of puncto manente circumacta recta linea circa circuli circumferentiam ad idem rursus con­
Heiberg (Gr 2, p. XX), Bessarion’s manuscript was the principal source of Regiomontanus’ vertetur unde ferri incepit, descriptam a recta linea superficiem quae conficitur ex binis super-
copy of the Conics, Nuremberg, Stadtbibl. Cent. V, App. 6. At any rate, we can note Regio­ ficiebus ad verticem invicem positis, quarum utraque augetur in infinitum descriptricis (!)
montanus’ codex as the second of the fifteenth-century copies. rectae lineae in infinitum productae vocant mathematici conicam superficiem. Verticem
MS V is first unambiguously mentioned in the papal inventory of 1475 (Devreesse, Le vero ipsius ibi eminens punctum. At axem per punctum et centri circuli actam rectam
Ponds grec, p. 60, item 351). Heiberg believed that MS V was the source of another fifteenth- lineam. (2) Conum porro figuram comprehensam a circulo et circumferentia (/ ea) quae inter
century copy, Paris BN suppl. gr 451 (Gr 2, p. XIX). MS v ’s presence in the papal library is not verticem circulumque est conicae superficiei, verticem coni punctum quod superficiei
revealed until the catalogue prepared by Fabio Vigili of Spoleto (about 1510; see Devreesse, quoque est vertex, axem vero a vertice ad circuli centrum actam rectam lineam, at basim cir­
Le Fonds grec, p. 163), unless, of course, the codex which Bessarion borrowed was MS v. culum. (3) Conorum porro rectos vocant qui ad rectos angulos ad basim axes habent, scalenos
MS v was the source of Giorgio Valla's copy of the Comes, Modena, Bibi. Estense II D 4(see qui non ad rectos angulos basibus axes habent. (4) Omnis inflexae lineae quae in plano uno
Gr 2, p. XXI) and apparently also of Oxford, Bodl. Libr., Canon, gr. 106 {ibid. ), the last two est diametrum rectam lineam vocant quae ab inflexa ducta linea omnes actas in linea rectas
of the fifteenth-century copies mentioned by Heiberg. I shall not detail here the large number lineas recta aliqua linea parallelos bifariam dispescit, verticem vero lineae extremum rectae
of copies made of the text in the sixteenth century, but see Gr 2, pp. X I-X IV , MSS 3 ,4 ,9 , lineae quod ad lineam est structim utique ad diametrum agatur quaevis parallelus. (5)
10, 12, 13, 14, 15, 19, 21, 24, 25; for their sources, see pp. XVI-XXI. Similiter binae inflexae lineae in plano uno positae diametrum insuper dixere obliquam
2 As previously noted, this is Modena, Bibl. Estense II D 4. It contains Valla’s ex libris and quae recta linea binas secans lineas omnes productas in utraque linea ad quampiam rectam
an indication of its later ownership by Alberto Pio di Carpi (Gr 2, p. XII). Like MS v it in­ lineam bifariam secat. Vertices autem linearum diametri ad lineas extrema. Rectum (/
cludes both Eutocius’ commentary and the text of the Conics. As we shall see later (see foot­ Rectam) autem angulum (/ lineam) qui (/ quae) positus (/ posita) inter binas lineas cunctas
note 5 below) some of Valla's knowledge of conic sections came from his ownership of Gr MS parallelos actas rectas lineas rectae alicui lineae admissas intra lineas bifariam secuerit,
A of Archimedes' works and Eutocius’ commentaries thereon. structim ad diametrum adiguntur singulae paralleli. (6) Syzyges vocant diametros binas in­
:i G. Valla, De expetendis et fugiendis rebus opus (Venice, 1501), Bk. XIII, chap, iii, sig. flexae lineae, et duas inflexarum linearum rectas lineas, quarum utraque diameter existens
x iiii recto-[x vi recto]: “ De comica (!) sectione. [1] Geminus in sexto mathematicarum alteri parallelos bifariam dispescit. (7) Dein axem vocant inflexae lineae et binarum inflexarum
praeceptionum ait priscos mathematicos conum definientes orthogonii trianguli ambitum linearum rectam lineam quae cum sit diameter lineae linearumve ad angulos rectos paral­
media una circa rectum angulum conos omnes rectos fieri arbitratos, unamque in singulis lelos secat. (8) Syzyges vocant axes inflexae lineae et binarum inflexarum linearum rectas
sectionem; in orthogonio quidem parabolen appellatam, in amblygonio autem hyperbolen, lineas quae cum sint diametri syzyges ad rectos angulos parallelos invicem secant.
at in oxygonio ellipsin; atque ita nominatas apud ipsos sectiones est invenire. Quemadmodum
238 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 239

“ 13] Si namque rursus ut in proxime memoratis descriptionibus secans planum kel. Com­ circumferentiam; nam si intra capiatur eadem evenerit absurditas aliudque ab eo positum est.
munis autem eius sectio et basis coni kfl, communis rursus ipsius kel plani et abc trianguli ef, Coniungantur m a , m b , mf, ponaturque ut c ad d ita am ad mb. Est igitur ut e d add ita quod ab
diameterque vocatur sectionis. Plus (/ in Gr MS E st.; Omnium) igitur sectionum subiicit kl ad e d ad id quod est ab c et quod ab am ad id quod est ab m b . Atqui ut e d ad d ita positum est af
rectos angulos ipsi bc basi ipsius abc trianguli. Relinquitur autem si e f parallelus fuerit ipsi ad/b; et ut igitur a f adfb quod ab am ad id quod est ab mb. Etiam ob ea quae prius sunt demon­
ac parabolen fieri kel in superficie coni sectionem; sin coincidet ipsi ac lateri ef extra verticem strata, si ab b ipsi am parallelum duxerimus, demonstrabitur af ad fb quod ab a f ad id quod
coni velut ad d, fiet kel sectio hyperbole; sin intus coincidat ipsi ac, ef, fiet (correxi ex fiet ef) est abfm . Atqui demonstratum est etiam af adfb quod est ab a f ad id quod est abfh. Aequalis
sectio ellipsis, quam thyraeon vocant. In totum igitur paraboles diameter parallelus est uni ergo fh ipsi fm , quod est impossibile.
lateri trianguli; hyperboles autem diameter coincidit lateri trianguli ut in partibus quae sunt ad “ [6] (Prop. 1.1) Ad (/ Ab) verticis (/ vertice) conicae superficiei actae rectae lineae ad
verticem coni; at ellipsis ipsius diameter lateri trianguli coincidit quemadmodum in partibus puncta quae in superficie in ipsa sunt superficie. Sit conica superficies cuius vertex a punctum,
quae sunt ad basim. Nec nos lateat parabolen et hyperbolen augeri in infinitum posse, at non capiaturque aliquod punctum in conica superficie b, coniungatur quaepiam recta linea acb.
ellipsim, quandoquidem omnis ellipsis in sese ut circulus vergit. Aio acb rectam lineam esse in superficie. Quod si fieri possit, ne sit, sitque describens super­
“ [4] Carpit Euclidem Pergaeus Apollonius, non, ut Pappus et quidam alii censuerunt, quod ficiem recta linea de, circulus autem in quo feretur ed sit ef. Si igitur manente a puncto de recta
medias duas non invenerit proportionales. Invenit siquidem Euclides mediam unam pro­ linea feretur ad ipsius ef circuli circumferentiam, ibit per b punctum, eruntque binarum
portionalem, at ut ipse quoque ait de binis mediis, querendum sibi non existimavit in ele­ linearum eadem extrema, quod alienum est. Non ergo ab eodem a ad b coniuncta linea recta
mentis, sed Apollonius de duobus punctis (/ in Gr MS E st.; mediis) quaerere Eutotio visus est in eadem superficie non est; in superficie ergo. Manifestum quod si a vertice ad aliquod punc­
Ascalonitae. In altero autem libro qui non modo ad nos sed ne quidem ad Eutotium pervenit tum intra superficiem coniungatur intus aliqua recta linea, casuram intra superficiem conicam;
incessit Euclidem quem de locis inscripsit ab Euclide scriptis. et si extra connectatur, extra superficiem fore.
“ [5] Priscorum namque geometrarum fuit consuetudo dicere in quaestionibus cum non “ [7] (Prop. 1.3) Si conus plano secetur a vertice, sectio triangulum est. Sit conus cuius
modo ab uno puncto verum etiam in pluribus fit quaestio, velut si in ordine rectae lineae datae vertex quidem a punctum, basis vero bc circulus, secetur plano aliquo ab a puncto, faciatque
finitae velint punctum invenire a quo acta perpendicularis in datam media sit proportionalis sectiones in superficie ab, ac lineas, in basi autem bc rectam lineam. Aio abc triangulum esse.
sectionum id sane locum vocant; nec enim unum modo punctum est faciens quaestionem, sed Nam quoniam ab a in b coniuncta est communi[s] sectio est secantis plani et coni superficiei,
locus totus quem habet circumferentia eius qui est circa diametrum datam rectam lineam recta linea igitur est ab ; similiter autem etiam ac ; est autem etiam bc recta linea. Triangulum
circuli. Nam si in data recta linea semicirculum descripseris, quod in circumferentia punctum igitur est abc. Si ergo conus plano aliquo secetur a vertice, sectio triangulum est.
sumpseris a quo per perpendicularem ducas in diametrum, feceris quaestionem quod “ [8] (Prop. 1.5) Si conus scalenus plano secetur ab axe ad rectos angulos basi, secetur
problema nominamus. Similiter data recta linea si quis invenire constituatur extra ipsam etiam altero plano ad rectos quidem angulos ab axe triangulo, auferendo autem a vertice
punctum a quo coniunctae ad extrema rectae lineae aequales inter se invicem erunt, atque in triangulum simile ab axe triangulo, subcontrarie autem posito, sectio circulus est. Vocatur
hoc non punctum unum modo problema facit, sed quantum obtinet quae acta est a bifariam vero huiusmodi sectio subcontraria. Sit conus scalenus cuius vertex a punctum, basis autem
secta ad rectos angulos; nam si seces bifariam datam rectam lineam et a bifariam facta sec­ bc circulus, seceturque plano ab axe recto ad bc circulum, faciat utque abc triangulum. Secetur
tione ad angulos rectos ducas, quod in ipsa punctum sumpseris faciet quod instituimus. nimirum etiam altero plano ad rectos angulos existenti ipsi abc triangulo, auferendo autem
Tradit autem Apollonius in loco resolutorio de subiecto. Binarum datarum rectarum linearum
triangulum ad a punctum [a]kg, simile quidem abc triangulo, subcontrarie vero positum,
in plano et puncto (/ in Gr MS Est. ; punctis) et ratione data inaequalium linearum describi hoc est ut aequalis sit qui sub akg angulus ei qui sub abc; et faciat sectionem in superficie
posse circulum ut a datis punctis ad circuli circumferentiam fractas lineas rectos (/ rectas) ghk lineam. Aio circulum esse ghk lineam. Capiatur namque puncta quaepiam in ghk, bc
rationem habere eandem cum dato. Sint namque data puncta a, b, ratio autem c ad d, ut sit c
linearum quae sint h , l , et ab h , l punctis in id quod est ex abc triangulo planum perpendiculares
maior. Oportet quod proposuimus efficere hoc modo. Coniungatur ab producaturque in
agantur, cadent nimirum in communes sectiones planorum. Cadant ut f h , Im; parallelus
partes b , fiatque ut d ad c et c ad aliam aliquam maiorem quam sit d, ac sit ut putae d; rursusque
igitur est fh ipsi Im. Agatur utique e x / ipsi bc parallelus dfe; est autem etiam fh ipsi Im paral­
fiat ut [e] est ad ab insuper d ad bf et c adg. Non dubium quine media sit proportionalis ipsius
lelus; ergo exfh,de planum parallelus est basi coni. Circulus ergo est cuius diameter de. Ergo
e d et d, necnong ipsarum af, fb. Et centro/, intervallo autem g, circulus describatur kh. Non
aequale quod sub ipsis df,fe ei quod est exfli. Et quoniam parallelus est ed ipsi bc, qui sub
dubium quin secet kh circumferentia ab rectam lineam; nam recta linea media proportionalis
ade angulus aequalis est ei qui sub abc ; at qui sub akg ei qui sub abc positus est aequalis; et
est ipsarum af,fo (/fb). Capiatur in circumferentia verbi causa punctum h, et coniungantur
qui sub akg igitur ei qui sub ade est aequalis; atqui sunt etiam qui ad/punctum aequales in
ha,hb, hf. Aequalis igitur est hf ipsi g ; et ob hoc est ut a f ad fh ipsumfh ad f b . Et circa eundem
vertice. Totum ergo simile est dfg triangulum ipsi kfe triangulo. Ergo ut ef ad fk ita g f ad fd.
angulum qui sub hfb sunt proportionales. Simile igitur est afh ipsi hbf triangulo et aequalis
Ergo quod sub efd aequale demonstratum est ei quod est ab eh; et quod sub ipsis kf,fg igitur
qui sub fhb angulus ei qui sub hab. Agatur itaque per b ipsi ah parallelus bl. Quoniam igitur
aequale est ei quod est ab/b. Similiter demonstrabuntur omnes quae ab gbk linea ing£ actae
est ut a f adfh ipsa hf adf b , ut etiam igitur prima a f ad tertiamfb quod ab a f ad id quod est abf h .
perpendiculares, quae possint aequale ei quod sub sectionibus gk. Circulus ergo est sectio
Atqui ut a f ad fb ipsa ah ad bl; et ut igitur quod est ab a f ad id quod est abfh verum ut a f ad fb
ipsa ah ad bl; et ut igitur quod est ab a f ad id quod est ab fh ipsa ah ad bl. Rursus quoniam cuius diameter gk.
aequalis est qui sub bhf ei qui sub hab, est autem etiam qui sub ahb ei qui sub hbl aequalis, “ [9] (I. Defs. II-1) Hyperboles et ellipsis utriusque diametri bifariam facta sectio centrum
vicissim namque id, reliquus igitur reliquo aequalis est, et simile est ahb ipsi bhl et propor­ sectionis appellatur, a centro ad sectionem procidens ex centri (/ centro) sectione (/ sec­
tionis). (2) Similiter etiam oppositarum bifariam facta sectio lateris obliqui centrum appella­
tionalia sunt latera quae circa aequales angulos ut ah ad hb ipsum hb ad bl, et ut quod ab ah ad
id quod est ab bh ipsum ah ad bl. At erat etiam ut ah ad bl quod [a]b a f ad id quod est ab f h ; ut tur. (3) A centro acta instructim immissam quae mediam habeat rationem speciei laterum et
ergo quod ab [a]/ad id quod est abfh ipsum ab ah ad id quod est ab bh, et ob hoc ut a f ad.fh ip­ bifariam dissecta a centro secunda diametros a geometris appellatur.
sum ah ad hb. Verum ut af ad fh etiam e d ad c et c ad d; et ut igitur c ad d ipsum ah ad hb. “ [10] (1.17) Si in sectione coni a vertice lineae acta fuerit linea recta in instructim immis­
Itidem demonstrabuntur omnes quae ab a, b punctis ad circumferentiam circuli fractae ean­ sam, extra sectionem cadet. Sit coni sectio cuius diameter ab. Aio a vertice nempe ab a puncto
dem habent rationem cum a (! in Gr MS Est.; c), d. Aio nimirum ad aliud punctum non in instructim immissam actam lineam rectam extra sectionem casuram. Quod si fieri possit,
existens in eadem circumferentia non fieri rationem eorum punctorum quae sunt ab a, b con­ cadat introrsum ac. Quoniam igitur in coni sectione sumptum est verbi causa punctum c , igitur
nexis rectis lineis eadem (/ eandem) ei quae est ipsius c ad d. Quod si fieri possit fiat ad m extra a c puncto intra sectionem acta in instructim immissam (correxi ex in missam) committitur
240 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 241

[1] Eutocius, Commentary on the Conics of Apollonius, on Bk. I. Cf. [4] Eutocius, Commentary. Cf. ed., Heiberg, Gr 2, p. 186, lines 1-10.
ed., Heiberg, Gr 2, pp. 168 (line 12)-174 (line 15). This passage concerns In this passage Eutocius notes that Apollonius castigates Euclid, not be­
the pre-Apollonian method of the generation of conic sections by varying cause he had not found two mean proportionals, as Pappus and others
the vertical angle of a right cone. It also introduces for Latin readers thought, but for what Euclid wrote in a work on Loci. This work existed
Eutocius’ false etymologies of the terms parabola, hyperbola and ellipse. in Apollonius’ time, but, as Valla notes, was not extant in Valla’s time.
[2] Apollonius, Conics, Bk. I, First Definitions. Cf. ed., Heiberg, Gr 1, [5] Eutocius, Commentary. Cf. ed., Heiberg, Gr 2, p. 178 (line 19)-
pp. 6 (line l)- 8 (line 20). These are the definitions which Gerard of Cre­ 184 (line 20). This passage initially gives two examples to explain plane
mona also translated from the Arabic (Chap. 1, n. 5). loci, i.e., problems where the solution is effected not by one point alone
[3] Eutocius, Commentary. Cf. ed., Heiberg, Gr 2, pp. 174 (line 18)-176 but by a multiplicity of points. It adds a further example, with solution,
(line 16). This continues passage [1] and describes Apollonius’ generation of such a problem presented by Apollonius in his [De] loco resolutorio
of conic sections from any cone by making the cutting plane (which is (to use Valla’s translation of the title). Thus Valla has given in passages
perpendicular to the axial triangle) parallel to a side of the triangle (pro­ [1], [3], [4] and [5] what he conceives to be the heart of the introductory
ducing a parabola), or having it meet a side beyond the apex (producing matter of Eutocius’ Commentary.
a hyperbola) or below the apex (producing an ellipse). [6] Apollonius, Conics, Bk. I, Prop. 1. Cf. ed., Heiberg, Gr 1, pp. 8
(line 21)-19 (line 14).
ipsi ab diametro ac bifariam ab ipsa secabitur; ac igitur educta bifariam secabitur ab ab, quod [7] Ibid., Bk. I, Prop. 3. Cf. ed., Heiberg, Gr 1, pp. 12 (line 21)-14
est absurdum; educta namque ac extra sectionem cadent (! cadet). Non ergo ab a puncto in
instructim actam producta recta linea intra lineam cadet. Ergo extra; ob hoc attinget sectionem.
(line 7).
"[11] (11.38) Si oppositorum binae rectae lineae coincidentes confluxerint, quae a coin­ [8] Ibid., Bk. I, Prop. 5. Cf. ed., Heiberg, Gr 1, pp. 16 (line 24)-20
cidentia copulata fuerit mediae contactus connectenti diameter erit oppositorum que (line 8).
nominatur orthia, hoc est recta, obliqua vero ei copulata quae a centro acta ad contactus [9] Ibid., Bk. I, Second Definitions. Cf. ed., Heiberg, Gr. 1, p. 66,
connectentem. Sint oppositae sectiones a ,b , copulatae autem sectiones cq, qd, coniungatur lines 16-26. Note that these definitions had been partly reflected in the
cd, seceturque bifariam in e, et coniungantur e, q. Aio eq diametrum esse quae orthia nun­
Arabic fragment translated by Gerard of Cremona (Chap. 1, ns. 26-27).
cupatur, obliquam vero ipsi quae a centro ipsi cd parallelus acta est. Sit namque, si fieri pos­
sit, diameter ef, capiaturque punctum/; coincidet igitur dq ipsi ef. Coincidat ad/ , et coniunga­ [10] Ibid., Bk. I, Prop. 17. Cf. ed., Heiberg, Gr 1, p. 68, lines 1-18.
tur cf, committetur itaque cd sectioni; committatur ad a , et per a ipsi cd parallelus agatur ab. [11] Ibid., Bk. II, Prop. 38. Cf. ed., Heiberg, Gr 2, pp. 254 (line 27)-
Quoniam igitur diameter est ef et cd bifariam secat, etiam parallelos ipsi bifariam dispescit. 256 (line 21).
Aequalis ergo ag ipsi g b . Et quoniam aequalis est ce ipsi e d , estque in triangulo [g]fd, aequalis [12] Ibid., Bk. II, Prop. 39. Cf. ed., Heiberg, Gr 2, pp. 256 (line 22)-
igitur ag ipsi gk. Et perinde etiam gk ipsi gb est aequalis, quod factu impossibile. Non ergo
ef erit diameter. 258 (line 8).
"[12] (11.39) Si oppositorum binae rectae lineae se contingant coincidentes, quae ex centro One cannot help but being struck by the random quality of these ex­
et coincidentia contingentium se acta bifariam secat contactus connectentem rectam lineam. tracts, if the purpose of them was to give a precis of the fundamental
Sint oppositae sectiones a, b, et ipsarum a, b binae rectae lineae commissae ce, ed, con- concepts of conic sections. In part, Valla might have been trying to give
nectaturque cd, diameterque ducatur ef. Aio aequalem esse cf ipsi fd. Si non est, secetur cd passages that would extend and clarify the material on conic sections
bifariam ing, et connectaturge; igitur diameter est. At est etiam ef·, centrum igitur est e. Ergo
found in the medieval tracts on conic sections, and particularly in Witelo’s
coincidentia copulatorum in centro est sectionum, quod absurdum dictu est; cf ipsi f d igitur
aequalis est.” Perspectiva. For one could consider passages [1] and [3] a clarification
I have altered the punctuation and capitalization radically and I have also italicized the of the remarks on the generation of conic sections at the end of Proposi­
letters marking magnitudes. The numbers in brackets at the beginning of each paragraph are tion 1.98 of the Perspectiva (see above, Chap. 3, n. 14). One could also
those I have employed in the text above to differentiate the various passages taken from say that Apollonius’ more general description of conic sections given by
Eutocius and Apollonius. The numbers in parentheses refer to the numbers of definitions and
propositions appearing in Heiberg’s text. I have also noted in parentheses corrections that
Eutocius in Passage [3] caused Valla to select for inclusion Passages
should be made, and sometimes I have indicated that Valla’s error resulted from the reading in [6]-[8]. Passage [6], namely 1.1 of the Conics, does indeed lie at the heart
his Greek manuscript (abbreviated here as Est.). Finally, observe that Valla follows the Greek of proving Prop. 1.98 of the Perspectiva, which held that the section is
in assuming that parallelus is an adjective of only two endings, for he throughout uses it as a not triangular but is bound by a curved line. Furthermore, Passage [7],
feminine form. The reader should read these extracts against the Greek texts of the Heiberg Prop. 1.3 of the Conics, proves that an axial section is a triangle for any
edition specified above. For paragraph [1], consult the figures on Gr2, pp. 172-73; for [3], Gr
cone and not just for the right cone of Proposition 1.90 of the Perspectiva
2,p. 175; for [5], Gr 2, p. 181; for [6], Gr l,p . 9;for[7],Gr l,p . 15;for[8],Gr l,p . 19;for [10],
Gr 1, p. 69; for [11], Gr 1, p. 257; and for [12], Gr 1, p. 259. Valla makes letter changes to adapt (see above, Chap. 3, n. 11). Similarly, Passage [8], Prop. 1.5 of the Comes,
the Greek letters to the Latin alphabet. These changes will be immediately obvious to the shows that the converse of Proposition 1.100 of the Perspectiva (which
reader who collates Valla’s translation with the Greek text. is true for the right cones understood by Witelo; see above, Chap. 3,
242 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 243

n. 23) is not true for subcontrary sections of oblique cones, which are by Valla from Gr MS A (owned by Valla) made available once more
also circles but are not parallel to the base of the cone. The inclusion (after William of Moerbeke and Jacobus Cremonensis) the various ref­
of this passage would also show the reader that Proposition 1.103 of the erences to conic sections included therein.5 As we shall see later, this
Perspectiva is not true when the cutting plane (perpendicular to the axial
triangle) is situated subcontrariwise, a condition excluded by Witelo’s 5 The whole section on proportional means and the division of a sphere in a given ratio taken
assumption of a right cone (see above, Chap. 3, note 26). As I have already from Eutocius occupies a part of Bk. XIII, chap, ii, of the De expetendis et fugiendis rebus
opus, sign, [u vi recto]-x iii recto. It was rendered from Gr MS A, which Valla owned. I
noted, Passages [2] and [9] gives the first and second set of definitions
published part of this section in my Archimedes in the Middle Ages, Vol. 3 (Philadelphia,
that Gerard had, for the most part, translated from the Arabic. Valla 1978), pp. 469-75. For Moerbeke’s translation of the whole section, see ibid., Vol. 2 (Phila­
might have given their translation from the Greek for the sake of their delphia, 1976), 36rY-42vK. The passages with references to the Conics of Apollonius have
comparison with Gerard’s translation from the Arabic. But given these been extracted above in Chap. 3, n. 6. The whole section in the translation of Jacobus
possible connections with the medieval material, it is difficult to see how Cremonensis may be found in MS Venice, Bibl. Naz. Marc. f.a. 327, 13v-38v (cf. the ed. of
1544, which includes the translation of Cremonensis as corrected by Regiomontanus,
Passages [4] and [5] related to any of the previous medieval material. separate page nos. for Eutocius, pp. 14-42). For a discussion of Cremonensis’ translation,
Nor is it clear how Passages [ 10]—[ 12] do much to clarify the Apollonian see my Archimedes in the Middle Ages, Vol. 3, pp. 321-42, 1342-56. The passages from
propositions assumed by Witelo from Alhazen in Propositions 1.129,1.131, Valla’s De expetendis that use conics and which Werner was later to summarize and rework
IX.39 and IX.40 of Witelo’s Perspectiva (see above, Chap. 3, ns. 30, 33, are the following: (Bk. XIII, Chap, ii, sign, u viii r-v; cf. diagram in Heiberg, Archimedis
38 and 41). Hence, in spite of the possible limited use of Valla’s extracts opera omnia, Vol. 3, p. 79, and note that Valla uses the following Latin letters for the Greek
letters of the text: h = Θ, f = Z ,g = H ,c = Γ). “ [1] Ut Menaechmus. Sint datae duae rectae
which I sketched above, they will still strike the reader as a rather ill lineae a, e. Oportet utique duas medias proportionales invenire. Fiant sintque b, c et ponatur
chosen group of passages. One would have thought that Valla could have positione recta linea dg terminata in d, et ad d ipsi c aequalis ponatur df, et ducatur ad rectos
more usefully presented, after his passages [I]-[3] and [6]-[8], Proposi­ angulos hf, ipsi b aequalis ponatur fh. Quoniam igitur tres rectae lineae proportionales a,
tions 1.11-1.13 of the Conics that make clear the basic parameters of b ,c , quod sub ipsis a , c aequale est [ei] quod ex b . Ergo sub data d (/ Gr MS A; a) et c , hoc est
the three conic sections than the propositions that he did present. af(! Gr MS A; df), aequale est ei quod exb, hoc est ei quae est exfh. Quoniam (/ Gr MS A; in)
comparatione (/ parabole) igitur h per a (! Gr MS A \d ) descripta, actae sunt paralleli hk, ak
The reader should also note other short passages regarding conics given (! Gr MS A; dk). Et quoniam datum quod est sub b , c, aequale enim est ei quod est sub d (! Gr
by Valla. In the first, he mentions the asymptotic relationship of the MS A; a), e, datum itaque etiam sub khf Quoniam hyperbole itaque ipsum h in [non] coincidenti-
hyperbola and its asymptote as well as that of the conchoid and its asymp­ bus ka, af (/ Gr MS A; kd, df) est, datum igitur ipsi (/ Gr MS A; ipsum) h, vel etiam/. Componetur
tote.4 The fact that he uses the expression dictum mirabile in this con­ utique hoc modo. Sint quidem datae rectae lineae d (/ Gr MS A; a), e. Positione autem ag
nection is perhaps a reflection of the comment at the end of Conclusion 10 (/ Gr MS A; dg) terminata pera {! Gr MS A; d), et describatur pera (/ Gr MS A ;d) comparatio
(/ parabole) cuius axis quidem ag (/ Gr MS A; dg), rectum autem formae latus d (/ Gr MS A;
of the Speculi almukefi compositio to the effect that the eleventh con­ a), at deductae in ag (/ Gr MS A; dg) in recto angulo possint quae ad d (! Gr MS A; a) sunt
clusion concerning two such lines seems prima facie mirabilis (see above, appositae areolae latitudines habentes sumptas sub ipsis ad a (/ Gr MS A; d) punctum.
Chap. 4, Text, Concl. 10, Appendix on a Good Steel, line 108). The Describatur etiam sitque ah (/ Gr MS A; dh), et rectus ak {! Gr MS A; dk), et in non coin-
example of the conchoidal line and its asymptote was perhaps suggested cidentibus ka ,a f(! Gr MS A; kd, df) describatur excessus (/ hyperbole), a quo quae ad ka, af
to him by the passage on Nicomedes’ solution of the proportional means (! Gr MS A; kd,d f) actae facient areolam aequalem ei quod est sub d (/ Gr MS A; a), e; secet
utique parabolen, idest comparationem, quae qualis sit in primo libro [geometriae ( = lib. X,
problem in Eutocius’ Commentary on the Sphere and the Cylinder of cap. iv; vide supra n. 6)] in diffitionibus exposuimus. Secetur in A, et perpendiculares ducantur
Archimedes, a passage he presented to the reader later in Book XIII of hk,hf. Quoniam igitur quod est ex fh aequale est ei quod est subi/ (/ Gr MS Α;α), a f {! Gr MS
his De expetendis et fugiendis rebus opus (sign, x i recto-verso). Needless to A ;df), est ut d (! Gr MS A; a) adfh et hf adfa (/ Gr MS A ;fd). Rursus quoniam quod est sub d
say, the collection of solutions of the means problem and the problem (! Gr MS A; a), e aequale est ei quod est sub hfa (/ Gr MS A; hfd), est ut d (! Gr MS A; a) ad
fh et fa {! Gr MS A \fd) ad e. Atqui ut d (/ Gr MS A; a) ad fh elfh ad fa (! Gr MS A ;fd), etiam
of dividing a sphere in a given ratio translated from Eutocius’ Commentary
fa (! Gr MS A ;/</) ad e. Ponatur ipsi quidem hf aequalis gb (/ Gr MS A; b), ipsi autem a f {!
Gr MS A ;df) aequalis c ; est igitur utc/ (/G r MS A; a) ad b etb ad c etc ad e. Ipsae i/(/G r MS A;
4 Ibid., Bk. X, chap, lix, sign, [n viii verso]: “ Geminus autem linearum inquit quaedam sunt a), b, c, e igitur sunt deinceps proportionales, quod oportebat invenire.
definitae figuramque complectuntur, ut circulus et eclipsis (/ ellipsis) linea et cissoides et “ [2] Aliter. (Cf. diagram in Heiberg, ibid., p. 83). Sint datae duae rectae lineae quae ad
aliae complures, quaedam indefinitae in infinitumque eductae, ut recta linea et rectanguli rectos angulos ab,bc, fiantque ipsarum mediaed b , be et perinde sint ut cb ad bd itabd ad be et
coni sectio et amblygonii, et conchoides. In infinitum rursus productarum aliae quidem be ad b a , ducanturque ad rectos angulos ipsae df, ef. Quoniam igitur est ut cb ad bd et db ad
figuram nullam comprehendunt, ut linea recta et conicae iam dictae sectiones, quae vero be, ergo quod sub cbe, hoc est quod sub data et be, aequale est ei quod est ex bd, hoc est ef.
coeunt efficiuntque figuram in infinitum post educuntur, non coeuntium aliae in uno sunt Quoniam igitur quod sub data et ipsa be aequale est ei quod ab ef, ipsum/igitur attingit para­
plano, aliae non sunt quae in uno sunt plano, earum aliae aequali semper distant intervallo, bolen quae circa axem ipsius ( / Gr MS A; ipsum) be. Rursus quoniam est ut ab ad be et be ad
aliae intervallum semper imminuunt ut hyperbole ad rectam lineam, et conchoides ad bd, quod ergo sub abd, hoc est sub data et bd, aequale est ei quod est ex eb, hoc est e f (/ Gr
rectam lineam, haeque quod dictum mirabile annuunt neque annuunt imminuitur intervallum MS A; <//); igitur/attingit parabolen eum quae circa axem bd. Attingit autem etiam aliam
neque unquam coincidunt.” datam circabf(! Gr MS A; be). Datum igitur/et perpendiculares//,/?; data igitur z/, i*. Com-
244 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 245

collection of Valla’s was of considerable interest to Johann Werner. A above, he gives both the pre-Apollonian and Apollonian names of the
further passage on conic sections given by Valla mentions, under the three conic sections.
rubric Coni sectio, the following pertinent sections of a cone: the axial Before returning to the medieval traditions and their influence in the
triangle, a circular section parallel to the base of the cone, and the non­ early sixteenth century, I should briefly note that in about 1508 or shortly
parallel sections of parabola, hyperbola, and ellipse.6 As in passage [1] thereafter, and thus not long after the publication of Valla’s extracts,
the German cleric Andreas Coner, who owned and corrected William of
ponetur autem ita. Sint datae duae rectae lineae ad rectos invicem angulos quae ab, bc, et Moerbeke’s autograph of his translations of Archimedean texts,7 saw and
producantur in infinitum ex/?, et describatur circa axem be parabole, itaque deductas be posse
used the Greek text of Apollonius, for in connection with the text of
liquet quae circa bc. Rursus describatur circa axem db parabole ut deductae possint ad ab.
Secabunt nimirum seinvicem parabole (/); secent in /e t ex/ perpendiculares ducanturfd ,fe . Eutocius’ Commentary on the Sphere and the Cylinder o f Archimedes
Quoniam igitur in parabole (/) deducta estfe , hoc est db, ergo quod sub cbe aequale est [ei] he adds marginal corrections of the proposition numbers of the Conics
quod ex bd. Est igitur ut cb ad bd et db ad be. Rursus quoniam in parabole (/) deducta est (given by Eutocius) to bring them in line with the numbers appearing
f d , hoc est e b , quod ergo sub dba aequale est ei quod est ex e b . Est igitur ut db ad be et be ad in some manuscript of the Conics he had seen.8 There is also some evi­
ba. Atqui ut db ad be ita cb ad bd\ et ut igitur cb ad bd et bd ad be, etiam eb ad ba, quod
dence that Francesco Maurolico, the celebrated mathematician of Mes­
oportebat invenire. Ait autem Eutocius ascalonites parabolen ab inventore Milesio Isidoro
mechanico nominatam.” (Note that in this last sentence Valla misrepresents the added re­ sina, who was later to render the Conics (see the next chapter), had
mark, which in the Greek merely says that a parabola is described by a diabeta invented by absorbed something of the text of the Conics by 1534.9
Isidore; cf. Heiberg, ibid., p. 84, lines 8-11.) I should remind the reader that, as in the After Regiomontanus’ skillful response to the medieval traditions of
previous notes, I have here and in the succeeding passage punctuated and capitalized as the conic sections, the most important contribution to those traditions was
meaning demanded. The reader will also observe that Valla followed Gr MS A, which he
effected by Johann Werner (1468-1522), the well-known mathematician
owned, in all of the mistaken substitutions of d fort/ and vice versa. I have given the correct
readings in parentheses (cf. William of Moerbeke’s translation where he too has made the cor­ of Nuremberg.10 His contribution appeared in a Libellus super vigintiduo-
rections, see my Archimedes in the Middle Ages, Vol. 2, 37vI-N). bus elementis conicis that was published as the first part of the volume
The last passage of Valla’s translation from Eutocius’ Commentary on the Sphere and whose title is: In hoc opere haec continentur: Libellus Ioannis Verneri
the Cylinder of Archimedes in which conics were used and Werner was to rework occurs in the Nurembergen. super vigintiduobus elementis conicis. Eiusdem com­
De expetendis, Bk. XIII, Chap. 2, sign, x ii r (I have given only the first part of Dionysodorus’
mentarius seu paraphrastica enarratio in undecim modos conficiendi
solution of the problem of cutting a sphere in a given ratio; consult the diagram in Heiberg,
ibid., p. 155, and note that Valla uses c ,/a n d g to replace the Greek letters Γ, Z and H): ‘‘[3] eius problematis quod cubi duplicatio dicitur. Eiusdem commentatio in
Datam sphaeram plano secare ut segmenta ipsius ad se invicem rationem habeant datam. Sit Dionysidori problema, quo data sphaera plano sub data secatur ratione,
data sphaera, cuius diameter ab, data autem ratio quam habet cd ad de. Convenit nempe alius modus idem problema conficiendi ab eodem Ioanne Vernero novis­
secare sphaeram plano recto ad ab ut segmentum cuius vertex a ad segmentum cuius vertex sime compertus demonstratusque. Eiusdem Ioannis, de motu octavae
b rationem habeat quam habet cd ad de. Producatur ba in/ , ponaturque ipsius ab dimidia
sphaerae, tractatus duo. Eiusdem summaria enarratio theoricae motus
af, et quam habet rationem ce ad ed habeat fa ad ag, sitque ag ad rectos angulos ipsi ab, et
ipsorum/a, ag media proportionalis sumatur ah. Maior igitur ah quam ag\ et circa axemJb octavae sphaerae (Nuremberg, 1522).
quod (/ per) / descripta sit parabole, et perinde productas posse ad ag. Ibit igitur per h, Despite the facts (1) that Werner was competent in Greek, (2) that
quoniam utique quod est sub/dg aequale est ei quod est ex ah. Describatur itaque, sitque ut Regiomontanus’ Greek manuscript of the Conics of Apollonius was at
jhk, et per b transducatur ad ah, bk, secetque parabolen in A:, et perg circa non coincidentes
Jbk describatur hyperbole; secet nimirum parabolen inter//, k. Secet in/, et ex / in ab perpen­
7 For my earlier treatment of Coner, see Archimedes m the Middle Ages, Vol. 2, passim,
dicularis ducatur Im, et per g, l ipsi ab paralleli agantur gn, Ix. Quoniam igitur hyperbole est
and Vol. 3, pp. 525-38.
g l, non coincidentes abk, et paralleli ipsis agn ipsae mix, aequale est quod sub agn ei quod est
* Ibid., Vol. 2, commentary to 40vC -D, L, R, S; 41vD; 42rl, O, and P. Since Coner
sub mix per octavum theorema, idest praeceptum, secundi libri Apollonii conicorum ele­
probably worked on the Moerbeke translation when he was in Venice after his purchase of
mentorum. . . .”
the manuscript in 1508, he possibly had access to Bessarion’s manuscript of the Conics then
h De expetendis et fugiendis rebus opus, Bk. X, chap. Ixxxi, sign, o i verso: “ Coni sectio.
in the San Marco collection (see note 1) just as he had access to Bessarion's copy of the Greek
Conus desectus per verticem triangulum facit sectionem, in paralleli ritum basi desectus
text of Archimedes (MS E, see Archimedes in the Middle Ages, Vol. 3, p. 537).
[circulum tacit sectionem], conus autem non in paralleli modum sectus sed quod lineae pars
9 Ibid., Vol. 3, pp. 795, 947 (lines 11-12), 1023 (Prop. 1, line 8; Prop. II, lines 4-5 ; Prop.
coni sectio nominatur; coni autem sectionum una quidem rectangulus appellatur, altera autem
Ill, line 8), 1024 (Prop. IV, line 6).
amblygonius, tertia oxygonius. Oxygonius dicta ea ipsa connexa figuram facit arealem;
10 See the article of M. Folkerts on Werner in the Dictionary o f Scientific Biography, Vol.
vocatur autem a quibusdam ellipsis quasi circulus compressus; orthogonii vocatur para­
14 (New York, 1976), pp. 272-77, my Archimedes in the Middle Ages, Vol. 3, pp. 1164,
bole, amblygonii hyperbolae.” Cf. the even briefer passage earlier in the same book, chap, iv,
1165n, 1169n, and M. Cantor, Vorlesungen iiber Geschichte der Mathematik, Vol. 2, 2nd ed.
sign, [n v verso]: “ Tres autem canorum (/ conorum) sectiones sunt parabolae et hyperbolae
(Leipzig, 1913), pp. 452-59. For Werner's connection with Greek studies, see G. Bauch, Die
et eclipsis (/ ellipsis), idest comperatio (/), excessus, et defectus dictae, de quibus dicetur mox
Anfange des Humanismus in Ingolstadt in Historische Bibliothek, Vol. 13 (Munich/Leipzig,
suo loco.”
1901), p. 94.
246 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 247

Nuremberg, and (3) that Werner was a friend of Bernhard Walther, who silently quite a number of obvious typographical errors that appear in the
apparently had executed a Latin translation of the Conics that was also edition of 1522.
at Nuremberg, Werner’s tract on Conics shows no direct dependence on Before examining the text, proposition by proposition, I should make
the Conics but is, in fact, a work that depends primarily on the medieval a few general observations. The title was no doubt suggested to Werner
traditions of conic sections and on the various passages on conics that by the few references to Apollonius’ work, in the Valla passages and in
were transmitted in Giorgio Valla’s De expetendis et fugiendis rebus the medieval treatises, as De conicis elementis. And though he did not
opus. 11 However, as we shall see, Werner developed these sources in a directly employ Apollonius’ work, he certainly knew it to be the standard
decidedly novel way. work on conical elements and so in all likelihood decided to use a similar
Now let us look at Werner’s work in some detail. I present its text title for his own work. We should also initially observe that, in spite of
below, with an accompanying English translation. My purpose is to give Werner’s significant dependence on the medieval texts he had read, he
a text that reflects clearly the author’s intent. I have thus made some recast the medieval vocabulary in terms he learned from reading Valla’s
changes. As usual, I have changed the punctuation and capitalization. work and the translation made from the Greek of Euclid's Elements by
Though the 1522 edition was not completely consistent, I have followed Bartolommeo Zamberti (printed in Venice in 1505 and everywhere cited
its prevailing practices of omitting periods after the rubrics of the defini­ by Werner in his volume). Thus the medieval pyramis rotunda becomes
tions, divisions, and postulates and including periods after the rubrics of conus (which ordinarily in the medieval geometrical vocabulary simply
propositions, corollaries, and lemmas. In the English translation, I have meant an apex), pyramis declivis or pyramis obliqua becomes conus scale­
used some modern notation in order to epitomize the verbal exposition of nus, pyramidis ypothenusa becomes coni latus, proportio becomes ratio,
Werner, as indeed I have done throughout this volume. Thus most of the linea ordinis is transformed into linea structim acta or linea ordinatim
time I have presented proportions in the form of a l b = c / d. Hence “ ratio acta, both of which expressions for the ordinate he perhaps devised from
ipsius/g ad g eest sicut ed ad dc" becomes simply "fg/ge = ed /dc." Also, similar expressions used by Valla. Werner almost always used Valla’s
I have used an exponent to indicate a square. Thus “ quadratus ipsius bd” expression for the axial triangle of a cone: ab axe coni triangulus; and
becomes “bd 2.” The verbal statement of a rectangle measured by its sides he exclusively used only the Greek names for the conic sections: parabole or
has often been presented as a product. Hence “ quod rectangulum sub ghf parabola, hyperbole and ellipsis. However, asymptotes are always called by
factum sit aequale ei quod fit su b /e , ed rectangulo” becomes “gh-hf Werner non coincidentes (as they were by Valla, and, for the most part, by
= f e e d . " Finally, Werner’s extended phrase for adding of one straight Jacobus Cremonensis in his translation of the Archimedean texts) instead
line to another in a straight line is reduced to a simple sum. So "de, ec of Moerbeke’s (and thus Witelo’s) asymptote. Gone also is the medieval
pariter iunctis in directum” becomes "de + ec." Other examples will be term sagitta for the conic’s axis or a segment thereof. Only axis remains
evident to the reader as he peruses the translation in conjunction with the in Werner’s tract. Gone too is Gerard of Cremona’s expression linea
text. It is also clear from these various examples I have given that I have munani for the conic line in the surface of the cone. In its stead Werner
italicized the letters indicating magnitudes in both my text and translation, puts linea inflexa, which he took from Valla.
though they were unitalicized in the original text of 1522, where they were While Werner’s vocabulary was greatly influenced by recent transla­
often set off by periods. Note that the author is not consistent in pre­ tions from the Greek, it will be clear in our account below that in his treat­
senting the order of the letters representing magnitudes. Thus we find a ment of the parabola (Elements I-X V I) and of the hyperbola (Elements
line being designated indifferently as me or em in the same argument XVII-XXII) he depended markedly on two medieval texts.12In the case of
without any mathematical reason for the transposing of the letters. In the parabola, it should be noted that all of his propositions concerned a
general, I have followed the author in his inconsistency. I have corrected parabola formed from a right-angled right cone. Such was also the case in
the various medieval texts that concerned parabolas. But the medieval text
11 As will be evident from my analysis below, Werner’s tract is introductory to the tracts on that most influenced Werner’s treatment of the parabola was the Speculi
the duplication of the cube and the division of the sphere in a given ratio that follow in the almukefi compositio, which Werner seems to have read in the version pre-
volume of 1522. In the Introduction to the volume (dated 11 January, 1522) he tells us (sig. a ii
recto): “ Eisque [de cubi duplicatione] non immerito praemisi conica elementa, ut his discussa
densae obscuritatis nebula longe evidentiore patescerent intelligentia.” The composition of 12 There is no treatment of the ellipse in Werner's work. This is not surprising since Werner
Werner’s tract ought to be dated between 25 October, 1505 (on which date appeared the has no need of propositions concerning the ellipse in his succeeding works. The passage in
translation of the Elements of Euclid from the Greek by Bartolommeo Zamberti which was Eutocius’ treatment of the problem of dividing a sphere in a given ratio which made use of an
cited everywhere by Werner in verbatim fashion) and 11 January, 1522 (the above-noted ellipse (see Heiberg, Archimedis opera omnia, Vol. 3, p. 168, lines 6-13) was not translated
date of the introduction to Werner’s volume). by Valla and hence was not included by Werner in his treatment of that problem.
248 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 249

pared by Regiomontanus. Regarding the hyperbola, it was the De duabus plane were an oblique cone. The section in the surface of the cone is the
lineis translated by John of Palermo which exercised the preponderant linea inflexa. (Werner here follows Valla’s translation of καμπύλη γραμμή)
influence on Werner. But it is also clear that Valla’s work was not without which, as Werner says, the mathematicians call a “ conic section.’’ The
its collateral and even substantial influence on Werner’s tract. The suc­ reader will recall that Gerard of Cremona had used the expression linea
ceeding analysis will, I believe, not only support these general remarks on munani for this section (Chap. 1 above, n. 5, passage [4]), while John of
the sources influencing Werner, but will as well reveal Werner’s originality. Palermo, in translating the De duabus lineis, used obliqua linea (Chap. 2
Definitions, Divisions, and Postulates. The first definition of the cone above, Text A, Proemium). Werner’s Division 3, which distinguishes the
given by Werner is the Euclidian definition that has the cone arise in the three conic sections, is not unlike the remarks rendered by Valla from
rotation of a right triangle. This was the definition so often seen in the Eutocius’ Commentary on the Conics o f Apollonius (see above, n. 3, pas­
medieval tracts. The alternate definition given by Werner is the Apollonian sage [3]). Compare also Gerard of Cremona’s fragment from the Arabic
definition that describes the double cone. It was taken by Werner directly version of the Conics (Chap. 1 above, n. 18), Witelo’s Proposition 1.98
from Valla’s work (see above, note 3, passage [2(1)]). Once given, this (Chap. 3 above, at the end of n. 14), and the third definition of the Speculi
alternate definition plays no further role in Werner’s work.13 Werner’s almukefi compositio (Chap. 4 above, Text).
Definitions 2 -4 relate to the initial definition of the cone and need no spe­ Werner’s sixth and last definition deserves more extensive commentary
cial comment. Of the divisions that follow the first four definitions, the than the others. It commences with the axis of the conic section which it
reader should observe that Division 2 was taken by Werner from Valla (see designates as the line that bisects the straight lines drawn from the sec­
note 3, passage [2(3)]). It is perhaps surprising that Werner’s alternate tion at right angles to the axis. It notes as an alternate term “ diameter,” a
word (after scalenus) is inclinatus rather than Regiomontanus’ obliquus term used both in Gerard of Cremona’s fragment (see Chap. 1 above, note
(see above, Chap. 5, Text C, Concl. 1, line 9;Concl. 3, line 7). The oblique 5, passage [4]) and in Valla’s translation of the First Definitions of Apol­
cone plays no further role in Werner’s tract. The first postulate of Werner’s lonius’ Conics (see above, n. 3, passage [2(4)]). Werner, however, makes
work is essentially Proposition 1.1 of the Conics of Apollonius, and is more no distinction between an axis and a diameter. In the medieval works it was
or less like the version of it given by Valla (see note 3, passage [6]). It is rather exclusively called the axis or sagitta of the conic section. The ver­
also equivalent to remarks made by Witelo (see Chap. 3 above, n. 8). The tex is then defined in a manner similar to its description in the various pas­
second postulate, equivalent to Proposition 1.2 of the Conics, was not given sages of the previous works where the axis or diameter is treated. I have
by Valla, but appears in a somewhat different form in Proposition 1 of the already noted in my discussion of Werner’s terminology that the medieval
medieval De duabus lineis (see Chap. 2, Text A) and in Conclusion 2 of the linea ordinis became for him the straight line from the section drawn at
Speculi almukefi compositio (see Chap. 4 above, Text). The third postulate right angles to the axis structim or ordinatim. He adds that mathematicians
and its derivative fifth definition is equivalent to Proposition 1.3 of the also call them lines drawn secundum ordinem, thereby, I suppose, intend­
Conics of Apollonius and has its principal source in Valla’s work (see ing the medieval authors. See particularly Proposition 1 of Alhazen’s De
above, note 3, passage [7]), but consult also Gerard of Cremona’s fragment speculis comburentibus (Chap. 1 above, n. 33), where secundum ordinem
from the Arabic version of the Conics (Chap. 1 above, note 18, passage [1]), is used. For linea ordinis, see Gerard of Cremona’s fragment (Chap. 1
Conclusion 1 of the Speculi almukefi compositio (Chap. 4 above, Text), above, n. 5, passage [4]) and the Speculi almukefi compositio (Chap. 4
and Proposition 1.90 of Witelo’s Perspectiva (Chap. 3 above, n. 11). above, Text, 3rd def.). In adopting for his own usage the expression linea
Werner’s fourth postulate, equivalent to Proposition 1.4 of the Conics of structim acta, Werner was perhaps following Valla’s translation (see
Apollonius, was not rendered by Valla, but was Proposition 2 of the De above, n. 3, passage [2(4)]). Of all of these parts of Definition 6, it is Wer­
duabus lineis (Chap. 2 above, Text A), Conclusion 3 of the Speculi almukefi ner’s definition of latus rectum that is the most interesting. He gives al­
compositio (Chap. 4 above, Text), and Proposition 1.100 of Witelo’s ternate definitions. The first declares that, of those lines drawn ordinate-
Perspectiva (Chap. 3 above, n. 23). Werner’s fifth postulate, which postu­ wise [in a parabola], that one which is equal to its axial intercept is called
lates a conic section as that section produced by a plane that does not pass the latus rectum. The alternate definition notes that the latus rectum is
through the apex of the cone and is not parallel to the base, is apparently sometimes called “ the straight line according to which the lines drawn
in reference only to a right cone, since no mention is made of a subcontrary ordinate-wise are squared.” Let me first comment on the alternate defini­
circular section which ought to have been excepted if the cone cut by the tion. It reflects the abbreviatory expression used in Greek as early as
Archimedes: παρ' αν (or ήν) δννάνται, αί άπο τάς τομάς (or τής τομής).14
13 In view of the fact that the Apollonian definition of a cone played no further role in
Werner’s Libellus it is surprising that Cantor (op. cit. [in note 10 above], p. 456) should take
special notice of this definition. However, the plain fact is that no account of Werner’s work 14 See Τ. L. Heath, The Works of Archimede s (Cambridge, 1897; Dover reprint. New York,
has been produced that gives careful attention to Werner's sources. 1950), p. clxviii; C. Mugler, Dictionnaire historique de la terminologie geometrique des
250 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 251

This truncated expression stands for a statement like the following: “ a to Passage 1 (Chap. 5, Text D) and its accompanying diagram (Fig. 5.11). It
straight line to which as a base are applied rectangles having as their widths perhaps would have struck him immediately that in the half-parabola repre­
the axial intercepts of the lines drawn ordinate-wise from the parabola and sented in that figure the total axis ex is indeed equal to the ordinate zqx, and
being equal in area to the squares of the lines drawn ordinate-wise.” The each is equal to the given latus rectum of 100 parts. At any rate, the most
Latin phrase used by Werner he perhaps adopted from Valla’s rendering interesting consequence of Werner’s definition of the latus rectum as the
of the Greek abbreviatory phrase in the latter’s version of passages from ordinate equal to its intercept is that it allows him to turn completely
Eutocius’ Commentary on the Sphere and the Cylinder o f Archimedes (see away from the medieval concentration on the focus of a parabolic sec­
above, n. 5, passages [l]-[3 ], and also Werner’s rewriting of the passages tion. There is, in fact, no mention in this work by Werner of the focus (i.e.,
of note 5 in ns. 16, 18-19 below), or he possibly translated from the Greek of the medieval “ point of combustion” or “ point of concourse of rays” ).
text of the Conics itself (see note 14). We shall see later that Werner was well aware of the medieval interest
Returning to Werner’s initial definition of the latus rectum as the ordinate in the focal point when we examine an appendix to his treatment of the
equal to its intercept, we ought first to note that he seems to imply that problem of proportional means (see below, note 17). Finally, I should alert
mathematicians customarily so define the latus rectum. But I know of no the reader to the fact that when discussing the definitions of latus rectum
mathematician available to Werner that gives such a definition. Perhaps in this sixth definition, Werner is limiting himself to the latus rectum of a
all that he meant to say was that this ordinate is in fact equal to the line that parabola, although he does not so specify it.
mathematicians call the latus rectum but locate otherwise (as for example Elements I-IV . These propositions are auxiliary to Elements V -V I. In
as the chord through the focus, as some medieval authors locate it). At any the first one it is shown that, if we have an isosceles right triangle abc (see
rate, I would suppose that Werner read Conclusion 4 of the Speculi al- Fig. 6.2) from whose base be one third part dc is taken, then the line de
mukefi compositio and learned that the latus rectum was the double or­ parallel to ac is the mean proportional between bd and dc. In the second
dinate or chord that passed through the midpoint of the axial segment be­ element (Fig. 6.3) it is shown that, if dc is less than or more than one-third
tween the vertex of the parabola and the intersection of the section’s axis be, and if de is again drawn parallel to ac and e f is marked off so that ef
with the axis of the right-angled right cone from which the section is formed. = 2 ae, and if line fhg is drawn parallel to be, then fh I eh = eh / gh. If dc is
It was shown in this fourth conclusion that the latus rectum is twice that less than one third d c , thenfg is parallel to be above line be (Fig. 6.3); if it is
axial segment, and hence the axial intercept of the chord through the mid­ more than one third dc, then fg is parallel to be below be. In the third ele­
point equal to the latus rectum is one quarter of the latus rectum. Hence, ment (Fig. 6.4) it is shown that, if point d is located so that de, parallel to
when the axial intercept of an ordinate is one-quarter of the latus rectum, b e , is the mean proportional between ad and d c , thenf h , parallel to a c , is so
the ordinate is one half of the latus rectum. He would have equally learned intersected by the first parallel de at g that fg-gh = de-eg. Finally, in
from the proof of Conclusion 4 that the axial intercepts of ordinates are to Element IV (Fig. 6.5), if, as before, d is so located that de is the mean pro­
each other as the squares of the ordinates. Hence, if we take an ordinate portional between ad and d c , and if a line gh is drawn parallel to ac but be­
equal to the latus rectum, then its axial intercept must also equal the latus low ac and thus intersects the extensions of be, ed, and ba at h ,f , and g
rectum. Thus would follow Werner’s definition. Or Werner could have respectively, then gh-hf = fe ed.
more directly reached that definition by the second part of Conclusion 4 Elements V -VI. Werner proves the principal property of the parabola in
of the Speculi almukefi compositio, namely that an ordinate is the mean these two propositions, namely that the square of an ordinate is equal to the
proportional between its axial intercept and the latus rectum. In such case, rectangle whose sides are equal to the ordinate’s axial intercept and the
it is obvious that, if the ordinate is taken as equal to the latus rectum, then latus rectum. In Element V (Fig. 6.6) we have a right-angled right cone
the axial intercept must also equal the latus rectum. In modern terms, ify 2 abed whose axial triangle is abc and whose base is circle adc. Parabola
= px, where y is the ordinate and x is the axial intercept and p is the latus fed is fashioned by cutting the cone by a plane perpendicular to the axial
rectum, and ify - p, then * must also equal p. It could also be that if triangle. Axis fg of the parabola is taken as equal to the latus rectum and dg
Werner examined Regiomontanus’ version of the Speculi almukefi com­ in the base of the cone is the ordinate equal to fg , following the sixth defini­
positio (which I am fairly certain he did), then he would have seen the Table tion. In the case of Element V, line eh is any ordinate of the parabola that
falls on fg. If a plane parallel to the base of the cone is drawn through eh,
Grecs (Paris. 1958), p. 323; Clagett, Archimedes in the Middle A ges, Vol. 2, p. 497, 37vR “ ita
ut1 . . . B G " . It might be argued that Werner has taken the alternate definition of latus rec­
then circle iek is formed. Since dg is perpendicular to diameter ac of semi­
tum from the Greek text of Proposition 1.11 of the Conics of Apollonius (Gr 1, p. 42, lines circle adc and d is on the circumference of that semicircle, then dg is a mean
1-3). If this is true, it would be the only case of the direct use of the text of the Conics by proportional between ag and gc. Hence fg (equal to dg) is a mean propor­
Werner (as distinguished from his indirect use of that text via the De expetendis of Valla). tional between ag and gc .H ence by the third conic element gf-fh = ik · hk.
252 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 253

Now since eh is perpendicular to diameter ik of semicircle iek and e is in the extended to/ so that a f = ae. With dc drawn and extended, we then draw
circumference of the semicircle, thus eh2 = ih · hk . Therefore eh2 = gf-fh, fg parallel to ac. Thus angle dfg is a right angle, fg is equal to fd, and we
which was proposed for proof, since eh is any ordinate falling perpendicu­ have completed by this construction a right isosceles triangle dfg of the
larly on latus rectum g f andfli is the ordinate’s axial intercept. In Element kind constructed in Element I. Indeed it is evident that such a triangle in
VI (Fig. 6.7), the same thing is proved for any ordinate dg that falls on axis which the section’s axis ac is equal to the latus rectum of the parabola and
fg beyond the terminus h of the latus rectum./?*, that is to say, it is proved cuts the two sides of the triangle into segments that are related as 2 to 1 is
that dg2 = fg -fh. The proof is similar to that of Element V, except that in it the axial triangle that plays a crucial role in all of Werner’s consideration of
the fourth conic element is the one appealed to. While these proofs are the parabola. Now if a perpendicular fh is dropped to dg, it will bisect tri­
completely distinct from the proof in Conclusion 4 of the Speculi almukefi angle dfg·, and if half of that triangle (say the half designated as dhf ) is ro­
compositio, it is evident that the same property is being demonstrated and tated about fh the required cone is formed. Now what ought to be shown to
that crucial to both treatments is the property of the circle to the effect that complete this problem is that all of the points of the parabola do indeed fall
a perpendicular from the circumference to the diameter is the mean propor­ on the surface of the cone that has been constructed. This can be demon­
tional between the segments of the diameter produced by the perpendicular. strated in two parts (of which Werner gives only the first and that presented
Element VII. This is the axial property of the parabola alluded to in my in a rather loose fashion): (1) that ordinate be is a perpendicular from the
discussion of the sixth definition above, namely that the ratio of the squares diameter to the circumference of the circle which is the base of the cone so
of the ordinates is as the ratio of their axial intercepts. Werner proves it that point b lies on the circumference of the base circle of the cone, and (2)
easily by reference to the preceding elements (see Fig. 6.8). Where/α is the that the terminus of any ordinate of the parabola (and thus all points of the
latus rectum, we have to prove that be2 / dg2 = ea / eg. This is evident, given parabola) will lie on the circumference of a circle that is parallel to
since, by Element V, be2 = fa-ea, and, by Element VI, dg2 = ga-fa. the base of the cone and is in the surface of the cone (so that all the points of
Note that the axial property was proved before the proof of the basic the parabola will lie on the surface of the cone). To prove the first part we
property of the parabola in terms of its latus rectum in Conclusion 4 of the show that ordinate be lies in the same plane as the base circle of the cone
Speculi almukefi compositio (see Chap. 4 above, Text Concl. 4, lines and is perpendicular to the diameter dg. Then it is demonstrated that, since
18-40), while in Werner’s tract the axial property is derived from the basic ac is the mean proportional between dc and eg (by Element I15), cb (equal
property. I should also remind the reader that this axial property was given to ac) will also be the mean proportional between dc and eg and thus the
by Apollonius in Proposition 1.20 of the Conics, a fact not known by the circumference of the base circle of the cone will pass through point b and so
author of the medieval tract and not mentioned by Werner. point b of the parabola lies on the surface of the cone. While Werner does
Element VIII. Here Werner easily finds the latus rectum when a parab­ not bother to prove the second part of the proof, it can easily be done by us­
ola constructed from a right-angled right cone has been given (Fig. 6.9). He ing the basic property of the parabola proved by Werner in Elements V
simply constructs at the vertex and with the axis an angle of 45 degrees. and VI and using the property of the circle to the effect that a perpendicular
From the point where the angle side intersects the parabola he draws an from the diameter of a circle to its circumference is the mean proportional
ordinate to the axis. Thus isosceles triangle acd has been formed, with between the parts of the diameter cut off by the perpendicular.
cd = ad. Hence by the sixth definition cd or ad is the required latus rectum. Element X. Werner now proves with ease, on the basis of the seventh
This is to be compared with the medieval technique of finding the line that is element, that, if we are given two ordinates of a parabola arising from a
double the line to the axis (i.e. double the axial intercept between the right-angled right cone, and if we are also given the segment of the axis be­
vertex of the parabola and the point where the section’s axis intersects the tween the two ordinates, then the whole axis from the vertex to the second
cone’s axis). Or if one is given the parabola, its vertex and the focus, the ordinate will be given. Werner’s source for this proposition was, I believe,
medieval mathematician simply extends the ordinate through the focus to Passage [5] of Regiomontanus’ notes to the Speculi almukefi compositio
produce the latus rectum as a chord. (see Chapter 5 above, Text D), though Regiomontanus’ consideration of
Element IX. Werner here constructs the right-angled right cone from this problem was expressed in terms of an annular paraboloidal mirror
which a given parabola with a given axis can be formed. The construction and its focus. Needless to say, Werner, as usual, ignores the focus. Nor
is simple. It rests on constructing what will be the axial triangle dfg of the does he propose to find the latus rectum, which, however, is immediately
cone to be constructed (see Fig. 6.10) and then rotating half of this triangle evident in the problem. For, after the whole axial segment of the second
around the line that bisects this triangle to produce the required cone. The
15 In fact, Werner does not cite Element I but proves once more that ac is the mean pro­
triangle is constructed as follows. A plane perpendicular to the plane of the portional between dc and eg. This is strange since he has here constructed triangle dfg to
parabola is passed through the axis of the parabola ac and in that plane line fit the conditions of Element I, and so all he needed to have done was simply to cite the earlier
ad (equal to ac) is led at right angles to ac. This line ad is bisected at e and element.
254 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 255

ordinate has been determined, and with the second ordinate given, the pendicular determines the length of an ordinate of the parabola and thus
latus rectum will be given by the equation of the parabola determined in lies on the parabolic section. It will be obvious to the reader that this is
Elements V and VI. precisely what has been done by Werner as he determines successively
Element XI. Werner here instructs us how to describe a parabola in a the lengths bf, bi, etc. Hence my conclusion that Werner in all likelihood
given plane when the latus rectum of the parabola is given. In stating the depended on the Speculi almukefi compositio as presented and commented
problem he uses the alternate definition of the latus rectum given in the on by Regiomontanus. It is of interest to note that Werner concludes this
sixth definition, namely the straight line according to which the lines drawn element by remarking that “by the amount that the [equal] parts of the
ordinate-wise are square. I have already remarked in my comments to the latus rectum kn are narrower [i.e. smaller] by such an amount will the parab­
sixth definition that this is a truncated expression that relates to the original ola be more truly drawn.” And he adds that “any straight line joining
Greek expression of the parameter of the parabola (παρ’ ην δννάνται etc.). any two immediately neighboring points [i.e. neighboring termini of the
Werner’s procedure for describing the parabola was almost certainly taken ordinates] differs in a small degree and quite insensibly from the curved bit
by him from Passage [1] of Regiomontanus’ notes to the Speculi almukefi of the parabola which is bounded by the same points.”
compositio (see Chap. 5 above, Text D). Still it includes a certain refine­ Elements XII-XV. These elements are all auxiliary to Element XVI. In
ment beyond Regiomontanus’ procedure. First Werner lays out the latus Element XII (see Fig. 6.13), Werner demonstrates that if from some point
rectum ab on a straight line (see Fig. 6.12). Then he divides line ab into any d outside of the circle two lines are drawn—dc to the center of the circle
number of equal parts (e.g., ac, cd, and db). Let be be equal to one of the and bd tangent to the circle— and if a perpendicular be is dropped to cd,
parts into which ab is divided and let it be added to the end of ab, produc­ then the lines cd, ac, and ce are continually proportional, and he further
ing line ae. At point b an indefinitely long perpendicular is erected. Fol­ shows in the corollary that these lines are continually proportional accord­
lowing this, ae is bisected at point d and a semicircle afe is drawn with d as ing to the ratio of ad to a e . In Element XIII (see Fig. 6.14), he shows by the
the center and ad as the radius. The semicircle will cut the indefinitely long corollary to Element XII that, if two lines are drawn—d f to some point in
perpendicular in /. To straight line ae is added eg, also equal to one of the the circumference and db tangent to the circle— and if as before dc is drawn
parts into which ab is divided. Then ag is bisected at h and a new semicircle to the center and a perpendicular eb is dropped to line d c , and if lines fa , f e ,
is drawn with h as the center and ah as the radius. It cuts the indefinite line and fc are drawn, then df /fe = ad I ae. And in the corollary to Element
in point i. In the same way we add, one by one, equal parts to line ag and its XIII it is demonstrated that line a f bisects angle efd. Then in Element XIV,
successors, construct successive semicircles and so determine further Werner demonstrates how to draw a plane tangent to a cone through a side
points of intersection on the indefinite perpendicular. Then we take line of the cone. And he indicates in a corollary that if a plane is tangent to a
kn equal to ab (i.e. to the given latus rectum), dividing it into the same num­ cone through a side of the cone, it is also tangent to every curved line which
ber of equal parts into which ab was divided. On the termini of these parts is described in the surface of the cone and cuts the side of the cone, and the
(l, m, n) we erect indefinitely long perpendiculars above and below mn. common section of that side and such a curved line will be the point of
Following this, we cut off, above and below mn, on these indefinitely long tangency. Finally, in Element XV Werner shows how to draw a tangent to a
perpendiculars lines that are equal to bf, bi, etc. Through the termini of given point on a conic.
these lines cut off above and below mn (i.e. through points o andp , q and r , Element XVI. With these preliminary elements presented, Werner then
etc.) and through the vertex k a parabola may be drawn. Werner justifies presents a proposition that is equivalent to Conclusion 7 of the Speculi
his procedure by showing that each of the ordinates is the mean propor­ almukefi compositio (and thus to Proposition 1.35 of the Conics of Apol­
tional between its axial intercept and the latus rectum. It is evident that lonius), namely that if a tangent ec (see Fig. 6.17) is drawn to parabola
Werner has followed the third method mentioned by Regiomontanus in abc, and if axis ad of the parabola is extended to meet the tangent ec at
Passage [1] for determining the ordinates, namely the one that uses Conclu­ point e , and if ordinate cd is drawn, then the axial intercept ad will be equal
sion 5 of the Speculi almukefi compositio. This conclusion had shown that to the extension ae of the axis. With the parabola given, the right-angled
if (1) we are given the axis, the vertex and the latus rectum of a parabola, right cone is given by the ninth element; its apex is h and its base is fcg.
and (2) we take any axial segment beginning from the vertex, and (3) we add Further its axial triangle is fgh in a plane perpendicular to the plane of the
to the axial segment the latus rectum in a continuous straight line, adding parabola. And so fcg will be the circumference of a semicircle. Then c and
it along the axis, and (4) we bisect this conjoined line, erecting on the point h are connected and thus ch is a side of the cone. Werner first takes arc eg
of bisection a circle with half of the bisected line as a radius, then (5) that to be half of semicircle aeg, and so the plane cdh will be at right angles to
circle will cut an indefinitely long perpendicular erected at the terminus of axial triangle/g/*. Hence the ordinate cd falls on d, the center of circle fe g h ,
the axial segment, and the point of intersection of the circle and the per­ and so the plane of triangle cdh is at right angles both to the plane of triangle
256 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 257

fgh and that of circlefcg. Then a plane ceh tangent to the cone on line ch is with Proposition 4 of the earlier work, he will see that they are very closely
drawn by the fourteenth conic element. The common section of the plane alike, even through the corollary, and that in each case it is the axial prop­
of circle/cg and plane ceh is line ci, which is at right angles to the plane of erty of the hyperbola that is established. (As I noted in my discussion of
cdh. Now cdg is a right angle by hypothesis. Hence ci and dg are parallel. Proposition 4 of the De duabus lineis in Chapter Two above, this proposi­
Hence plane ceh is at right angles to plane cdh. But the plane of triangle tion is equivalent to the second part of Proposition 1.21 of the Conics of
fgh is also at right angles to the plane of cdh. Therefore line eh is the com­ Apollonius, a fact which neither the author of the De duabus lineis nor
mon section of the plane of triangle fgh and the plane ceh. Hence ci and eh Werner mentioned.) Similarly, a comparison of Werner’s Element XX with
are parallel. Since ci and dg are also parallel, hence eh and dg are parallel. Proposition 5 of the De duabus lineis demonstrates once more Werner’s
Now point e is common to plane ceh and plane/g/z; hence e lies on line eh. dependence on the earlier work. Elements XVIII and XIX are auxiliary
Since eh is parallel to dg, therefore angle ahe equals angle agd, and hence propositions, and almost certainly Element XVIII was suggested to Wer­
angle ahe is also equal to angle ahd. Finally the angles at a are right angles ner by the lemma at the end of Version B (see Chapter Two above). Since
and side ah is common to triangles aeh and adh. Hence the two triangles that lemma was not given in Version A of the De duabus lineis, it would be
are equilateral. And because the sides subtended by equal angles are equal, fair to conclude, as I have done, that it was the later version of the De
therefore ad = ae, which was to be proved. duabus lineis that Werner read.
But if arc eg is less than a quadrant (see Fig. 6.18), the plane tangent to the The reader might be puzzled as to why Werner should have had recourse
cone on the side he of the cone will meet with the diameterfg extended. Let to the proofs of the De duabus lineis rather than to the proofs given in Con­
it meet with the diameter at point i . Line hi will necessarily pass through clusion 11 of the Speculi almukefi compositio, which he no doubt also knew
point e. Ordinate cd by hypothesis is at right angles to diameterfdg. Now if in Regiomontanus’ version. I am sure that he turned to the earlier work
we rotate the plane of the axial triangle of the cone through 90°, so that it because he found there the property of a hyperbola in relationship to the
lies in the plane of the base circle, with all of the straight lines in the rotated asymptote in a form that he wished to use later, and which was not present
plane remaining in length and position as they were before rotation, then in the Speculi almukefi compositio. As I have noted in my description of
point h will fall on the circumference of circle fcg because angle fhg is a Proposition 5 of the De duabus lineis, the author developed in steps (7)-
right angle andfg is the diameter of the circle. Now ci is tangent to the circle (9) the following equations (see Fig. 2.5):
at c and cd is perpendicular to fgi, the line drawn from a point outside the (7) eh2 = mq2 + 2 mq-mn
circle through the center of the circle. Hence we can now apply the thir­
teenth conic element. By its corollary angle dhg is equal to angle ghi. But (8) eh2 = uc2 + l u c - z ' c
the angles at a are right angles by construction. Therefore, triangles adh (9) mq2 + 2 mq-mn = uc2 + lu c -z 'c .
and ahe are congruent. Therefore ad = ae, which was to be proved.
Lastly Werner constructs a similar proof for the third case, namely for These are precisely equivalent to the equations presented in Werner’s
the case where arc eg is greater than a quadrant, which proof I leave the lemma to Element XX (see Fig. 6.23):
reader to examine in the translation. The reader should remember, though gh2 = ko2 + 2 ko -kl,
Werner forgets to tell us, that once more the plane of the axial triangle of the
cone must be rotated through 90° to be in the plane of the base circle. So far gh2 = pr2 + 2 rp -pq,
as I can determine, Werner’s complex proof of this element is entirely ko2 + 2 ko-kl = pr2 + 2 rp -pq.
original with him. It certainly differs completely from the proofs found in
Since both tracts make the point that these relationships hold for every
the Speculi almukefi compositio and Regiomontanus’ notes thereto. The
proof of the first case seems to be reflected in a proof presented by Fran- point on the hyperbola, we have in effect established, in both tracts, the fol­
ceso Maurolico (see the next chapter, note 31). lowing general property of the hyperbola. The sum of the square of the ex­
tension of any ordinate of the hyperbola to its asymptote and two times the
Elements X V II-X X . Having to this point completed all of the elements
concerning the parabola, Werner now proceeds to the hyperbola. The ob­ rectangle which has one side equal to that extension and the other side
ject of these four elements is to show that a curved line (a hyperbola) and a equal to the ordinate is a constant for any given hyperbola. Werner singles
straight line (its asymptote) can be constructed so that the more they are out these relationships in the lemma to Element XX because he will use
this characterization of a hyperbola (along with its axial property presented
extended the closer they approach each other without however ever meet­
earlier in Element XVII) to demonstrate in Element XXI that he has in­
ing. The source of these elements is certainly Propositions 4 and 5 of Ver­
sion B of the De duabus lineis. If one compares Werner’s Element XVII deed constructed a rectangular hyperbola and to prove in Element XXII a
258 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 259

further property of the hyperbola. Hence we may say that the crucial axial it is proved that (3) he-ef = e d 2. Thus Werner has shown that points
and asymptotic properties of the hyperbola that Werner learned from the k and d are on a hyperbola with asymptote be whose axial property is
De duabus lineis lie at the heart of his whole treatment of the hyperbola. expressed by ik2 = hi-if and ed2 = he-ef and whose asymptotic property
We should note that, unlike the author of the De duabus lineis, Werner is expressed by equations (1) and (2). Following the same procedure we
concludes (see Fig. 6.23) the proof of Element XX by saying thatp is closer can establish other points like k and d on the rest of the lines drawn
to line mh than is k because ko > pr. It would have been preferable to go parallel through the points of the divisions of ef. They would all have
one step further and say that because ko > pr the perpendicular distance the same asymptotic property as k and d. Thus the locus of all such
from k is also greater than the perpendicular distance from/?, and thusp is points would, by the converse of the lemma of Element XX, be a hyper­
closer to line mh than is point k . This trivial step was also left out of Conclu­ bola with asymptote be. Werner actually says that if we connect the
sion XI of the Speculi almukefi compositio. But the step is so obvious that established points by means of straight lines an indexed line would be
it is surprising that its omission was of such great concern to Francesco fashioned which “ is scarcely different from a hyperbola whose vertex is
Barozzi (see the next chapter, Text B, marginal page numbers 178-79). / and axis is ef. . . . For the hyperbola that is described through points
Element XXL “ To describe a hyperbola through a given point and be­ d and / and does not meet be will pass through the points assigned in
tween two asymptotic straight lines that meet at a right angle. ’’ Or, to put it the said way in de, ik and the remaining parallels, by the same lemma
in another way: To draw a rectangular hyperbola through a given point. It of the twentieth conic element. ’’ Obviously he intends that the more points
will be immediately evident that this element is more restricted than the that are established in the manner of point k, that is, the smaller we make
similar problem given in Proposition II.4 of the Conics of Apollonius, the divisions of line ef, the closer we shall approach a true hyperbola.
where the asymptotes are said to meet at any chance angle. And indeed The other half of the hyperbola that does not touch ba is constructed in
Werner’s presentation of the problem has no reference to the Apollonian the same way. Hence by this procedure Werner has constructed a rec­
proposition. Werner has probably restricted the angle of the asymptotes tangular hyperbola through point d with asymptotes be and ba. A lemma
to a right angle because of the fact that in the various cases where he needs is added by Werner to show how the straight lines de, ik, etc. (i.e. the
to construct a hyperbola in his tracts on the problems of the finding of two ordinates of the hyperbola) whose squares are equal to the rectangles
proportional means and cutting a sphere in a given ratio, the asymptotes he-ef, hi-if, etc. can be easily and quickly found (see Fig. 6.25). It is
indeed meet at a right angle (see below, notes 16, 18, and 19). That is to say, similar to the procedure for determining the ordinates of a parabola in
it is only the rectangular hyperbola that he later needs to construct. Wer­ Element XI and demands no special comment.
ner’s procedure in Element XXI commences with the bisection of the right Element XXII. ‘‘Straight lines drawn from a hyperbola to the asymptotes
angle abc by means of line be (see Fig. 6.24). (In Apollonian terms b would form equal rectangular areas.’’ This can be made more precise by refer­
be the center of the hyperbola to be constructed.) Through the given point ence to Fig. 6.26. If we have a rectangular hyperbola def with asymptotes
d a line de is drawn at right angles to be and extended on one side to point a ab and ac, and if we draw from point e on the hyperbola to asymptote
and on the other to point c . Then bf is marked off on be and is of such a ac the perpendicular eg and to the other asymptote ab the perpendicular
length that b p = {de + ec) -dc. Hence the v ertex/of the future hyperbola eh, and from point / we draw lines fl and fk so that fl is parallel to
is thus obtained. Linefg is drawn through/ parallel to ce. Hence angle bfg eg and fk is parallel to eh, then the rectangle ageh = rectangle alfk. This
is a right angle. Since anglefbg is half of a right angle by construction, so is equivalent to saying that the product of the perpendicular distances
angle bgf is also half of a right angle. Accordingly triangle bfg is an from any point on a hyperbola to the asymptotes is a constant. This
isosceles right triangle, and thus b f = fg. And so f g 2 = {de + ec)-dc. element was added by Werner because it was to be used by him in his
[And since ec = dc + de], therefore f g 2 = dc2 + 2 dc-de. Now extend version of Menaechmus’ solution of the proportional means problem (see
fb to h so that bh = f b . (Thus we have constructed hf, which in Apol­ below, note 16). The proof is simple and depends on Element XX. First
lonian terms would be the transverse side or diameter of the hyperbola gm (equal to eg) is cut off on eg. Then through points e and m is drawn
to be constructed.) Then let fe be cut into any number of segments, and line men. Similarly line Ic (equal to fl) is cut off on line ac and line
through the points of section let parallels to ac be drawn. For example, cfdb is drawn. Because the angles at m and n (above the line) are each
take segment ei, and through point i draw line il parallel to line ce. And one half of a right angle, hence (1) me / eg = en I eh, the ratios each
let ik be cut off from il such that ik2 —hi-if. Here he is making use of being that of a diagonal of a square to its side. Similarly (2) en / eh
the axial property of the hyperbola developed in Element XVII and taken = fc Ifl. Then since bfk and ehn are similar triangles, (3 ) b f I en = fk I eh.
from Proposition 4 of the De duabus lineis. He then goes on to prove Werner says that, by a corollary to Element XX, (4) bf-fc = me-en. [I
that (1) Λ:/2 + 2 ik-kl = f g 2. By construction, (2) f g 2 = cd2 + 2 c d -d e , and add parenthetically that the corollary to which he refers was not actually
260 ARCHIMEDES IN THE MIDDLE AGES
EARLY SIXTEENTH CENTURY 261

framed by Werner but can be easily proved. For the equation in the
In the twelfth appendix to this work on the duplication of the cube, he
lemma of Element XX, as adopted to the rectangular hyperbola of Ele­
briefly discusses the parabolic burning mirror, noting the determination
ment XXI (Fig. 6.24), would be expressed by dc2 + 2 ed-dc = kl2 + 2 ki ■kl.
of its focus as given in a work on parabolic burning mirrors (presumably
This equation may be rewritten as (2 ed + dc)-dc = (2 ik + kl)-kl.
the De speculis comburentibus of Alhazen, or, less probably, the Speculi
In the different lettering of Element XXII as given in Fig. 6.26, bf is similarly
almukefi compositio).17 In that appendix he remarks that the focus is at a
situated as (2 ed + dc),fc as dc, en as (2 ik -I- kl), and me as kl. Thus with
the substitution of the equivalent magnitudes, bf-fc = me-en, as Werner finitionem sextam elementorum conicorum, deductae ab ipsa bef parabola ad ab axem pos­
has expressed the corollary to Element XX.] And (5) bf I en = em I fc sint areolas rectangulas quarum longitudo quidem ab, latitudines autem segmenta axis ab
= eg / fl. And so with (3) and (5) together, it follows that (6) fk / eh sumpta inter b apicem et structim deductas, per quintum elementum conicum, ad ab axem.
= eg I fl. And so it has been shown that the sides about equal angles Atque per d signum ad ipsas ab, be non coincidentes, per vigesimum primum elementum
conicum, hyperbole scribatur de, secans parabolen bef in e. Atque ipsis ab, bc paralleli
geh and Ifk are proportional and hence that the rectangles with these
agantur eg,eh ,eg quidem ipsi b c , sed eh ipsi a b , secans bc productam in h . Et quoniam ab axis
sides are equal.
aequalis est ipsi adf recto lateri paraboles bef, igitur eg structim acta ad ab potest id quod sub
So much, then, for this interesting tract of Werner’s. My rather extended ab, bg fit rectangulum per quintum elementum conicum. Ergo per propositionem xvii. libri
account of it has shown that the sources on which he depended were vi. elemen. Eu. tres rectae lineae ab, eg, bg sunt continue proportionales. Praeterea quia duo
essentially three: (1) the scattered passages from Valla’s De expetendis parallelogramma rectangula abed, egbh ad non coincidentes ab, bc ex d, e punctis ipsius
et fugiendis rebus opus, for various definitions, postulates and the objec­ hyperboles habent latera deducta, igitur per vigesimumsecundum elementum conicum
rectangula abcd, bgeh sunt sibi invicem aequalia. Ergo per propositionem xiiii. Ii. vi. ele­
tives of some propositions; (2) the Speculi almukefi compositio in Re­ mentorum Euclidis ratio ab ad eg ut gb ad bc. Sed ut antea fuerat demonstratum ab ad eg
giomontanus’ version, for the basic properties of a parabola generated est ut eg ad gb. Igitur quattuor rectae lineae ab, eg, gb, bc sunt continue proportionales.
from a right-angled right cone; and (3) the De duabus lineis, translated Datis ergo duabus rectis lineis ab, bc binae mediae proportionales compertae sunt eg, gb,
from the Arabic by John of Palermo, for the basic propositions concern­ quod oportebat efficere.
“ Aliter ut Menechmus per communes binarum parabolarum sectiones. Datis duabus rectis
ing the properties of a hyperbola. In a certain respect, we can say that Wer­
lineis binas medias continue proportionales invenire. Sint igitur datae duae rectae lineae
ner’s work was the last original tract produced out of the medieval traditions ab, bc [Fig. 6.28] quibus oporteat binas medias proportionales invenire. Igitur vertice b, axe
of conic sections, though, as we shall see in the next chapter, it was by et recto latere ab, ad quod ordinatim deductae possunt, per undecimum elementum conicum
no means the last work to reproduce the conclusions of the medieval parabole rectanguli coni scribatur bde. Rursus apice b , axe seu latere recto b c , ad quod struc­
traditions. Though I have emphasized the probable sources available to tim actae possunt, rectanguli coni parabole scribatur bdf, secans bde parabolen in d, a quo
ipsis ab, bc paralleli agantur dg, dh, quarum dg quidem secet ab in g, dh vero ipsam bc
Werner, I have not sufficiently stressed the fact that his selection of the
productam in h. Et quia per quintum elementum conicum quod est sub abg rectangulum
particular elements to be included in his work was significantly governed aequum est ei quod ex dg quadrato, igitur per propositionem xvii. libri vi. elemen. Eu.
by what was needed for the next two works of his omnibus volume of a b ,d g, gb sunt continue proportionales. Praeterea per sextum elementum conicum quadratus
1522: Commentarius seu paraphrastica enarratio in undecim modos con­ ex dh aequalis est ei quod ex hb, bc parallelogrammo rectangulo; igitur per propositionem
ficiendi eius problematis quod cubi duplicatio dicitur and Commentatio xvii. libri sexti elementorum Euclidis tres rectae lineae bh, dh, bc sunt continue propor­
tionales. Est autem dh aequalis ipsi bg et bh ipsidg aequalis; igitur sub eadem ratione d g , g b ,
et paraphrasis . . . in Dionysodori et Dioclis problema super sectione
bc sunt continue proportionales. Et quia ut pauloante fuerat ostensum ab ad dg est ut dg ad
sphaerae sub data ratione (the title as given before the work rather than gb, igitur per propositionem xi. Ii. quinti elementorum Eu.— Quae eidem sunt eaedem ra­
that given in the general title of the volume, which see in the Bibliography tiones adinvicem sunt eaedem— quattuor rectae lineae ab, dg, bg, bc sunt continue pro­
below), works that he composed on the basis of the passages that Valla portionales. Datis ergo duabus rectis lineis ab, bc binae mediae continue proportionales
translated from Eutocius’ Commentary on the Sphere and the Cylinder d g ,g b sunt compertae, quod oportuit demonstrare.” As usual I have punctuated as I saw fit
o f Archimedes. In the first work he refers to his own work on conic and italicized letters representing magnitudes.
17Ibid., sign, [h iv recto]: “ Appendix duodecima. Speculum concavum concavitate
elements only in the solution attributed to Menaechmus and in the solution parabolica, quam describit parabole circumacta defixo eius axe, solum a tota concavitatis
that succeeds it, in which solutions Werner cites his sixth definition, superficie solis radiis ad unum punctum axis resilientibus ignem incendit. Eandem autem
and his fifth, sixth, eleventh, twenty-first, and twenty-second elements.16 concavitatem oportebit fieri ab ea parabole quae in rectangulum et erectum incidit conum,
qualem quidem parabolen undecimum elementum describere docet. Et quia radii solis prope
tellurem paralleli sunt per xi. appendicem, ideo ipsi cadentes in speculum concavum para­
16 Werner, Commentarius seu paraphrastica enarratio in undecim modos conficiendi
bolica concavitate ab ea resultant ad unum tantum punctum, quod iuxta demonstrationem
eius problematis quod cubi duplicatio dicitur, sign, g ii recto— verso: “ Ut Menechmus per
authoris libelli de eodem speculo distat a vertice poraboles (/) qua idem speculum fuit
sectionem paraboles et hyperboles. Datis duabus rectis lineis binas medias continue pro­
cavatum quarta parte lateris recti eiusdem paraboles. At in speculis concavis concavitate
portionales invenire. Sint ergo duae datae rectae lineae ab, be [Fig. 6.27] rectum com-
sphaerica radii incidentes ad diversa puncta . . . reflectuntur. . . . Qui denique speculum
praehendentes angulum abc inter quas oporteat binas medias proportionales invenire.
parabolicae concavitatis preparare velit, huic necessaria est cognitio undecimi elementi
Compleatur parallelogrammum abcd. Et ab axe, vertice vero b, per undecimum elemen­
conici quo talis parabole describi docetur.” I think it likely that the treatise on the parabolic
tum conicum parabola scribatur bef, cuius rectum latus adf aequale sit ipsi ab, sicut per dif-
mirror mentioned by Werner was Alhazen’s De speculis comburentibus rather than the
262 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 263

distance from the vertex of the parabola equal to one fourth of the length Werner’s Libellus super vigintiduobus elementis conicis does not seem
of the latus rectum. He also mentions there Element XI of his own tract. to have been widely cited, perhaps because it was rendered obsolete
In the second short work on the division of the sphere, he gives Dionyso- by the various translations of the Conics of Apollonius and sundry works
dorus’ solution and mentions in his account “the converse of [his own] using these translations which I shall mention in the next chapter. How­
fifth or sixth conical element,” his twenty-first, and his twenty-second ever, there exists a rather pedestrian and unoriginal De speculo parabolico
element (which latter he identifies with the eighth theorem of the second composed on the basis of Werner’s Libellus and Alhazen’s De speculis
book of the Conics of Apollonius).18 This identification with a proposition comburentibus and now extant in a single Viennese manuscript (National-
from the Conics of Apollonius he took not from Apollonius’ work itself bibl. Cod. 5277, 224r-30r). Its preface imitates compendiously that of
but directly from the Eutocian account of Dionysodorus’ solution, which Alhazen’s De speculis comburentibus but also reflects the twelfth appendix
he (Werner) had summarized from Valla’s work (see note 5 above). In of Werner’s tract on the duplication of the cube (see note 17 above).20
fact, the naming of this proposition as the eighth of the second book of
the Conics, without further comment, removes any doubt that Eutocius Decet igitur dispescere sphaeram plano ad ab recto ut segmentum cuius vertex a ad seg­
(through Valla) was his source, for this proposition in all of the manu­ mentum cuius vertex b rationem habeat quam de ad ef. Et fiat ut d f ad fe sic ac ad g. Et
producatur ab in a partem usque in h, sitque ah aequalis ipsi ac, et ipsi ab in b ad rectos an­
scripts of the Conics extant in the Renaissance was not the eighth but the
gulos excitetur bk, et fiat ut bh ad g sic ab ad bk. Compleaturque parallelogrammum abkl.
twelfth. Further, that Valla was his source is clear from the fact that when Rursus bkm sit dupla ipsius bk, et perficiatur bhmn parallelogrammum. Et per m ipsis bh,
he mentions this “eighth” theorem, he uses Valla’s exact phrase: “per oc­ hn non coincidentibus per xxi. ele. conicum scribatur hyperbole mop. Praeterea per xi.
tavum theorema, idest praeceptum, secundi libri Apollonii conicorum conicum elementum scribatur parabole cuius axis bm et vertex b, ad quam vero structim
elementorum” (again see note 6 above), thus including Valla’s explana­ actae possint sit aequalis ipsi bc. Sitque talis parabole boq, secans mop hyperbolen super o.
E tex o ad ab perpendicularis agatur or. Dico quod planum veniens perr et adab rectum secat
tory phrase “ idest praeceptum.” To the solutions of Dionysodorus and
datam sphaeram sub ratione ipsius de ad ef. Ipsi denique bc aequalis fiat rcu, perficianturque
Diodes (in which latter he gives no reference to conic sections), Werner tria parallelogramma mbct, orux, orhs. Et quia per constructionem ut d f ad fe sic ac ad g,
adds a solution of his own.19In that solution he refers to the fifth, eleventh, igitur solidum parallelepipedum cuius basis quadratus ipsius ab, altitudo vero ipsi ac aequalis,
twenty-first, and twenty-second elements of his own tract. ad solidum parallelepipedum cuius eadem basis, altitudo autem ipsi g aequalis, rationem
habet quam d f ad fe. Quoniam autem ex hypothesi ut bh ad g sic ab ad bk, igitur per primam
propositionem lib. vi. ele. Eu. quadratus ipsius ab ad rectangulum aM/ existit ut bh ad g. Igitur
Speculi almukefi compositio, because, like Alhazen but unlike the author of the Speculi solidum parallelepipedum cuius basis aequalis quadrato ab, altitudo vero aequalis ipsi ^ , par
almukefi compositio, Werner contrasts the parabolic mirror with the spherical mirror. See the est solido cuius basis rectangulum abkl, altitudo vero ipsi bh aequalis, per propo. xxxiiii.
De speculis comburentibus, ed. of Heiberg and Wiedemann, pp. 218-19. lib. xi. ele., bases enim ipsis fastigiis sunt reciprocae. Et quia duorum parallelogrammorum
18 Werner, Commentatio et paraphrasis . . . in Dionysodori et Dioclis problema super abkl et bmtc latera per constructionem sunt reciproca iuxta eundem communemque angulum
sectione sphaerae sub data ratione, sign, i i recto— verso: “ Ut Dionysodorus. Datam abm constituta, igitur per propo. xiiii. Ii. vi. ele. Eu. eadem parallelogramma abkl, cbmt
sphaeram plano secare ut ipsius segmenta rationem adinvicem habeant datam. Sit data sunt aequalia. Duo igitur solida quorum bases sunt parallelogramma abkl, cbmt, altitudines
sphaera cuius diameter ab [Fig. 6.29], data autem ratio quam habeat cd ad de. Convenit autem ipsi bh aequales, paria sunt per propo. xxxi. Ii. xi. ele. Igitur ex communi sententia—
nempe secare sphaeram plano recto ad ab ut segmentum cuius vertex a ad segmentum cuius Quae uni sunt aequalia etc. — solidum parallelepipedum cuius basis cbmt, altitudo vero ipsi
vertex b rationem habeat quam cd ad de. Producatur ba in/ , ponaturque ipsius ab dimidia bh aequalis, aequatur solido parallelepipedo cuius basis quadratus ipsius ab, fastigium autem
af. Et quam habeat rationem ce ad ed eandem habeat a f ad a g , sitque ag ad rectos angulos ipsi ipsi g aequale. Praeterea, quia duo parallelogramma bhmn et orhs compraehenduntur actis
a b . Et ipsarum/ α , ag media proportionalis sumatur ah; maior igitur ah existit quam a g . Et si ab hiperbole (/) mop ad non coincidentes bh, hs rectis lineis, igitur per ultimum elementum
circa axemfb descripta fuerit parabole cuius structim deductae possint ad ag, ipsa parabole conicum duo parallelogramma bhmn et orhs sunt aequalia. Et quia per propositionem xvi. Ii.
ibit per h per constructionem atque per conversionem quinti aut sexti elementi conici, vi. el. Eu. — Si sub extremis compraehensum rectangulum etc.— igitur ut bh ad hr sic ro ad
quoniam per propositionem xvii. libri vi. ele. Eu. quod est sub fag aequale est ei quod est ex bm. At ex hypothesi atque per propo. primam li. vi. ele. Eu. ut or ad bm sic parallelogram­
ah. Describatur itaque parabole haec per elementum conicum xi., sitqueflilk, et per b traduca­ mum orux ad cbmt parallelogrammum. Igitur solidum parallelepipedum cuius basis orux
tur ipsi ab ad rectos angulos bk, secans parabolen flilk in k signo. Et per g circa non coinci- parallelogrammum, fastigium autem hr, aequatur solido cuius basis cbmt, altitudo autem
dentes fbk per xxi. elementum conicum describatur hyperbole gl, quae nimirum secabit bh. Atqui per quintum elementum conicum quadratus ipsius br aequatur parallelogrammo
parabolen inter h,k. Secet igitur in/, et ex / in ab perpendicularis ducatur Im, et per g , l ipsi ah orux. . . .” After this point there is no further citation of the conic elements and so I have
paralleli agantur gn, Ix. Quoniam igitur hyperbole est gl, non coincidentes autem abk, et omitted the remainder of this long proof.
paralleli ipsis agn sunt ipsae mix, igitur aequale est quod sub agn ei quod sub mix per octavum 20 Vienna, Nat.-bibl. cod. 5277, 224r: ” [P]arabolicum speculum est cuius superficiem
theorema, idest praeceptum, secundi lib. Apollonii conicorum elementorum seu per xxii. ele. concavam describit parabole coni rectanguli circunducta de fixo eius axe. Hoc speculum
libelli de elementis conicis. . . . ” From this point on there is no more use of conic elements multo fortius celeriusque quam sphaerice concavam incendit. Cuius ratio est quia omnes
and hence I have not given the remainder of the long proof. radii solares in superficiem eius concavam incidentes, si directe ipsum soli opponatur, ad
19Ibid., sign, i iii recto— [i iv recto]: “ Ut Ioannes Verner Nurembergensis. Datam punctum unum reflectuntur. In speculo autem sphaerice concavo soli radii in eundem cir­
sphaeram plano secare ut ipsius segmenta rationem adinvicem habeant datam. Sit data culum in speculo descriptum incidentes, ad unam eius partem reverberantur, id quod
sphaera cuius diameter ab [Fig. 6.30], et centrum c, data autem ratio quam habeat de ad ef. Euclides in sua specularia demonstrat. Quod sit punctum illud concursus radiorum para-
264 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 265

The definitions of the anonymous work were taken from Werner’s Libel­ comburentibus, Proposition 10 having been taken from the eleventh ap­
lus.21 The same is true of Propositions 1-2 and 4 -5 , Proposition 3 being pendix of Werner’s treatment of the duplication of the cube.23 In addition
little more than a translation of the definition of a parabola into a con­
ducta erit parabole. . . . Propositio quarta ( = Werner, Lib., Elem. V -V I). Si a parabola
structional problem, and Proposition 6 being a statement of the basic
rectanguli coni ad eius axem ordinarim agatur linea recta, quadratum eius erit aequale
optical theorem which asserts the equality of the angles of incidence and rectangulo quod continetur sub latere recto eius paraboles et sub linea contenta inter illam
reflection, whose demonstration the author attributes to Euclid’s Specu­ ordinatim actam et verticem paraboles. Haec propositio est quinta et sexta elementorum
laria.22 Propositions 7 -9 and 11-13 originated in Alhazen’s De speculis conicorum cur ab eius demonstratione supersedo, sola oculari demonstratione contentus.
Propositio quinta ( = Werner, Lib., Elem. XVI). Si linea recta contingit parabolam rectanguli
bolico speculo incidentium, et quamobrem sit necesse omnes radios incidentes, ad punctum coni, sique eadem contingens et paraboles axis in eandem partem producuntur quousque
unum resilire, et quo nam modo speculum ipsum commodius paretur est animus in hoc libro concurrant, erit axis producti pars exterior, quae scilicet paraboles vertici dictique concursus
docere (supra scr. m. 2: hoc libello docebimus), praemissis quibusdam ad eam rem intel- puncto interiacet, aequalis illi portioni axis quae inter eundem paraboles verticem et punctum
ligendam necessarius.” I have made a few changes in punctuation and capitalization, axis in quod cadit linea ordinatim acta a puncto contactus interiacet. Haec propositio est 16.
and have expanded ς to ae. In this and the succeeding notes I have occasionally given some elementorum conicorum. Ideo solam figuram oculis subiiciemus; demonstrationem apud
corrections that are in a second hand, which I have labelled m. 2. Incidentally, I shall not in­ Vernerum invenies. Proposito sexta. Speculis concavis radius incidens et reflexus faciunt
clude the proofs of the various propositions, though I shall extract remarks that mention angulos aequales cum linea contingente quae est communis sectio superficiei planae con­
Werner’s work. tingentis superficiem speculi in puncto in quo radius incidit et superficiei planae protractae in
21 Ibid., 224r-v: “ Definitiones. [1] ( = Wemer, Libellus, Def. 1) Conus seu pyramis est qua sunt radii, scilicet incidens et reflexus. Hanc propositionem Euclides demonstrat de
figura quae fit quando rectanguli trianguli manente uno eorum quae circa rectum sunt angulum speculo sphaerice concavo propositione prima suae speculariae. Cuius demonstratio potest
latere circumductus triangulus in idem rursus unde sumpserit exordium revolvitur. [2] facile etiam applicari ad speculum parabolice concavum. ” The reference by m. 2 to Rogerum
( = Werner, Lib., Def. 2) Axis coni est linea recta circa quam triangulus vertitur, [add. m. 2: is, I believe, a reference to Conclusion 1 of the Speculi almukefi compositio, which, the reader
Triangulus ab axe coni est communis sectio conicae superficiei et plani secantis conum per will recall, was attributed in two manuscripts to Roger Bacon (see Chap. 4 above, Text,
verticem secundum eius axem (del. et add.: super eius axe).] [3] ( = Werner, Lib., Div. 1) Variant readings for the title).
Conorum alius rectangulus, alius obtusiangulus, alius acutiangulus. Si manens recta linea 23 Ibid., 226v-30r: “ Propositio septima ( = Alhazen, De spec, comb., Prop. 1; cf. Chap. 1
fuerit reliquae quae circa rectum est angulum circumductae aequalis, conus erit rectangulus. above, n. 33 and the edition of Heiberg and Weidemann cited there). In sectione parabolica si
Si vero minor, obtusiangulus. Sin maior, acutiangulus. [4] ( = Werner, Lib., Post. 5) Conica de axe eius separetur quarta pars lateris recti, incidatque in eam linea aliqua aequidistans axi
sectio est communis differentia conicae superficiei et plani neque per verticem coni venientis et altera resiliat ad punctum axis (add. m. I?) quod terminat quartam partem lateris recti, hae
neque aequidistantis basi. [5] ( = Werner, Lib., Div. 3) Conicarum sectionum alia dicitur duae lineae continent cum linea contingente parabolam super illud (dei. m. 2?) punctum in
parabole, quam Mukefi Arabes appellant, alia hyperbole, alia ellipsis [note: the Arabie name quod (dei. m. 2?) incidit (corr. m. 2 in incidentem) aequidistans (dei. m. 2?) duos angulos
for parabola was not given by Werner]. Parabole est quando planum secans conum ad aequales. . . . (note: the proof contains the synthetic proofs given by Alhazen but omits the
triangulum ab axe erigitur [ad rectos angulos], planique ac trianguli ab axi communis sectio analytic proofs). Propositio octava ( = Alhazen, De spec, comb., Prop. 2, ed. of Heiberg and
reliquo trianguli ab axe lateri aequidistat. Hyperbole est quando plani secantis et trianguli ab Wiedemann, p. 224). Si parabole, axe fixo, volvatur donec redeat ad locum a quo moveri
axe communis sectio reliquo eiusdem trianguli ab axe lateri producto ultra coni verticem caepit fiet superficies parabolice concava. Et si ab eius axe egrediatur quaecunque super­
coincidit. Ellipsis fit quando plani secantis et trianguli ab axe communis sectio cum reliquo ab ficies plana, erit communis sectio superficiei concavae et planae parabole aequalis et similis
axe trianguli latere intra conum coincidit, [6] ( = Werner, Lib., Def. 6) Axis sectionis [add. ei cuius motu ipsa superficies concava producta est. Utraque pars huius propositionis per se
m. 2: conicae] est linea recta in duo aequalia dividens rectas [add. m. 2: a conica sectione] ad patet, neque ulla eget demonstratione, est enim definitioni quam propositioni similior.
se perpendiculariter cadentes. Vertex sectionis est supremus punctus axis. Linea ordinarim Propositio nona (= Alhazen, De spec. comb. Prop. 3, ed. cit., p. 225). In omni superficie
acta est linea recta a sectione [add. m. 2: conica] ad axem eius ad rectos angulos deducta. parabolice concava si de axe separetur quarta pars lateris recti paraboles eius cuius cir­
Latus rectum est quae ex lineis ordinarim actis sola aequalis axis portioni ad se terminatae.” cumductu haec superficies producta est, et si incidat in eam linea recta aequidistans axi,
22Ibid., 225r-26v: “ Propositio prima (=Werner, Lib., Elem. VIII). Data parabola coni reflectaturque ad punctum finale lateris recti quartae partis, continent hae duae lineae,
rectanguli, eiusque axe dato, latus eius rectum invenire. . . . Cuius demonstratio apud scilicet incidens et reflexa, duos angulos aequales cum linea contingente quae est communis
Ioannem Vernerum Nurmbergensem absolutissime habetur. Propositio secunda ( = Werner, sectio superficiei planae secantis (dei. m. 2?) contingentis superficiem parabolica[m] in
L ib., Elem. XI). Dato latere recto, parabolen coni rectanguli in plano dato describere. . . . puncto incidentiae et superficiei planae protractae in qua sunt lineae incidens et reflexa. . . .
Huius demonstrationem a Vernero require. Propositio tertia. Data cono rectangulo para­ Propositio decima (Werner, Comm, seu paraphr. enarratio in undecim modos e tc., sign, h iii
bolam aliquam resecare. Sit conus datus cuius vertex a , basis circulus bde [Fig. 6.31], Ducatur recto). Radii solares apud td^am sunt aequidistantes sive a centro sive ab alio puncto
in basi diameter [mg. add. m. 2: be], cuius extremitates cum duabus a vertice rectis lineis procedant. Huius propositionis duplicem propositionem (del. m. I?) assert demonstrationem
iungantur. Deinde de super illas duas lineas et diametrum basis secetur conus bifariam. doctissimus Mathematicus Ioannes Vernerus Nurmbergensis, ex quo petas si demonstra­
Eritque in hemicono triangulus qui alias dicitur triangulus ab axe abe [mg. add. m. 2: Hic vide tione dedaris. Propositio undecima ( = AIhazen, De spec, comb., Prop. 4, ed. cit., p. 226).
Rogerum], cuius duo latera ab, be divide per aliquam lineam proportionaliter, quae linea sit Quando superficies specularis parabolici concava soli directe opponitur, omnes radii ei
de, ita ut sit proportio ad ad db sicut ce ad eb. Sic erit linea de aequidistans reliquo lateri incidentes ad unum punctum resiliunt, quod distat a vertice per quantitatem quartae partis
trianguli, scilicet ac, per secundam partem secundae sexti Euclidis. Deinde a puncto e in basi lateris recti paraboles eam superficiem producentis. . . . Propositio duodecima ( = Alhazen,
hemiconii (!) duc perpendicularem ad lineam bc eamque protrahe usque ad circumferentiam De spec. comb. Prop. 5, ed. cit., p. 228; cf. Spec, almukefi compos., Conci. 10). Speculum
semicirculi basis et sit eh. Postea iunge puncta h , d filo aliquo subtili valde intento super parabolicum construere. . . . Propositio tertia decima( = Alhazen, De spec, comb., ed. cit.,
superficiem coni, secundum quod filum ducas lineam in superficie coni. Linea haec ultimo pp. 228-29). . . .”
266 ARCHIMEDES IN THE MIDDLE AGES EARLY SIXTEENTH CENTURY 267

to the use of Werner’s Libellus in this anonymous tract we should note sections, nor indeed in the revived Apollonian tradition that was soon
its influence on Girolamo Cardano’s treatment of the parabola (see the to follow. I shall limit myself to describing the method for drawing a
next chapter, note 60) and its possible influence on Maurolico’s De lineis parabola (Fig. 6.32).26 Durer first draws the axial triangle abd which repre­
horariis liber tertius, written in 1553 (see the next chapter, note 31). Fur­ sents the cone. The line ace represents the plane through the axis of the
thermore, Francesco Barozzi in an omnibus work on the relation of a cone and is perpendicular to the base. The line bd represents the base
hyperbola to its asymptote borrows heavily from Werner’s treatment of circle of the cone. The parabolic section is represented by the lincf gh,
this problem.24 which line is equal to the axis of the parabolic section. Hence the whole
Not long after the appearance of Werner’s volume, Albrecht Durer triangle represents a projection on the axial triangle of the cone and its
gave a brief treatment of the description of conic sections in his Under- section in elevation. Line fgh is divided into twelve equal parts and the
weysung der Messung (Nuremberg, 1525). It has only the briefest re­ eleven points of division are numbered as shown in the diagram.
marks on the definitions of the three conic sections, which do not appear Through these points of division we first draw horizontal lines perpendic­
to have been influenced by Werner’s work.25 In Diirer’s preliminary re­ ular to line ace, the lines through points 1 and 2 proceeding to side ad,
marks, the ellipse is described as the section produced by a plane that while the lines through points 3-11 proceed to side ab. Then through
cuts the cone (which appears from the diagram to be a right cone, but not these same points of division we draw vertical lines parallel to line ace,
a right-angled cone) obliquely, the parabola as the section produced from these lines representing planes that are perpendicular to the base and
the same kind of cone by a cutting plane that is parallel to one side of parallel to the axis. Directly below the triangle representing the pro­
the cone, and the hyperbola as the section produced from the same kind jection of the cone in elevation is drawn a circle equal to the base circle
of cone by a cutting plane that is parallel to the axis of the cone. The that represents the projection of the cone on the base circle. This
generation of the hyperbola is the only one that is unusual. So far as circle Durer calls the Grund (in Latin, fundamentum) of the cone.
I know, the only work that could have been available to Diirer in which Thus in effect we have the cone now projected both in elevation and
such a generation of the hyperbola appears is the Speculi almukefi com­ plan. Then the vertical lines through the numbered points of division
positio, Conclusion 11, and I suspect that he had read that work in the along line fgh in the triangle are extended so that they cut the circle
Regiomontanus version (see Chap. 5 above, Text C). Durer’s definition below, thus producing an equivalent number of sections of the circle,
of the ellipse is not very precise, but presumably he meant, as his later
26 Durer, Unterweysung der Messung, sign. E iiii recto— verso. Cf. Institutiones geo­
drawing of the ellipse implies, that all of the elements of the cone’s sur­
metricae,ρ . 31: “ Parabola eoferme modo quo ellipsis fit. Primo protraho pyramidem abcde
face had to be cut to produce the ellipse. Perhaps he was influenced [Fig. 6.32], atque in ea lineam perpendicularem a, secoque pyramidem a summo deorsum per
by the remarks made toward the end of Proposition 1.103 of the Perspectiva basim eius, ita quod haec sectio sit parallela lateri pyramidis ab, voceturque superne/ et
of Witelo (see Chap. 3 above, n. 26). inferne gh . Postea partiorfgh in 12 spacia aequalia et duco lineas transversas per omnia puncta
Durer follows these general remarks on the conic sections with an in­ ipsius fg h , ac per illa puncta quae sunt versus illud latus ad traho transversas ab erectaa usque
ad idem latus. Sed per illa quae sunt ex altero latere duco transversas a perpendiculari a ad
genious method of drawing them. The method is precisely the same for
latus pyramidis a b . His absolutis facio fundamentum pyramidis directe sub ea [i.e. pyramide],
each section and has no counterpart in the medieval traditions of conic cuius centrum sit a et circumferentia bcde, et sino ex omnibus divisionibus sectionis fgh
lineas perpendiculares cadere deorsum per rotundum fundamentum, in quo ipsis etiam adiicio
24 See the next chapter, Text B, Extract III, where Barozzi has derived his first seventh, suos numeros, quemadmodum prius in fundamento factum est. Iam ut etiam in ellipsi
eighth, and ninth demonstrations from Werner and where he has refashioned his first three edocuimus, aperto circino ad quantitatem lineae transversae 1 in pyramide, pono pedem in
propositions leading up to the demonstration from Werner’s eighteenth, seventeenth and centrum a fundamenti et alium in lineam perpendicularem, et inde scribo arcum versus d et
nineteenth elements. His critique against Werner (Text B, marginal page numbers 175-79) is ultra ipsum usque ad eandem lineam 1. Sic procedo per omnes lineas notatas numeris donec
not of great significance. ad gh venio videbiturque ilico parabolae sectio in fundamento depresso. Quo facto describe
25 Durer, Unterweysung der Messung, sign. E iii verso. A facsimile published in Zurich in parabolam ex iam protracto fundamento, hoc pacto. Primo duco lineam transversalem, cui
1966 gives ready access to the German text. For comparison I give the Latin translation by J. erigo ad angulos rectos lineam quae sit superne /tan tae quantitatis quanta est sectio fgh in
Camerarius, the title of which is ordinarily shortened to Institutiones geometricae (Paris, pyramide. Deinceps accipio ex fundamento latitudinem gh et traduco eam in lineam trans­
1532) (for fuller title, see the Bibliography below), p. 29: “ Veteres nobis monstraverunt tres versam, sic quod erecta/stet in medio, et extremitates utrinque signo literis g et h , dividoque
pyramidis sectiones quae inter se differunt, nec tamen habent similem cum basi pyramidis erectam /in duodecim partes aequales, quibus suos numeros adiicio superne incipiendo, et
circumferentiam: alias posset etiam pyramis per medium secari, fieretque sectio ipsi pyramidi per singulas divisiones produco lineas transversas tantae longitudinis quanta opus est, et
similis, quod non curatur: at quaelibet aliarum trium sectionum facit singularem lineam, eas transferro ex fundamento omnia intervalla inter sectiones arcuum et rectarum linearum bd
lineas protrahere docebo. Prima igitur sectio vocatur Ellipsis, ea secat pyramidem oblique, contenta ad perpendicularem / , ex cuius utroque latere signo transversas. Deinde traho
et basi nihil aufert, est tamen ei una parte proprior, altera vero remotior. Secunda sectio est lineam parabolae de puncto ad punctum quemadmodum hic etiam delineavi [in Fig. 6.32].”
parallela uni lateri pyramidis, et nuncupatur a doctis Parabola. Tertia sectio est aequidistans Following this passage, Durer gives a short passage ( Unterweysung, sign. E iiii verso·,Institu­
axi pyramidis, eam appellant Hyperbolen." tiones, pp. 32-33) that shows how to form a burning mirror from this parabolic line.
268 ARCHIMEDES IN THE MIDDLE AGES
Johann Werner
and these lines in the circle are numbered in the same way as they
were in the triangle. Then one foot of a compass is placed on the point The Booklet on Twenty-two Conic Elements
in line ace where that line is intersected by the horizontal line through
point 1. The other foot is placed on the point where that same horizontal
line intersects side ad. Preserving this opening of the compass, we then a Hi Γ / LIBELLUS IOANNIS VERNERI
place one foot of the compass on the center a of the circle below and Nurembergensis super vigintiduobus elementis Conicis.
draw an arc that cuts a chord on the vertical line in the circle designated
by number 1. In the same way the compass is successively opened by dis­ DIFFINITIO PRIMA
tances equal to the horizontal lines drawn through the remaining points
of division in the triangle, and with these successive openings chords are Conus est figura quae fit quando rectanguli trianguli manente uno eorum
successively intercepted on the vertical lines in the circle that are desig­ 5 quae circa rectum sunt angulum latere circumductum triangulum in idem
nated by the succeeding numbers. Also represented on the figure is a chord rursus unde sumpserit exordium circumvolvitur. Et si manens recta linea
intercepted on the line designated / in the circle by means of a compass aequa fuerit reliquae, quae circum rectum est angulum circumductae
opening equal to the horizontal distance from/ to line ace in the triangle. orthogonius seu rectangulus erit conus. Si vero minor ambligonius seu
Also represented is chord gh equal to the base of the parabolic segment. obtusiangulus. Sin autem maior oxygonius seu acutiangulus. Aliter [Fig.
It is clear that each of these chords intercepted on the vertical lines 10 6.1]. Conus est figura quae fit si a puncto aliquo ad circuli circumferentiam
in the circle represents (1) the diameter of a circle produced by a plane qui in eodem plano non est, et eidem puncto recta linea coniuncta utrinque
that cuts the cone and the parabolic section and is parallel to the base porrecta et puncto manente circumacta recta linea iuxta circuli circum­
of the cone and at the same time (2) a chord (i.e. double ordinate) of ferentiam, donec ad idem rursus convertetur unde ferri incoepit. Descrip­
the parabolic segment, a fact that Diirer could have learned from any of tam itaque a recta linea superficiem, quae conficitur ex binis superficiebus
the previous treatments of the parabola or merely by inspection of the 15 ad verticem invicem positis quarum utraque augetur in infinitum descrip­
cone and its section. Thus we are now prepared to draw the parabola. tione rectae ad utramque partem in infinitum productae, Mathematici
On a continuation of base line bd in the triangle we lay off line gh equal conicam vocant superficiem. Verticem vero ipsius eminens ibidem punc­
to the line gh that represents the last vertical line in the circle and thus tum. At axem, per punctum illud eminens et centrum circuli actam rectam
the line that is the base line of the parabolic segment. And on the middle lineam. Coni denique basim circulum illum.
point of that line we erect the section’s axial line that terminates in /.
This axial line we divide as before into twelve equal parts; through each 20 DIFFINITIO SECUNDA
point of division is drawn a traverse horizontal line equal to the chord Axis coni est manens quaedam recta linea, uti in prima coni diffinitione,
intercepted in the correspondingly numbered line in the circle, the middle circa quam lineam rectam triangulum vertitur.
points of the horizontal lines corresponding to the middle points of the
chords. Then through all of the termini of these horizontal lines and points DIFFINITIO TERTIA
/ , g and h we can draw a parabola.
We have thus completed our account of the treatment of conic sections a iii v Basis coni est circulus, iuxta primam coni diffinitionem, sub cir/cum-
in the first quarter of the sixteenth century, and are now prepared in the 25 ducta recta linea descriptus.
final chapter to examine the remaining influence of the medieval traditions
of conic sections in the next generation or so. DIFFINITIO QUARTA
Vertex seu fastigium culmen sive apex coni est punctus ille summus fixae
circa rectum angulum rectae lineae velut in prima Coni diffinitione. Denique
axis, basis, et vertex Coni aliter definiuntur in secunda definitione Coni.

30 Divisio Prima
Conorum alius orthogonius seu rectangulus, alius amblygonius seu ob­
tusiangulus, alius oxygonius seu acutiangulus; huius divisionis particulae
ex prima coni diffinitione patescunt.
269
270 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 271

Divisio Secunda verticem concidit. Ellipsis autem fit quando eadem communis dictorum
planorum sectio cum reliquo ab axe coni trianguli latere intra conum concidit.
35 Conorum alius rectus, alius scalenus seu inclinatus. Rectus est qui ad
rectos ipsi basi angulos axem habet. Scalenus seu inclinatus qui non ad
7o DIFFINITIO SEXTA
rectos ipsi basi angulos habet axem. Haec divisio in quamlibet trium
specierum primae divisionis cadere potest. Conicae sectionis axis, seu ut alii dicunt, diameter, est recta linea quae
super se ad rectos deductas angulos a conica sectione rectas lineas bifariam
secat. Huius axis extremum in conica sectione punctum, vertex dicitur
Postulatum Primum
conicae sectionis. Ad axem vero sectionis ad rectos angulos deductae
40 Si Coni verticem atque signum in basis circumferentia aut alibi in conica 75 rectae lineae structim seu ordinatim actae vel deductae a Mathematicis
superficie utcumque susceptum recta coniungit linea, ea in conica existit vocantur, nonnunquam quoque eas secundum ordinem ductas Mathe­
superficie. Eandem rectam lineam plerique Mathematici coni latus ap­ matici nuncupant. Quae denique ex his ordinatim seu structim ductis
pellant. aequalis fuerit axis portioni, apud se terminatae, latus rectum, aliquando
etiam recta linea ad quam structim actae possunt a Geometris solita est
Postulatum Secundum 80 appellari.
45 In conica superficie recta linea duos praeter fastigium punctos con-
ELEMENTUM CONICUM PRIMUM.
nectens intra conum cadit.
Dati rectanguli trianguli isoscelis recto angulo subtensam ita dividere ut a
Postulatum tertium puncto divisionis alteri lateri acta sit inter subtensae segmenta media pro­
portionalis.
Si planum per coni culmen seu verticem secet conum, communis sectio
5 Esto triangulus rectangulus et isoscelis a b c [Fig. 6.2] rectum habens
conicae superficiei et secantis plani triangulus rectilineus existit.
angulum/? a c, cuius subtensa/) c cui dematur tertium d c; similiter ex a b
a iiii V latere ter/tium unum quod sit a e auferatur; connexisque d e dico quod
50 DIFFINITIO QUINTA
subtensa/) c divisa sit superi/, a quo ipsi a c lateri parallelus d e acta media
Ab axe coni triangulus est facta in conica superficie communis sectio est proportionalis inter/) d, d c segmenta ipsius b c subtensae. Et quoniam
quando planum secat conum super axe, talis autem communis sectio 10 per propositionem xlvii. libri primi elementorum Eu. quadratus ipsius b d
conicae superficiei atque plani secantis rectilineus existit triangulus per duplus est quadrati d e , igitur ratio quadrati b d ad ipsius d e quadratum est
tertium postulatum. ut b d ad d c , per constructionem autem b d dupla est ipsius c d. Igitur per
corolarium secundum propositionis xx. Ii. vi. ele. — Si tres lineae propor­
55 Postulatum quartum tionales fuerint, erit sicut prima ad tertiam sic quae a prima fit species ad
15 eam quae a secunda similis et similiter descripta— tres rectae lineae b d , d e,
Si planum coni basi parallelum conum secuerit communis sectio plani
d c sunt continue proportionales. Igitur dati rectanguli trianguli et reliqua
secantis atque conicae superficiei circulus est.
ut supra quod oportebat efficere. At a c , d e paralellas (!) esse liquet ex pro­
positione ii. lib. vi. el. Est enim ut a e ad e b sic c d ad d b.
/ Postulatum quintum
Si planum secans conum per eius verticem non venerit neque basi Corolarium.
60 parallelum extiterit, communis sectio eiusdem plani et conicae superficiei
inflexa quaedam est linea quam Mathematici conicam vocant sectionem. 20 Inde etiam perspicuum est quod tres rectae lineae b d, d e, d c propor­
tionales sunt iuxta rationem dimetientis quadrati ad suam costam, idest
Divisio tertia secundum rationem potentia duplam.

Conicarum sectionum alia parabole, alia hyperbole, alia Ellipsis. Para­

ELEMENTUM CONICUM SECUNDUM.
bole est quando planum secans conum ad planum trianguli ab axe erigitur,
65 horumque planorum communis sectio ad reliquum huius ab axe trianguli In subtensa trianguli rectanguli isoscelis dato puncto qui auferat partem
latus parallelum fuerit. Hyperbole autem est quando dictorum planorum maiorem aut minorem tertio uno ipsius subtensae, atque ab eodem puncto
communis sectio cum reliquo ab axe trianguli latere producto ultra coni alteri laterum parallelus si fuerit acta, eam alia ipsi subtensae parallela
272 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 273

5 sic secare quod superior eius portio sit media proportionalis inter secundae circa rectum angulum latera in easdem partes producta, erit rectangulum sub
parallelae segmenta. sectionibus secundae paralleli factum aequale ei rectangulo quod fit a prima
In triangulo itaque rectangulo a b c et isosceli ex subtensa b c punctus parallela producta in eius particulam quae inter basis segmenta media existit
d aufferat c d [Fig. 6.3]. In primis minorem tertio uno totius b c, et per d proportionalis.
ipsi a c paralellus (/) agatur d e , secans latus a b super e , atque ex b e dematur io Sit igitur ut prius triangulus rectangulus isoscelis a b c [Fig. 6.5], circa
io e f dupla ipsius a e, ipsique b c subtensae parallelus acta f g secet a c rectum angulum a b c duo latera α b ,b c habens aequalia, et in subtensa a c
quidem super g, d e vero super h; et quia a g, e h paralleli sunt per con­ sit d signatus punctus per quem ipsi b c acta parallelus d e sit inter a c basis
structionem, igitur per secundam propositionem li. vi. ele. / h ad h g est ut segmenta a d , d c media proportionalis. Atque d e in partes d producta in
e f ad a e . At ex hypothesi e f dupla est ipsius a e ; igitur/ h dupla est ipsius rectum quoad libet usque ad /; atque per/ ipsi subtensae a c parallelus sit
a vr g h. Sed per xlvii. propositionem li. i. el. / quadratus ipsius f h duplus est 15 acta g f h, secans duo circa rectum a b c angulum latera a b , b c in easdem
15 quadrati e h, nam e f , e h sunt aequales et angulus f e h rectus. Igitur ratio partes producta super g, h signis, e d vero in easdem partes eiectam in/ .
ipsius f h ad g h est ut ratio ipsius f h ad h e duplicata; ergo tres rectae Dico quod rectangulum sub g h f factum sit aequale ei quod fit sub f e, e d
lin eae/h,e h,g h sunt continue proportionales per corolarium praecedentis rectangulo. Et quia quae est ratio ipsiusf g ad e f eadem est etiam ratio ipsius
elementi. In subtensa igitur trianguli et reliqua ut supra. Sin autem d c d e a d d c, est autem c d aequalis ipsi / h, per corolarium primi huius ele-
maior extiterit tertio uno totius subtensae b c, ergo f g cadet infra sub- 20 menti, utrobique enim est ratio potentia dupla, igitur per proposi. xvi. li. vi.
20 tensam b d c, atque inde propositum eodem modo ut ante conficietur. ele. Eu.— Si quattuor rectae lineae proportionales et cetera— rectangulum
sub g f , f h factum est aequale ei quod s u b /e , e d fit rectangulo. Si igitur in
ELEMENTUM CONICUM III. isoscelis rectanguli trianguli basi seu subtensa punctus fuerit assignatus a
quo et reliqua ut supra quod oportebat demonstrare.
Si in trianguli isoscelis rectanguli basi punctus signetur a quo alteri circa
rectum angulum lateri parallelus agatur inter basis segmenta media existens
ELEMENTUM CONICUM V.
proportionalis, deinde ipsi basi alia utcunque acta fuerit parallelus priorem
5 secans parallelum atque trianguli latera reliqua, erit quod fit sub segmentis Si a parabola recti rectangulique coni ad ipsius paraboles axem ordinatim
secundae paralleli rectangulum aequale ei rectangulo quod fit sub prima [a vi r] acta ceciderit in rectum parabolae latus, erit quadratus / eiusdem ordinatim
parallela atque illa ipsius particula quae lateri iuxta rectum angulum atque actae aequalis ei rectangulo quod fit sub recto latere eiusque assumpto inter
secundae adiacet parallelo. 5 paraboles fastigium atque ordinatim actam segmento.
Sit triangulus rectangulus isoscelis a b c [Fig. 6.4], rectum habens an- Sit ergo in recto rectanguloque cono a b c d [Fig. 6.6], cuius vertex b, basis
10 gulum a b c atque circa eundem angulum a b c duo aequalia latera a b ,b c a c d, parabole d e f , huius axis/ g , vertex/ , sitque/ g latus rectum eiusdem
et basim a c , in qua a puncto d ipsi b c acta sit parallelus d e , secans a b latus paraboles d e f. Itaque d g structim acta sit aequalis ipsi f g. Atque in para­
super e , fueritque d e inter a d ,d c basis a c segmenta media proportionalis. bola d e f datum sit e , signum a quo ad f g latus rectum structim deducatur
Rursus ipsi a c basi sit acta parallelus f g h, secans a b super/et parallelum io e h. Dico quod quadratus ipsius e h ordinatim actae est aequale ei rectangulo
d e in g et latus b c super h. Dico quod rectangulum factum sub f g, g h sit quod fit sub g f , f h . Planum igitur aliquod ipsi a c d basi parallelum secet
15 aequale ipsi rectangulo sub d e, e g facto. Quoniam autem ratio ipsius f g ad super e h conum a b c d ; igitur per quartum postulatum huius plani secantis
g e est sicut e d ad d c sive ad aequalem g h per corolarium primi elementi et conicae superficiei communis sectio i e k erit circulus. Et sit ab axe coni
conici, utrobique enim est ratio potentia dupla, igitur, per propositionem xvi. triangulus a b c, qui circulum e k bifariam secat; atque eorundem trianguli
li. vi. el. Eu. — Si quattuor rectae lineae proportionales fuerint et cetera— 15 ab axe coni et circuli e i k communis sectio i k diameter est eiusdem circuli
quod fit s u b / g, g h rectangulum aequale est ei quod fit sub d e, e g i e k, necessarioque meabit per h signum. Et quia a d c quoque semicirculus
20 rectangulo. Si igitur in trianguli isoscelis rectanguli basi punctus signetur est, et per constructionem d g ad rectos angulos est ipsi a c diametro eius­
et reliqua ut supra quod oportuit demonstrare. dem semicirculi a d c, atque d g aequalis ipsi/ g ex hypothesi et sexta diffini­
tione huius, et quia per propositionem xxi. li. iii. el. Eu. angulus a d c rectus,
av v / ELEMENTUM CONICUM IIII. 20 et d g ipsi a c ad rectos est angulos: igitur, per corolarium propositionis viii.
li. vi. eorundem ele. Eu., d g inter a g , g c media existit proportionalis. Est
Si in rectanguli trianguli isoscelis subtensa punctus fuerit assignatus, a quo autem per constructionem d g aequalis ipsi / g ; igitur/ g media proportionalis
alteri circa rectum angulum lateri parallelus acta sit media existens propor­ existit inter a g, g c. Et i h k parallelus ipsi a c basi trianguli a b c ab axe
tionalis inter basis segmenta, in qua ut supra in partes subtensae producta, coni. Ergo per tertium elementum conicum quod fit sub g f , f h rectangulum
5 puncto assumpto ab eo ipsi basi seu subtensae parallelus agatur duo secans 25 aequale est ei quod fit sub i h k rectangulo. Est autem i e k circumferentia
274 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 275

semicirculi velut patuit, et e h per diffinitionem vi ad rectos angulos ipsi i h k, ELEMENTUM CONICUM VII.
et per propositionem xxxi. li. iii. ele. Eu. i e k angulus rectus. Ergo per
A recti rectangulique coni parabola quarumlibet duarum structim
corolarium propositionis viii. li. vi. ele. Eu. e h media est proportionalis
actarum quadratis ratio est ut ratio earum quae sunt ipsis conterminae axis
[a vi v] inter / h , h k. Igitur per propositionem / xvii. li. eiusdem vi. el. Eu. quadratus
portionum.
30 ipsius e h est aequalis ei quod fit sub i h, h k rectangulo, quod per iam
5 Sit igitur recti rectangulique coni parabola α b c d [Fig. 6.8], cuius axis
ostensa aequale est ei quod fit sub g / , / h rectangulo. Quadratus igitur ipsius
α e f g, in quem structim actae sint b e , d g. Dico igitur quod ratio quadrati
e h aequalis est ei quod fit sub g / , / h rectangulo. At per constructionem/g
b e add g quadratum sit sicut e a a d a g . Sit ergo latus rectum a f paraboles
axis latus est rectum parabolae d e f , atque ad idem latus rectum ab e , signo
a b c d. Et quia per quintum aut sextum conicum elementum quadratus
paraboles d e f, e h structim deducitur. Ergo si a parabola recti rectangu-
[a vii v] ipsius b e aequalis est ei quod fit sub f a e rectangulo, si/militer quadratus
35 lique coni ad ipsius paraboles axem structim acta ceciderit et reliqua ut supra
10 ipsius d g aequatur ei quae fit sub g a, a f areolae rectangulae, sed per
quod oportebat demonstrare.
primam propositionem li. vi. el. Eu. rectangulum sub f a, a e ad
ELEMENTUM CONICUM VI. rectangulam areolam sub g a f rationem habet quam a e ad a g haec
namque rectangula sub eadem sunt altitudine quae est a / , latus rectum
Si a recti rectangulique coni parabola sit ad ipsius paraboles axem struc­ paraboles per constructionem seu ex hypothesi. Ergo eisdem rectangulis
tim deducta cadens extra latus rectum, erit quadratum structim deductae ei 15 aequalia quadrata ipsarum b e, d g rationem habent quam a e ad a g . Igitur
aequale rectangulo quod fit sub recto latere atque ea axis portione quae a recti rectangulique coni parabola et reliqua ut supra quod oportebat
5 structim deductae atque paraboles fastigio adiacet. demonstrare.
Sit igitur in rectangulo rectoque cono a b c d [Fig. 6.7] parabole d e / ,
cuius a x is/ g et a d signo structim acta sit d g, secans axem paraboles super ELEMENTUM CONICUM VIII.
g signo. Et latus rectum eiusdem parabolae sit e h, quod per diffinitionem
Data quae in cono recto et rectangulo fit parabola eiusque axe dato, para­
sextam conterminae axis particulae / h semper est aequale. Hic quoque
boles rectum latus dare.
10 obiter est notandum quod apud priscos geometras utraque duarum e h , f g
Sit ergo data recti rectangulique coni parabole a b c [Fig. 6.9], cuius axis
( / / h) recti lateris appellationem invenitur habere, ea, potissimum ut puto,
5 a d , vertex a , et ad axem a d atque ad datum in eo signum a , per proposi­
ratione quia sunt aequales. Sed iam a digressione hac revertendo ad institu­
tionem xxxiii. li. i. ele. Eu., constituatur angulus rectilineus d a c recti
tum, sit ab axe coni triangulus a b c. Dico itaque quod quadratus ipsius d g
dimidio aequalis, et α c secet parabolen a b c in c puncto, a quo ad axem
sit aequalis rectangulo quod fit sub g f , f h . Intelligamus itaque per e h
a d ordinatim agatur c d. Et quia in triangulo a c d angulus a d c rectus est
[a vii r] aliquod planum basi ipsius conia b e d / parallelum secare eundem conum.
et c a d recti dimidius ex hypothesi, ergo per propositionem xxxii. li. i. ele.
16 Huius igitur plani et trianguli a b c ab axe coni communis sectio i h k paral­
10 Eu. angulus quoque a c d recti dimidius est, quoniam trianguli rectilinei
lela est a c subtensae rectanguli trianguli a b c. Eiusdem denique plani
tres interiores anguli duobus sunt rectis aequales. Isoscelis (/) ergo est tri­
atque inflexae seu conicae superficiei communis sectio est i e k circulus per
angulus a e d per propositionem vi. eiusdem primi lib. ele. Eu. Et quia c d,
quartum postulatum. Igitur per propositionem xxxi. lib. iii. elemen. Eu.
structim deducta ad a d axem, aequalis existit ipsi a d, per diffinitionem
20 angulus / e k rectus est. Et quia per diffinitionem sextam e h ipsi i h k ad
igitur sextam seu recti in parabola lateris c d seu aequalis a d rectum est latus
rectos existit angulos, ergo per corolarium octavae propositionis li. vi. ele.
15 datae paraboles a b c. Data igitur quae in cono recto et rectangulo fit para­
Eu. e h media est proportionalis inter i h, h k. At e h per diffinitionem recti
bola et reliqua ut supra quod oportebat efficere.
lateris aequalis est ip s i/ h. Ergo/ h media proportionalis est inter / h, h k.
Igitur per quartum conicum elementum quod fit sub α g, g c rectangulum
ELEMENTUM CONICUM IX.
25 aequale est ei quod fit sub g / , / h rectangulo. Sed ei quod fit sub α g, g c
rectangulo, aequalis est quadratus ipsius d g. Est enim α d c circumferen­ Data recti rectangulique coni parabola eiusque axe dato, ipsum conum
tia semicirculi et per propositionem xxxi. li. iii. ele. Eu. angulus α d c rectum et rectangulum dare.
rectus est et per constructionem d g ipsi α c ad rectos angulos. Igitur Sit ergo data recti rectangulique coni parabola α b [Fig. 6.10], cuius axis
quadratus ipsius d g aequalis est ei quod fit sub g f axe paraboles et recto 5 a c ; atque per elementum conicum octavum rectum detur parabolae α b
30 latere f h . Sed ordinatim acta d g per constructionem et hypothesim in [a vi» r] latus, sitque b c; atque per propo/sitionem xxiii. li. i. ele. Eu. ipsi α c axi
axem paraboles d e f cadit extra rectum latus f h . Igitur si a recti ad α signum rectus construatur angulus c a d, cuius planum rectum sit ad
rectangulique coni parabola ad eius axem structim acta ceciderit extra latus ipsius paraboles a b c planum. Sit etiam α d recta linea aequalis ipsi α c , et
rectum erit quadratum, et reliqua ut supra quod oportuit demonstrasse. connexa c d ipsa α d bifariam secetur in e , et α d in partes α producatur
276 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 277

10 usque a d / sic quod a f aequalis sit ipsi a e; atque ad a c parallelus agaturf g, erit sicut d e ad e a . At ex hypothesi in hac proportione tribus datis terminis,
et d c in partem c producta incidat in/ g super g signo. Et quoniam per con­ io quartus terminus a e axis portio datur. Totus igitur axis a e d paraboles
structionem/# actaestada c, igitur per secundam propositionem li. vi. el. a b c datur. Ergo a parabole recti rectangulique coni duabus ordinatim
Eu. d c ad c g est ut d a ad a / . Est autem d a dupla ipsius a /; igitur et d c deductis datis, et reliqua ut supra quod oportuit demonstrare.
dupla est ipsius c g. Et quia per propositionem xlvii. lib. i. ele. Euc. quadra­
15 tum ipsius d c duplum est ad ipsius a c quadratum, ergo ratio ipsius d c b ir / ELEMENTUM CONICUM XI.
quadrati ad ipsius a c quadratum est ut d c ad c g. Et quoniam per corolarium
Data recta linea ad quam structim actae in parabola possunt, ipsam de­
propositionis xx. lib. vi. ele. Eu. — Similes rectilineae figurae ad invicem
scribere in dato plano parabolam. Recta linea ad quam ordinatim deductae
in dupla sunt ratione similis rationis laterum— igitur ratio d c ad c g dupla
possunt alio nomine latus parabolae rectum dicitur, per diffinitionem
est rationis ipsius d c ad c a. Igitur per diffinitionem duplicatae rationis
5 sextam. Pari ratione parabole ibi quae ad describendum proponitur ea est
20 a c, seu aequalis b c, media proportionalis existit inter d c et c g. Et quia
quae in rectum rectangulumque conum incidit.
per constructionem angulus b e d rectus est, atque per corolarium proposi­
Ad quam igitur structim actae possunt recta et data linea sit a b [Fig.
tionis viii. li. vi. el. Eu. angulus d b g rectus est per imaginationem videlicet
6.12]. Et ipsa a b in quotlibet aequales secetur partes a c, c d, d b, atque
connexis d b, b g, igitur per propositionem xxxi. li. iii. ele. d g dividue
earum cuilibet aequalis ad a b in directum adiiciatur b e , atque in b signo
secta in h , et centro h , intervallo autem d h , scriptus circulus d b g ibit per
io ipsi a e ad rectos excitetur angulos b / , quae in partes / in infinitum sit
25 b signum. Coniuncta deinde f h, et quia angulus ad / rectus est, atque d f
producta; atque ipsa a e dividue secta super d , et d centro, spacio autem
aequalis ipsi / g , ergo et anguli ad h recti sunt, et uterque duorum an­
a d, semicirculus scribatur a f e, secans b f perpendicularem in / signo.
gulorum qui super/ g sunt basi rectanguli trianguli f h g dimidius est recti.
Rursus ipsi b e aequalis adiungatur e g , sitque tota a e g, quae iterum
Igitur partilis trianguli f h g fixo f h, latere atque eodem particulari trian­
bifariam secetur h signo, quo centro, atque intervallo a h, rursus semi-
gulo f h g circumducto donec eo redeat unde moveri coeperat, ipse conum
15 circulus scribatur a i g, dispescens perpendicularem b f super i signo.
30 describet rectum et rectangulum per primam diffinitionem, cuius quidem
Atque in hunc modum quotlibet ipsi e g aequales in directum adiiciantur
coni basis est circulus d b g . Ipse denique triangulus f g h circumactione
atque semicirculi scribantur secantes perpendicularem b f in singulis
[a viii v] sua paulatim seu successive / secabit parabolen a b in omnibus suis punc­
biv punctis ultra/ punctum. / Praeterea alia quaedam recta linea in subiecto
tis. Quod autem circulus d b g basis existat coni circumactione f g h tri­
assumatur plano k l m n, aequalis ipsi a b existens, sitque in partem n
anguli confectus liquet ex eo: quoniam circulus d b g erectus est ad planum
20 infinita, habens partes k 1,1 m, m n, numero et magnitudine aequales ipsis
35 trianguli d f g, quia per constructionem planum a b c erectum est ad
a c , c d, d b partibus, et sic deinceps in infinitum quotlibet assumptis
planum trianguli d f g, ex hypothesi autem angulus a c b rectus est, igitur
partibus aequalibus. Atque per signa /, m , n, et alia deinde comitantia signa
b c recta linea erigitur ad planum trianguli d f g per diffinitionem tertiam
ipsi k l m n rectae perpendiculares agantur, quae in utramque partem ip­
li. xi. el. Eu.— planum ad planum rectum est et cetera. At planum circuli
sius k l m n sint quoque in infinitum productae, et perpendiculari per /
d b g transit super b c rectam. Igitur per propositionem xviii eiusdem li.
25 actae utrinque duae aequales ipsi b f auferantur sintque / o, l p. Sic quoque
40 xi. ele. planum circuli d b g rectum est ad planum trianguli d f g. Ergo idem
perpendiculari per m ductae duae rectae demantur aequales, hinc quidem
circulus d b g basis est coni quem partilis triangulus f g h sua circumac­
m q , inde vero m r, sic ut utraque ipsarum m q , m r sit aequalis ipsi b i per­
tione descripsit. Igitur f d b g conus datus est et rectus rectangulusque in
pendiculari. Id fiat quousque libeat, et eisdem punctis, velut o, q , p , r , ex
cuius conica superficie parabole a b data describitur. Data ergo coni
utraque parte ipsius k l m n per rectas lineas iugatis, descriptam esse quae
rectanguli parabola, datur et conus in cuius conica superficie eadem para­
30 proponitur parabolam sic constabit. Esto igitur/: l m n recta linea aequalis
45 bole describitur, quod demonstrasse oportuit.
ipsi a b lateri recto dato. Et quia per constructionem a f e semicirculus
et per propositionem xxxi. li. iii. ele. Eu. angulus a f e rectus est, imagina­
ELEMENTUM CONICUM X.
tis videlicet seu per imaginationem ductis duabus rectis lineis a f , f e, igitur
A parabola recti rectangulique coni duabus structim deductis datis atque per corolarium propositionis viii. li. vi. eorumdem ele. b f est media pro-
inter ipsas axis segmento dato, totus eiusdem paraboles axis dabitur. 35 portionalis inter a b et b e . Est autem k n aequalis ipsi a b recto lateri dato,
Sit ergo data parabole a b c [Fig. 6.11], a qua structim deductae b e, c d et / o aequalis b f perpendiculari, et k / ipsi b e aequalis. Igitur utraque
atque inter eas axis segmentum d e dentur. Dico quod totus paraboles axis ipsarum l o , l p media proportionalis existit inter k n et k /. Igitur per pro­
a e d detur. Et quoniam per septimum elementum conicum ratio quadrati positionem xvii. lib. vi. ele. quod fit sub k n , k l parallelogrammum rectan­
c d ad quadratum ipsius b e est sicut d a ad a e, dirimendo igitur excessus gulum aequale est ei quod fit ex / o aut ex / p quadrato. Per similem denique
quadrati ipsius c d super ipsius b e quadrato ad eiusdem b e quadratum 40 argumentationem probabitur quod fit sub k n, k m rectangulum esse
278 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 279

aequale quadrato ipsius m q seu aequalis rectae lineae m r. Haud aliter totam. Tres igitur rectae lineae c d, a c , c e continue sunt proportionales
idem ostendemus de reliquis perpendicularibus per puncta ipsius k n secundum rationem ipsius a dada e, atque ita corolarium existit manifestum.
productae actis. Igitur per conversionem quinti aut sexti elementi conici
inflexa linea q o k p r conica est sectio quae parabole dicitur quaeque in ELEMENTUM CONICUM XIII.
45 conum cadit rectum rectangulumque qualem dato recto latere nonum Si ab aliquo puncto extra datum circulum suscepto ad eundem circulum
elementum conicum docet construere. Data igitur recta linea a b, seu duae deducantur rectae lineae, altera ad centrum, altera circulum tangens,
aequali k n, ad quam structim actae l o, l p, m q, m r, et reliquae structim et a contactu supra ad centrum deductam perpendicularis agatur, atque a
actae possunt, in dato plano descripta est parabola<7 o k p r, quod oportuit 5 puncto in circumferentia eiusdem circuli utcumque assumpto duae rectae
b ii r effecisse. Obiter notandum est quod quanto partes ipsius k n / recti lateris coniungantur lineae, altera quidem ad dictum punctum extra circulum, al­
50 arctiores assumuntur, tanto verius proposita parabole scribetur. Verum tera vero ad terminum dictae perpendicularis: erit earundem a circum­
quaevis recta linea iugans proxima quaeque duo puncta ab inflexa para­ ferentia dati circuli deductarum ratio ut rectae lineae quae in deducta ad
boles particula quae eisdem finitur punctis parum admodum ac penitus centrum circuli assumpto extra puncto et circulo adiacet ad eam rectam
insensibiliter differt. Praeterea notandum est quod paraboles portio quae 10 quae eodem circulo atque praedicta perpendiculari comprehenditur.
fastigio proxima existit modico ac pene insensibili distat intervallo a cir- Manentibus itaque eisdem sublectionibus et figuratione praecedentis
55 cumferentia eius circuli qui scribitur super centro puncto existente mediae elementi, in circumferentia circuli a b [Fig. 6.14] suscipiatur utcunque/sig­
divisionis lateris recti atque spacio medietatis eiusdem recti lateris, velut id num a quo connectantur d f , e f . Dico quod ratio ipsius d f ad/ e sit sicut
manifeste liquet ex subiecta descriptione. a d ad a e. Connectatur ergo c / ; et quia in duobus triangulis c d f , c e f
15 latera circum communem angulum e c f sunt proportionalia, nam per
ELEMENTUM CONICUM XII. praecedens elementum ut d c ad c f sic c f ad c e, igitur duo triangula
c d / , c e f sunt aequiangula per propositionem vi. Ii. vi. ele. Eu.— et anguli
Si ab aliquo puncto extra datum circulum suscepto duae rectae agantur aequales quibus proportionalia subtenduntur latera. Igitur ut c f ad c e sic
lineae, altera quidem ad centrum, altera vero tangens eundem circulum, d f ad e f , sed ut c / ad c e sic α d ad α e per Corolarium praecedentis ele-
atque ab ipso contactu supra deductam ad centrum perpendicularis agatur: b iii r menti conici. / Ut igitur α d ad α e sic d f ad e f. Similiter probabitur quod a
5 erunt deducta ad centrum et semidiameter circuli atque ad centrum de­ 21 quocumque alio praeter / puncto connexae ad d, e duae rectae lineae
ductae particula quae centro et perpendiculari adiacet continue propor­ rationem habeant quam α d ad α e. Si igitur ab aliquo puncto extra datum
tionalis (/). circulum suscepto et reliqua quod oportuit demonstrasse.
Esto itaque datus circulus a b [Fig. 6.13], cuius centrum c, atque extra
circulum a b susceptus utcumque punctus d, a quo ad c centrum quidem
Corolarium.
10 connectatur recta linea c d , secans circulum a b super a , tangens autem
a b circulum super b signo sit acta b d. Rursus a d (/ b) contactu super c d 25 Hinc etiam perspicuum fit quod coniuncta α f bifariam secat angulum
perpendicularis sit b e. Dico quod c d deducta ad centrum c, et semi- e f d, ita ut angulus α f d sit aequalis angulo α f e. Nam ut ostensum fuit
diameter a c circuli a b, atque c e sint continue proportionales. Coniun- ut d f ad e f sic α d ad α e, igitur per secundam partem tertiae proposi­
gantur itaque b c. Et quoniam per propositionem xvii. Ii. iii. ele. angulus tionis li. vi. ele. Eu. recta α f dividue secat e f d angulum. Idem constabit
15 c b d rectus est, atque anguli ad e signum recti per diffinitionem perpen­ de omni angulo facto, si ab aliquo signo in circumferentia α b sumpto ad
dicularis, erit igitur ut c d ad b c sic b c ad c e. Ipsa autem a c aequalis est 30 d, e signa duae rectae lineae connectantur. Ergo Corolarium manifestum est.
ipsi b c, utraque enim ex centro circuli a b. Tres igitur rectae lineae c d
deducta ad c centrum circuli a b, atque eiusdem circuli [semi]diameter ELEMENTUM CONICUM XIIII.
a c, et e c sunt continue proportionales. Si igitur ab aliquo puncto extra In dato cono per lineam rectam a coni vertice actam ad basim planum
20 datum circulum et reliqua ut supra quod oportuit demonstrare. agere tangens eundem conum.
Sit ergo conus α b c d [Fig. 6.15], cuius basis circulus α e d (! α c d).
Corolarium. 5 Sitque coni α b c d vertex b , a quo ad basim α c d ad eius circumferentiam
b» v / Hinc etiam patet quod tres rectae lineae c d, a c , c e sint continue a e d i n signo d connexa sit b d, linea quae per primum postulatum recta
proportionales iuxta rationem ipsius a d ad a e. Nam per propositionem est, existens in conica superficie coni a b c d, propositumque sit per rec­
xix. Ii. v . ele. Eu. sicut tota c d ad a c totam sic ex c d ablata a c ad c e sub- tam lineam b d planum agere tangens conicam superficiem coni a b c d
25 latam ex a c. Igitur reliqua a d ad a e reliquam est sicut tota c d ad a c super recta linea b d. Igitur per d signum recta linea agatur d e tangens cir-
280 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 281

10 cumferentiam a c d super d. Perspicuum itaque est quod planum in quo rectum rectangulumque, igitur et idem conus datur per nonum elementum
duae rectae lineae b d, d e deducuntur tanget conicam superficiem coni conicum. Detur itaque et s i t / c g h, cuius vertex h , axis d h, basis/ c g.
a b c d super recta linea/? d. Ipsae namque b d , d e rectae lineae in eodem Et ab axe coni triangulus sit f g h, secans planum a c d super paraboles
consistunt plano per propositionem li. xi. el. Eu. Igitur in dato cono per 15 axe a d ad rectos angulos. Erit itaque f c g circumferentia semicirculi.
lineam rectam ad circumferentiam basis a coni vertice actam b d deducitur Connexa denique c h, erit super inflexa superficie coni f c g h. Sit autem
15 planum b d e, tangens inflexam superficiem coni a b c d super b d recta primum c g periferia dimidium semicirculi f c g, seu quadrans totius cir­
linea, quod oportuit efficere. cumferentiae circuli, uti subiecta habet figuratio. Igitur a signo c ordinatim
deducta cadit in d, centrum circuli f c g. Erit itaque planum trianguli c d h,
Corolarium. 20 erectum ad duo plana, trianguli vid elicet/g h ab axe coni et circuli/ c g
per propositionem xviii. li. xi. ele. Eu. Et per c h rectam lineam planum
b iii v Hinc quoque liquet quod si planum aliquod tangat subiectam coni/cam agatur c e h, tangens f c g h conum super e h (! c h) recta per ele. conicum
superficiem super data recta linea, idem planum tanget etiam omnem in-
xiiii; denique plani circuli f c g atque co n u m /c g h tangentis plani c e h
20 flexam lineam quae in conica superficie describitur, secans dictam rectam
communis sectio sit c i, quae etiam erigitur ad planum trianguli c d h. Nam
lineam, et erit communis sectio duarum linearum rectae et inflexae signum
25 c i per constructionem tangit circulum/ c g in c signo et c d deducta est a
contactus.
centro d ad c, contactum paraboles, per propositionem xviii. li. iii. ele. Eu.
ELEMENTUM CONICUM XV. Ex hypothesi denique rectus est angulus c d g \ paralleli igitur sunt c i , d g.
Igitur sicut d g erigitur ad planum trianguli c d h, sic c /. Ergo per proposi­
Ad inflexam lineam datae conicae cuiusvis sectionis in dato puncto tionem xviii. li. xi. ele. Eu. conum f c g h tangens planum erigitur ad
tangentem rectam lineam agere. b ϋϋ v planum trianguli c d h. Et quia planum trianguli/g h ab axe coni / velut
Sit igitur in dato cono a b c d [Fig. 6.16], cuius basis circulus a c d, 31 patuit etiam erigitur ad trianguli c d h planum, igitur eiusdem trianguli
5 conica quaedam sectio qualiscumque f g h , atque super inflexa eiusdem f g h ab axe coni atque plani c e h communis sectio e h ad planum eiusdem
sectionis linea f g h datum punctum h , per quem oportet agere rectam trianguli c d h erigitur per xix. propositionem eiusdem li. xi. ele. Eu. Paral­
lineam quae tangat eandem inflexam lineamf g hi n h signo. Igitur a vertice leli ergo sunt c i, e h, per propositionem vi. eiusdem li. xi. ele. Eu. Atqui
b conia b c d per idem punctum h producatur recta linea b h d, secans coni 35 c i parallelus quoque existit ipsi d g uti patuit; ergo et e h parallelus est
basim in eius circumferentia super d signo et per b h d conicae superficiei eidem d g. Et quoniam c e per constructionem tangit parabolen a b c in c
10 conia b e d planum/? d e applicetur, tangens eandem conicam superficiem signo et e signum commune est plano e c h et plano f g h trianguli ab axe
super b h d recta linea per xiiii elementum conicum. Atque plani in quo coni, igitur e signum necessario constituitur in communi sectione e h ip­
consistit sectio f g h atque plani b d e communis sectio sit hk. Et quia recta sorum planorum f g h et c h e. Atqui velut patuit e h parallelus est ipsi
h k sectio communis eorundem planorum, solum h punctum commune 40 d g·, igitur per propositionem xxix. li. i. ele. Eu. angulus a h e aequalis est
habet cum inflexa linea f g h. Igitur per diffinitionem et per corolarium a g d angulo; igitur etiam aequalis angulo a h d. Anguli denique ad a recti
15 praecedentis elementi conici h k recta tanget inflexam lineam f g h super h sunt, atque duobus triangulis a e h, a d h commune latus a h. Igitur
signo. Igitur ad inflexam lineam conicae sectionis f g h per signum h datum duo triangula a d h, a e h sunt aequilatera per propositionem xxvi. lib. i.
tangens recta linea acta est, quod oportuit efficere. ele. Eu. Et quia latera sunt aequalia quae aequalibus subtenduntur angulis,
45 ergo a d aequalis est ipsi a e . Igitur si in dato signo quae in rectum rectan­
ELEMENTUM CONICUM XVI. gulumque cadit conum parabolen et reliqua ut supra quod oportuit demon­
Si in dato signo quae in rectum rectangulumque cadit conum parabolen strare.
recta linea tangat, eademque tangens et paraboles axis in easdem produ­ Sin autem c g circumferentia quadrante minor extiterit [Fig. 6.18],
cantur partes quousque concurrant, erit producti axis pars exterior quae igitur planum tangens conum f c g h super h c recta linea coincidet cum
5 paraboles vertici dictique concursus signo adiacet aequalis ei axis portioni 50 dimetiente f d g in partes g producto. Coincidat itaque super / signo. Con-
b ϋϋ r quae eidem paraboles vertici et / a contactu structim actae interponitur. nexaque h i necessario meabit per e signum. Nam dicti plani quod tangit
Sit igitur data parabole a b c qualis proponitur cuius axis a d [Fig. conum super h c recta atque plani ab axe trianguli f g h communis sectio
6.17]. Et ipsam parabolen a b c tangat recta linea e c super c signo. E tc e est h i recta. Igitur e signum necessario constituitur super h i recta linea.
tangens axisque a d producti in partes a coincidant super e signo, atque Coniungatur deinde c d structim acta, quae ex hypothesi atque per con-
10 ex contactu c super paraboles axem a d structim acta sit c d. Dico quod a d 55 structionem ad rectos est angulos ipsi/ d g dimetienti. Et si ab axe coni
sit aequalis ipsi a e. Et quia ex hypothesi parabole data a b c cadit in conum triangulum/ h g subiectum fuerit in eodem plano circuli f c g, erit h punc-
282 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 283

tus super circumferentia f c g propterea quod angulus f h g rectus est et ELEMENTUM CONICUM XVII.
bv r f g subtensa dimetiens / circuli f c g . Connectatur denique d g. Et quia c /
tangit circulum /c g super c, et ex c ad/ g i per centrum circuli actam ad Si in dato cono ab axe trianguli latus unum ultra coni verticem produca­
60 rectos angulos c d agitur, ergo per superius ostensa elementa conica duo tur, atque a termino eiusdem lateris producti ad basim ipsius ab axe trian­
anguli d h g et g h i sunt aequales. Et quia per constructionem anguli ad a guli recta quaedam linea ducatur, secans alterum eiusdem trianguli latus,
sunt recti et duobus triangulis a h d, a h e latus commune a h, igitur duo 5 et in eadem recta linea intra conum ex duobus contingentibus signis ad
trianguli a d h, a h e sunt aequianguli et aequilateri per propositionem planum ipsius ab axe trianguli duae excitentur perpendiculares conicae oc­
xxvi. lib. i. ele. Eu. Ergou d aequalis est ipsia e. Si igitur in dato puncto et currentes superficiei, erit ratio rectanguli facti sub eadem linea recta usque
65 reliqua ut supra. ad primam perpendicularem acta et eius portione quae eidem perpendicu­
At si c g [Fig. 6.19] circumferentia circuli quadrantem exuperet, ergo lari atque ab axe trianguli alteri lateri adiacet ad quadratum eiusdem per-
c i secabit dimetientem/ g in partes/productum. Rursus itaque perspicuum 10 pendicularis sicut ratio rectanguli compraehensi sub eadem recta linea
fit h i, d a rectas lineas in partes a, h {! h, a) productas coincidere invicem [b vi r] ducta usque ad secundam perpendicularem et eius / particula quae ad ean­
ad eundem e punctum, nisi quispiam velit assere quod eadem recta eandem dem secundam perpendicularem et latus ipsius ab axe trianguli terminatur
70 rectam lineam in duobus secet signis quod absurdissimum est et a perito ad quadratum eiusdem secundae perpendicularis.
geometra maxime alienum. Nam si sic, duae rectae lineae superficiem con­ Sit itaque conus a b c [Fig. 6.20], cuius basis b c circulus, et ab axe coni
cluderent quod neutiquam fieri potest repugnante communi geometrarum 15 triangulus a b c, eiusque latus ab in partem a verticis quantumlibet pro­
sententia. Suscipiatur itaque circuli f c g centrum et sit k. Coniungan- ducatur usque ad d signum, a quo recta linea agatur d e usque ad basim ab
turque h, k. Et quia per constructionem triangulum/ g h isoscele[s] est, axe coni trianguli b c, secans a c latus eiusdem trianguli super/; atque in
75 ergo anguli ad k recti sunt. Et quia h k, k g ex centro sunt circuli/ c g , ipsae e f linea recta duo utcunque puncti summantur g , h, a quibus ad planum
igitur sunt aequales. Igitur in triangulo h k g uterque duorum angulorum trianguli α b c duae excitentur perpendiculares g i, h k, conicae occur-
qui sunt ad basim g h recti dimidius est. Et per h signum ipsi/ g parallelus 20 rentes superficiei super i, k signis. Dico quod ratio rectanguli sub d g f
h / agatur. Et quiag / i et h l paralleli sunt, ergo per propositionem xxix. compraehensi ad quadratum perpendicularis g i est sicut ratio rectanguli
libri primi elementorum angulus e h l aequalis est/ i h angulo. Praeterea sub d h , h f contenti ad quadratum perpendicularis h k. Protrahantur ergo
80 per elemen. coni. xii. tres rectae lineae i k, k f , d k sunt continue propor­ per g , h signa ipsi b e c duae paralleli / g m , n h o, apud α b , a c latera tri­
tionales. Connectatur deinde d h; ergo ut i k ad k h sic k h ad d k. Et quia anguli a b c terminatae. Intelligantur deinde duo plana super / g m , n h o,
bv v duobus triangulis h i k, d k h commu/nis est angulus d k h rectus et cir­ 25 secantes conum a b c atque parallela ipsi b c basi. Igitur per postulatum
cum eundem angulum, velut ostensum est, proportionalia latera, igitur per quartum communes sectiones eorundem planorum atque conicae super­
propositionem vi. Ii. vi. elemen. Eu. duo trianguli h i k, k d h aequianguli ficiei erunt circuli / i m , n k o, quorum dimetientes l m , n o. Erit itaque
85 sunt, et anguli aequales quibus eiusdem rationis latera subtenduntur; ergo rectangulum sub l g , g m compraehensum aequale quadrato ipsius g i per­
angulus d h k aequalis est h i k angulo. Sed eidem h i k angulo iam ostensus pendicularis, atque rectangulum sub n h, h o compraehensum aequale
fuit aequalis c h / angulus; igitur angulus d h k aequalis est ipsi e h / angulo. 30 quadrato perpendicularis h k, per corolarium propositionis viii. libri vi.
Sunt auteml h g e t g h k anguli aequales, uterque enim aequalis est, velut elemen. Eu. Duo enim anguli sub l i m , n k o recti sunt per propositionem
patuit et per propositionem xxix. Ii. i. elemen. Eu., angulo g h k (/ h g k). xxxi. Ii. iii. eorundem ele. Et quoniam per propositionem xxiii. eiusdem li.
90 Igitur ex communi sententia— Si aequalibus addantur aequalia et cetera, [b vi v] vi. ele., rectangula parallelogramma / adinvicem rationem habent ex
compositi ex aequalibus duo anguli e h g, g h d sunt aequales. In duobus lateribus compositam, igitur ratio areolae contentae sub d g , g f ad areolam
autem triangulis α e h, α h d anguli qui ad α recti sunt et commune latus 35 compraehensam sub l g , g m componitur ex duabus rationibus quarum
α h. Igitur per propositionem, xxvi. Ii. i. elemen. Eu. duo triangula α h e et una est ipsius d g ad g /, altera ipsius f g ad g m. Similiter ratio rectanguli
α h d sunt aequiangula et aequilatera, necnon α d, a e latera quae aequalibus sub d h , h f facti ad rectangulum sub n h o compraehensum componitur ex
95 subtenduntur angulis aequalia. Ex hypothesi autem e c recta linea tangit ratione ipsius d h ad h n ct ratione ipsius f h ad h o . At eadem est ratio
parabolen a b c super c signo. Si igitur in dato signo, quae in rectum rec- ipsius d g ad g / quae ipsius d h ad h n. Sic quoque eadem est ratio ipsius
tangulumque conum cadit parabolen recta linea tangat, eademque tangens 40 f g ad g m quae ip siu s/h a d h o . Igitur ratio rectanguli sub d g/con tenti ad
et paraboles axis in easdem producantur partes quousque invicem concur­ rectangulum sub / g m comprehensum seu ad quadratum perpendicularis
rant, erit producti axis pars exterior quae paraboles vertici dictoque con- g i est sicut ratio rectanguli sub d h, h f facti ad rectangulum sub n h, h o
100 cursui adiacet aequalis ei axis portioni quae eidem vertici atque a paraboles contentum seu ad quadratum perpendicularis h k ex communi sententia:
contactu structim actae interponitur, quod hucusque oportuit demonstrasse. rationes eaedem sunt quae eisdem componuntur rationibus. Igitur si in
284 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 285

45 dato cono ab axe trianguli latus unum ultra coni verticem producatur et rectanguli a b c, lateri videlicet b c h quadrati, igitur e g, latus quadrati
reliqua ut supra quod oportuit demonstrare. e f g , erit aequale ipsi b h, lateri quadrati c b h, et tota d e g recta minor
15 erit quam ipsa a b h, ex communi sententia— Si aequalia inaequalibus ad­
Corolarium. dantur et cetera. Igitur aggregatum ex d e /rectangulo et quadrato/ e g
minus erit aggregato ex a b c rectangulo et quadrato c b h per proposi­
Hinc etiam fit perspicuum quod h k maior est quam g i. Nam n h maior tionem i. Ii. vi. ele. Eu., quod contrarium est hypothesi. Nam a c h , d f g
est quam g l et h o maior quam g m . Igitur quod fit sub n h,h o rectangulum aggregata adinvicem subiiciuntur aequalia. Eodem rursus argumentationis
50 maius est facto sub l g, g m rectangulo. Quadratus autem ipsius h k 20 genere probabimus e /latitudinem, seu latus quadratif e g, non esse minus
aequalis est, velut patuit, ei quod fit sub n h o et quadratus ipsius g i latitudine b c , seu latere quadrati c b h. Igitur quadratus f e g, respectu
aequalis ei quod fit sub l g, g m rectangulo. Igitur quadratus ipsius h k longitudinis d e, brevioris erit maior quadrato c b h, in comparatione ad
maior est quadrato ipsius g i liquet itaque corolarium, videlicet h k esse prolixiorem a b longitudinem. Si igitur quo (/ duo) data rectangula in­
maiorem ipsa g i, ex communi sententia: latera sunt maiora quorum aequalium latitudinum (/ longitudinum) et reliqua ut supra quod oportebat
55 quadrati sunt maiores. 25 demonstrare.

ELEMENTUM CONICUM XVIII. ELEMENTUM CONICUM XX.

Rectangulo aliquo atque quadrato lineaque recta datis, dabitur quadra­ Duas producere lineas alteram rectam, alteram inflexam quae hyper­
tus ad quem quadratus eiusdem rectae erit sub ratione dati rectanguli ad bole coni sectio est, quae quanto amplius producuntur eo magis vicis-
quadratum datum. sim appropinquant nunquam coincidentes etiam si in infinitum producantur.
5 Datum igitur rectangulum sit a [Fig. 6.21], datus quadratus b, et c recta 5 Sit itaque conus a b c [Fig. 6.23], habens basim b e d circulum; atque a
linea data. Deinde per propositionem ultimam lib. ii. elementorum Eu. coni axe triangulum sit a b c, cuius a b latus producatur in partem a usque
dabitur d, recta linea potens aream a rectanguli. Et per propositionem in e, quoad id libet; et a puncto e intra triangulum a b c recta agatur linea
xii. Ii. vi. elemen. Eu. fiat ratio ipsius c datae rectae lineae ad rectam lineam e / , secans a c latus super g, basim autem b c in/sig n o , atque plano ab axe
e velut est ratio rectae lineae d ad latus quadrati b . Et quoniam ex hypothesi trianguli a b c ad rectos excitetur angulos g h recta linea, et super e g h
[b νϋ r] quadratus ipsius d rectae aequalis est rectangulo a I et per propositionem io planum eat, secans triangulum quidem a b c supere / , conicam vero super­
ii xxii. Ii. vi. eorundem elemen. — Si quattuor rectae lineae proportionales ficiem super g i d linea. Erit itaque g h recta linea contingens g i d sec­
fuerint, et ab eis rectilinea similia similiterque descripta proportionabilia tionem; planum igitur e g h ad rectos secat angulos a b c trianguli ab axe
erunt, igitur quadratus ipsius d, seu aequalis area ipsius videlicet quad­ planum. Intelligatur deinde planum e g he x parte g h in infinitum produci,
ranguli a , ad quadratum b rationem habet quam quadratus ipsius c ad e similiter conum ex parte b e d basis. Et quia idem planum e g h per coni
15 quadratum. Igitur dato rectangulo a , atque quadrato b , datus est quadratus 15 verticem non vadit, ergo recta linea connectens g, i, seu quaecunque alia
ipsius e , ad quem quadratus ipsius c est sicut rectangulum a ad b quadratum sectionis g i d signa, intra conum cadit per secundum postulatum. Ergo
quod oportuit efficere. linea g i d inflexa est, super qua k sumpto puncto ab eo ad planum trianguli
a b c perpendicularis agatur k /, quae necessario cadit in e / , communem
[b viii ■] sectionem trianguli a b c et plani e g h. Praeterea linea / recta e g super m
ELEMENTUM CONICUM XIX.
20 dividue secetur. Atque quae fuerit ratio rectanguli sub e l g contenti ad
Si duo data rectangula inaequalium longitudinum quadratis suarum lati­ quadratum ipsius / k perpendicularis eadem etiam per ele. co. viii. fiat ratio
tudinum iungantur, fuerintque haec duo aggregata invicem aequalia, erit quadrati ipsius g m ad ipsius g h quadratum. Connectatur denique m h
quadratus aggregati maioris longitudinis minor quadrato aggregati brevioris recta linea. Dico quod recta linea m h et g i d inflexa linea, si protrahantur
5 longitudinis. in infinitum, nunque coincident, et quanto amplius producantur, eo magis
Sint igitur data duo rectangula: a b c [Fig. 6.22], cuius a b longitudo 25 invicem propinquant. Concurrant autem coincidantve, si id fuerit pos­
prolixior, et d e / , cuius longitudo d e brevior, atque rectangulo a b c in sibile, super i puncto, a quo ad e g f perpendicularis agatur i n quae neces­
directum additus sit quadratus b c h ipsius latitudinis b c. Similiter rectan­ sario cadit in e g f, communem sectionem e g h plani et trianguli a b c, et
gulo d e f in directum iunctus sit quadratus e f g, atque aggregatum a c h ad planum eiusdem trianguli perpendicularis seu erecta. Quoniam e g h
10 aequale extiterit d f g aggregato. Dico quod e / , latitudo rectanguli d e f, planum super plano trianguli a b c erigitur, erit ergo ratio rectanguli sub
maior sit b c , latitudine rectanguli a b c. Si enim e / , latitudo rectanguli 30 e n, n g contenti ad quadratum ipsius i n sicut quadrati g m ad quadratum
[b νϋ v] d e f, latus existens quadrati e f g , aequalis extiterit ipsi latitudi/ni b c ipsius g h . Quare etiam sicut ratio quadrati m n ad quadratum ipsius i n. Et
286 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 287

quoniam per propositionem ix. li. v. elemen. Eu.— Quae magnitudines ad Igitur quadratum ipsius A:o et duplum eius quod fit sub o k , kl sunt aequalia
eandem, eandem habent rationem aequales adinvicem sunt, ergo rectangu- quadrato ipsius r p et duplo eius quod fit sub r p, p q. Patet igitur lemma
[b viii v] lum sub e η, n g contentum erit aequale quadra/to ipsius m n, utrunque 20 seu assumptum.
35 enim ad quadratum i n eandem habet rationem quam quadratus ipsius g m
ELEMENTUM CONICUM XXI.
habet ad g h quadratum, quod impossibile est per propositionem vi. libri ii.
elementorum Eucli. Quoniam e g bifariam dividitur in m eique in rectum Ad datas rectas lineas non coincidentes quae ad rectum se contingunt
adiicitur# n, igitur per eandem propositionem vi. li. ii. elemen. Eu. quadra­ angulum per datum signum hyperbolen describere.
tum ipsius m n superat id quod fit sub e n, n g rectangulum quadrato ipsius Sint igitur datae duae rectae lineae a b, b c [Fig. 6.24] se invicem con-
40 g m. Patet itaque pars prima propositi elementi conici. Scriptae nanque 5 tingentes in puncto b , rectumque compraehendant angulum α b c et datum
sunt, velut ostensum est, recta m h et inflexa# i d, nunquam coincidentes signum d. Intentioque sit per signum d ipsis α b, b c rectis datisque lineis
quantumcunque producantur. hyperbolen describere non coincidentem. Igitur α b c rectus angulus
Deinde dico quod quanto plus m h et g i d linea[e] producantur, tanto ci v dividue / secetur per propositionem ix. li. i. elemen. Eu. producta recta
magis sibi invicem appropinquant. Protrahatur ergo linea k /, incidens ipsi linea b e . Et a signo d ipsi b e ad rectos angulos agatur d e , incidens in ip-
45 m h super o puncto atque in g i d, hyperbole seu conica sectione, post k 10 sam b e super e signo. Eadem denique perpendicularis d e in utrasque
assumatur p signum, a quo super e g f perpendicularis agatur p q, quae in partes producta secet ipsam quidem α b in α et ipsam b c super c puncto.
partes p acta occurrat ipsi m h productae super r. Et quia quadratus o k Et ex b e auferatur b / , quae per propositionem ultimam libri ii. elemen.
atque duplum eius quod fit sub o k, k l aequalia sunt quadrato p r et duplo Euclidis possit areolam parallelogrammi seu rectanguli cuius longitudo
eius quod fit sub/? r , p q. Et quia m / ad / o est ut m q ad <7 r, est autem m q aequalis extiterit ipsis d e, e c pariter iunctis in directum et latitudo ipsi
50 maior quam m l, igitur q r maior est quam l o. Et quia per xvii. elementum 15 d c aequalis. Et per/ signum producatur/ # parallelus ipsi c e et secans b c
conicum p q maior est quam k /, igitur per xix. elementum conicum k o lineam super #. Et quoniam per constructionem angulus b e c rectus est et
maior est quam p r. Ergo signum p propius est rectae lineae m h productae / # parallela ipsi e d e , igitur per propositionem xxix. li. i. elemen. Eu.
quam signum o (f k). Horum autem utrunque signorum o (! k),r ( /p) existit angulus b f g rectus est. Ex hypothesi autem angulus f b g dimidius recti
in hyperbolica sectione g i d. Et quoniam idem de omni alio puncto quod in existit, igitur per propositionem xxxii eiusdem libri primi elemen. angulus
55 eadem obliqua linea hyperbolicae sectionis g i d extiterit eodem modo 20 b g f recti dimidius est. Et quia duo qui ad basim b g sunt anguli trianguli
demonstrari poterit usque in infinitum, igitur quanto amplius recta linea b f g sunt aequales, igitur per propositionem vi eiusdem primi libri elemen.
m h et inflexa linea hyperbolicae sectionis g i d producantur eo amplius / # aequalis est ipsi b / , cuius quadratus aequalis est per constructionem
appropinquant, quod secundo demonstrare oportuit. parallelogrammo habenti longitudinem quidem aequalem ipsis c e, e d
pariter additis in rectum, latitudinem vero aequalem ipsi c d. Igitur quadra-
Lemma seu assumptum. 25 tus ipsius / # aequalis est parallelogrammo rectangulo habenti longi­
Quod autem stantibus praemissis hypothesibus et constructione quadra­ tudinem quidem aequalem ipsis c e, e d in rectum pariter additis, lati­
tum ipsius k o et duplum eius quod fit sub o k, k l sint aequalia quadrato ip­ tudinem autem aequalem ipsi c d, ita quod quadratus ipsius/ # aequalis est
sius p r et duplo eius quod fit sub r p, p q, quotienscunque hyperbole ad quadrato ipsius c d et duplo eius quod fit sub c d , d e. Producta deinde b e
5 non coincidentem describitur, sic liquebit. Nam ratio rectanguli sub e 1,1 g in partem b usque ad h signum sitque b h aequalis ipsi b / . Ipsa denique
ad quadratum ipsius k l est sicut ratio quadrati ipsius g m ad quadratum ip­ 30 e / in aliquot utcunque scindatur sectiones, atque per sectionum signa ipsi
sius g h , atque idcirco sicut ratio quadrati ipsius / m ad ipsius / o quadratum α c paralleli peragantur, secantes α b, b c datas lineas, quanto autem
cir / per propositionem iiii. li. vi. elemen. Eu. Erit igitur per propositionem arctiores ipsius e f sectiones capiuntur tanto exactius scribetur per d sig­
xix. li. quinti ele. Eu. ratio quadrati ipsius# m, quod est differentia qua/ m num hyperbole non coincidens α b, b c rectis. Earundem quoque sec­
10 quadratus excedit rectangulum sub e l, l g, ad ipsius o k quadratum et tionum prima sit e i, et per / ipsi e d e parallelus sit actus / k /, secans b c in
duplum eius quod fit sub k o, k l sicut ratio quadrati ipsius g m ad ipsius 35 / puncto, atque ex ipsa i l dematur i k potens parallelogrammum rectan­
# h quadratum. Igitur per secundam partem propositionis nonae li. v. ele. gulum sub h /, i f contentum. Et quia per propositionem vi. li. ii. elemen.
Eu. quadratus ipsius # h aequalis est quadrato ipsius k o et duplo eius quod Eu. f h bifariam secatur in b, atque ipsi/ /i in rectum ad d itu r//, ergo
fit sub o k, k /, nam ad utrunque eorum quadratus ipsius # m refertur sub quadratus ipsius b i aequalis est ei quod fit sub h i, i f et quadrato ipsius
15 eadem proportione. Similiter quoque demonstrabitur quod quadratus ip­ cnr b / . / Et quia / k l aequalis est ipsi b f i, igitur quadratus ipsius / / aequalis
sius # h aequalis est quadrato ipsius p r et duplo eius quod fit sub r p , p </; 40 est parallelogrammo sub hi , i f contento et quadrato ipsius b f. At ex
at ex communi sententia, quae uni sunt aequalia adinvicem sunt aequalia. hypothesi quod fit sub h i, i f aequale est quadrato ipsius / k. Igitur ex com-
288 ARCHIMEDES IN THE MIDDLE AGES VERNERI DE ELEMENTIS CONICIS 289

muni sententia— si aequalibus auferantur aequalia quae reliquuntur aequalia io xxxi. Ii. iii. elemen. Eu. angulus m p n rectus est atque ab eo ad m n per­
sunt— quadratus ipsius k / atque quod bis fit sub i k , k l sunt aequales quad­ pendicularis agitur o p , ergo o p perpendicularis media proportionalis est
rato ipsius b / seu ipsius/ g. Pari modo demonstrabimus quadratum ipsius inter m o , o n. Est autem m n ex hypothesi aequalis ipsi h e, et η o aequalis
45 d e esse aequalem parallelogrammo quod fit sub h e, e f rectangulo. Nam ipsi e f . Et velut ostensum est d e media proportionalis est inter h e, e f per
b e aequalis est ipsi e c per propositionem vi. Ii. i. ele. Eu. quoniam in propositionem vi. Ii. vi. elemen. Eu., ergo o p aequalis est ipsi d e. Rursus
triangulo b c e rectangulo anguli qui ad basim b c sunt aequales, eorum 15 η o in totidem et aequales partes secetur ipsius e f sectionibus, quarum n r
enim uterque recti dimidius existit, et quia per vi. propo. Ii. ii. elemen. sit aequalis ipsi e i . Et centro item q , et spacio q r, scribatur semicirculus
eorundem quadratus d e aequalis est ei quod fit sub h e, e f atque quadrato r s t, secans o p perpendicularem super s et m q superi. Dico itaque quod
50 ipsius b / , et per iiii propositionem eiusdem secundi libri quadratus ipsius 0 s sit aequalis ipsi i k. Nam η o aequalis est ipsi e f , e t n r equalis ipsi e i.
c e est aequalis duobus quadratis ipsarum c d, d c atque duplo eius quod Atqui ex communi sententia— Si aequalibus auferantur aequalia quae
fit sub c d, d e, igitur quadratus ipsius b f atque rectangulum sub h e, e f 20 relinquuntur erunt aequalia, ergo o r est aequalis ip s i//. Pari argumenta-
contentum aequales sunt duobus quadratis ipsarum c d, d e atque duplo c iii r tione constabit s t (! o t) esse aequalem ipsi h i. Igitur tota r o t / aequalis
eius quod fit sub c d, d e rectangulo. Sed per constructionem quadratum est ipsis h i / , in rectum pariter compositis. Igitur o s aequalis est ipsi i k,
55 ipsius b / , aut ipsius aequalis lineae / g , aequale est quadrato ipsius c d potenti rectangulum sub h i, i f contentum. Pari denique et aliae potentes
atque duplo eius quod fit sub c d, d e. Igitur ex communi sententia— si similia rectangula dabuntur. Perspicuum ergo est propositum lemma seu
aequalibus etc.— rectangulum quod fit sub h e, e f aequale est quadrato 25 assumptum.
ipsius d e. Si demum in reliquis parallelis per puncta divisionum ipsius e f
ELEMENTUM CONICUM XXII.
ductis similia signa qualia sunt d, k pari constructione constituantur, eaque
60 rectis connectantur lineis, inflexa quaedam creabitur linea hyperboles Ab hyperbole non coincidentibus actae rectae lineae rectangulas areolas
haud absimilis, cuius fastigium/ , axis vero e / , per conversionem lemmatis comprehendent aequales.
elementi conici xx. Nam hyperbole per d , f signa descripta et non coin- Sint igitur rectae lineae a b , b c [Fig. 6.26] non coincidentes ipsi hyper-
cidens ipsi b c meabit per puncta in parallelis ipsis d e, i k et reliquis dicto 5 bolae d e f , atque ex e , / punctis ipsius hyperboles d e f non coincidentibus
modo signata per idem lemma elementi conici xx. Et velut ipsius hyper- a b , b c (/ a c) parallelae agantur e g , e h j l j k . Dico quod duo rectangula
65 boles d k f dimidium constitutum est in partem c, ita quoque reliquum a g e h, a l f k sint aequalia. Ex c g igitur ipsi e g aequalis dematur g m .
dimidium constituetur in partem a. Et inflexa linea f k d atque recta b c, Connexaquem e et in partem e producta secet a b super n. Et iterum / c sit
in partes c d productae, nusquam coincident per ea quae circa elementum aequalis ipsi/ /, protractaque c / in partem/ secet rectam quidem a b super
conicum xx. fuerunt ostensa atque per eius corolarium. Datis igitur duabus io b , hyperbolen autem d e f super d . Et quoniam uterque duorum angulorum
rectis lineis a b, b c, quae se invicem contingunt in puncto b, rectum con- ad m , n signa per constructionem recti dimidio aequalis est, igitur ut m e
c Hv tinentes angu/lum, per datum inter easdem lineas d signum hyperbole ad e g sic e n ad e h . Utrobique enim ratio est diametri ad costam quadrati.
71 d k f descripta est, non coincidens ipsis a b, b c rectis, etiam si eadem Eandem quoque rationem pari modo probabimus esse inter c f , f L Et
hyperbole atque eaedem rectae lineae a b, b c in partes a d c in infinitum quoniam duo triangula b f k, e h n aequiangula sunt, ergo per proposi-
producantur, quod oportuit demonstrare. 15 tionem iiii. Ii. vi. elemen. Eu. ut b f ad e n s i c f k ad e h. Et quia per coro­
larium vigesimi elementi conici rectangulum sub b / , / c contentum aequale
est compraehenso sub m e, e n rectangulo, igitur per secundam partem
Lemma seu assumptum.
propositionis vi. Ii. vi. elemen. eorundem/? / a d e « est sicut e m a d c /, seu
Ut autem rectae lineae ipsa sub/i e, e / e t sub h i, i f et similiter contenta sicut e g a d / /. At iam ostensum fu it/ k esse ad e h ut b / ad e n \ ergo u t / k
rectangula potentes breviter ac ferme eodem inveniantur denturque mo­ 20 ad e h sic e g a d / / per propositionem xi. Ii. v. elementorum Eu.— Quae
mento, tali utendum est compendio. c iii v uni eaedem sunt / rationes et cetera. At qui circum aequales angulos g e h,
5 Sumatur itaque m n [Fig. 6.25] recta aequalis ipsis h e, e f in directum 1 f k reciproca sunt latera; igitur per propositionem iiii. Ii. vi. elemen.
compositis. Sitque η o aequalis ipsi e f ipsaque m o aequalis sit ipsi h e. eorundem duo parallelogramma α g e h, a l f k sunt aequalia. Ab hyper­
Deinde ex o signo ipsi m n ad rectos excitetur angulos o p . Et m n dividue bolae igitur non coincidentibus actae rectae lineae rectangulas areolas
secetur in q, et centro q, intervallo autem m q , scribatur semicirculus 25 compraehendunt aequales, quod oportuit demonstrasse.
m p n, secans perpendicularem o p in p signo. Et quia per propositionem Explicit libellus de Vigintiduobus elementis conicis.
290 ARCHIMEDES IN THE MIDDLE AGES WERNER’S ON CONIC ELEMENTS 291

The Booklet of Johann Werner of Nuremberg on Twenty-Two Second Division

Conic Elements Of cones, one is right and another is scalene or oblique. The right cone is
that which has [its] axis at right angles to [its] base. The scalene or oblique
First Definition
cone is that which does not have [its] axis at right angles to [its] base. This
The cone is the figure which arises when one of the sides about the right division is appropriate for any of the three species of the first division.
angle of a right triangle remains fixed and the triangle is rotated until it re­
turns again to the place from which it began. And if the fixed straight line First Postulate
is equal to the other side of those about the right angle— the side that was
rotated, the cone will be orthogonal or right-angled. If it is less, the cone If a straight line joins the apex of a cone and a point in the circumference
will be amblygonal or obtuse-angled. But if it is more, it will be oxygonal or of the base or any point taken elsewhere in the conical surface, that [line]
acute-angled. lies in the conical surface. A great many mathematicians call this same
[It is defined] in another way [see Fig. 6.1]. The cone is a figure which straight line the side of the cone.
arises if a straight line is drawn in both directions from some point to the
circumference of a circle which is not in the same plane as the point and, Second Postulate
with the point fixed, if the straight line is rotated on the circumference of
A straight line connecting [any] two points on the surface of the cone ex­
the circle until it returns again to the same place from which it began to
cepting the apex falls inside the cone.
move. And so the mathematicians call “ a conical surface” the surface de­
scribed by the straight line, which surface consists of two surfaces mutually
[i.e. oppositely] placed with respect to the apex and each of which surfaces Third Postulate
is increased indefinitely by the description of the straight line in each direc­ If a plane cuts a cone through the culminating point or apex of the cone,
tion when the straight line has been produced indefinitely. The high point the section common to the conical surface and the cutting plane is a recti­
[that remains] in the same place [they call] its apex. The straight line drawn linear triangle.
through that high point and the center of [each] circle, [they call] its axis.
[Each] circle [they call] its base. Fifth Definition

Second Definition The axial triangle is the common section produced in the conical surface
when the plane cuts the cone through the axis. Moreover such a section
Following the first definition of the cone, the axis of the cone is the cer­ common to the conical surface and the cutting plane is a rectilinear triangle,
tain straight line which remains fixed and around which the triangle is turned. by the third postulate.

Third Definition Fourth Postulate

Following the first definition of the cone, the base of the cone is the circle If a plane parallel to the base of a cone cuts the cone, the section common
described by the rotated straight line. to the cutting plane and the conical surface is a circle.

Fourth Definition Fifth Postulate

Following the first definition of the cone, the vertex or culminating point If a plane cutting a cone does not pass through its apex and is not parallel
or apex of the cone is that highest point of the fixed side [of those sides] to the base, the section common to this same plane and the conical surface
about the right angle. Finally, the axis, base, and apex of the cone are de­ is a certain curved line which the mathematicians call a conic section.
fined in another way in the second definition of the cone.
Third Division
First Division
Of conic sections, one is a parabola, the second is a hyperbola, and the
Of cones, one is orthogonal or right-angled, a second is amblygonal or third is an ellipse. A parabola exists when the plane cutting the cone is
obtuse-angled, and the third is oxygonal or acute-angled. The particular erected [at right angles] to the plane of the axial triangle and the common
members of this division are manifest from the first definition of the cone. section of these planes is parallel to the remaining side of the axial triangle.
292 ARCHIMEDES IN THE MIDDLE AGES WERNER’S ON CONIC ELEMENTS 293

A hyperbola exists when the common section of the said planes meets the if from this same point there is drawn a line parallel to another of the sides,
remaining side of the axial triangle produced beyond the apex of the cone. another line parallel to the subtended side will so cut it (the first parallel)
An ellipse arises when the same common section of the said planes meets that its upper segment is the mean proportional between the segments of
the remaining side of the axial triangle within the cone. the second parallel.
And so in an isosceles right triangle a b c , let point d cut off c d in sub­
Sixth Definition tended side be [see Fig. 6.3]. In the first case let it (cd) be less than one third
The axis of the common section, or, as others say, the diameter, is the of the whole b e ; and through d let a line de be drawn parallel to a c , cutting
straight line which bisects the straight lines that are drawn from the conic side ab at e ; and from be let there be taken ef, which is double a e ; and let f g
section at right angles to it (the axis). The terminal point of its axis in the drawn parallel to the subtended side be cut ac at g and d e a t h . And because
conic section is called the vertex of the conic section. The straight lines ag and eh are parallel by construction, therefore, by Proposition VI.2 of
drawn at right angles to the axis of the section are called by mathemati­ the E l e m e n t s , f h I hg = e f / ae. And by hypothesis e f = 2 ae. Therefore,
cians [lines] drawn or produced “ in order” or “ordinate-wise” ; and f h = 2 gh. But, by Proposition 1.47 of the Elements, f h 2 = 2 e h2, for e f
sometimes mathematicians also call them [lines] drawn “ according to = eh and angle f e h is a right angle. Therefore, f h I gh = f h 2 1h e 2 and there­
order. ’’ Finally, that line of those drawn ordinate-wise or in order which is fore the three lines f h , eh, and gh are continually proportional by the
equal to the segment of the axis terminated at it, or sometimes also the corollary of the preceding element. Therefore, with [a point given] in the
straight line according to which the lines drawn ordinate-wise are squared, subtended side of [an isosceles right] triangle, etc., as above. But if d c is
is customarily called the latus rectum by geometers. more than one third of the whole subtended side b e, therefore f g will fall
below the subtended side b d e , and thence the proposed [problem] will be
The First Conic Element effected in the same way as before.
In a given isosceles right triangle, to divide the side subtending the right
The Third Conic Element
angle so that a line drawn from the point of division to the other side is the
mean proportional between the segments of the subtended side. If a point is designated in the base of an isosceles right triangle, from
Let abc be an isosceles right triangle [see Fig. 6.2] having a right angle which point a parallel is drawn to one side [of those] about the right angle,
bac and whose subtended side is b e, from which we let a third part dc be a parallel that is a mean proportional between the segments of the base,
taken. Similarly, from side ab let one third be taken, namely ae. And with then [if] another line is drawn anywhere parallel to the base and cuts the
de connected, I say that the subtended line be be [so] divided at d [that] first parallel and the remaining sides of the triangle, the rectangle which
the line de drawn from it (d) parallel to side ac is the mean proportional be­ arises from the segments of the second parallel is equal to the rectangle
tween segments b d and dc of the subtended side be. And since, by Proposi­ which arises from the [whole] first parallel and that segment of it which lies
tion 1.47 of the Elements of Euclid, b d 2 = 2 d e 2, therefore b d 2 / d e 2 = b d / between the side next to the right angle and the second parallel.
d c , for, by construction, b d = 2 cd. Therefore, by the second corollary to Let ab c be an iscosceles right triangle [see Fig. 6.4] having right angle
Proposition VI.20 of the Elements (“if three lines are proportionals, the a b c , and about the same angle abc the two equal sides ab and b e, and the
first to the third will be as the figure described on the first to the similar and base a c , in which, from point d , de is drawn parallel to be and cuts side ab at
similarly described figure on the second”), the three straight lines b d , de, e and de is the mean proportional between the segments a d and dc of base
and dc are continually proportional. Therefore, in a given [isosceles] right a c . Again letf g h be drawn parallel to base ac and cut ab at/and the parallel
triangle, etc. as above, which it was necessary to effect. And thatac and de de at g and side be at h Λ say that the rectangle made from/? and gh is equal
are parallels is clear from Proposition VI.2 of the Elements. For ae / eb to the rectangle made from de and eg. But since f g / ge = e d / d c , orf g / ge
= c d / db. = e d / gh, by the corollary of the first conic element [above] (for each is the
ratio of v T / 1): therefore, by Proposition VI. 16 of the Elements of Euclid
Corollary (“ If four straight lines are proportionals etc.” ), f g - g h = de eg. There­
Thence it is also clear that the three lines b d , d e , and dc are proportionals fore, if a point is designated in the base of an isosceles right triangle etc., as
according to the ratio of the diameter of a square to its side, that is accord­ above, which it was necessary to demonstrate.
ing to the ratio whose square is 2/1.
The Fourth Conic Element
The Second Conic Element
If a point is assigned in the subtended side of an isosceles right triangle,
With a point given in the subtended side of an isosceles right triangle, from which point a parallel is drawn to one side [of those] about the right
which point cuts off more or less than one third of this subtended side, and angle and the parallel is the mean proportional between the segments of the
294 ARCHIMEDES IN THE MIDDLE AGES WERNER’S ON CONIC ELEMENTS 295

base, and the parallel is extended in the direction of the subtended line as the Elements of Euclid angle adc is a right angle, and (5) dg is at right angles
above, and with a [second] point assumed from which a parallel to the base to ac. therefore, by the corollary to Proposition VI.8 of the same Elements
or subtended side is drawn so that it cuts the two sides about the right angle of Euclid, dg is the mean proportional between ag and g c . But by construc­
when they have been extended in the same direction, the rectangle made tion dg is equal to fg. Therefore (6) fg is the mean proportional between ag
from the segments of the second parallel is equal to the rectangle which and gc. And ihk is parallel to ac, the base of the cone’s axial triangle abc.
arises from the [whole] first parallel extended and the part of it which is the Therefore, by the third conic element, gf-fh = ih -hk. But iek is the cir­
mean proportional between the segments of the base. cumference of a semicircle, as was evident; and eh, by the sixth definition,
Leta^c, as before, be an isosceles right triangle [see Fig. 6.5] having the is at right angles to ihk·, and, by Proposition III.31 of the Elements of
two equal sides ab and be about right angle abc, and in the subtended side Euclid, iek is a right angle. Therefore, by the corollary to Proposition VI.8
ac let there be the designated point d, through which de is drawn parallel to of the Elements of Euclid, eh is the mean proportional between ih and hk.
be and is a mean proportional between the segments ad and dc of base ac. Therefore, by Proposition VI. 17 of the Elements of Euclid, eh2 = ih -hk,
And [let] de [be] produced rectilinearly in the direction of d as far as one and it has just been demonstrated that ih-hk = gf-fh. Therefore, eh2
likes up to/ . And let ghfbe drawn through/ parallel to the subtended side - gf'fh. But by construction fg is the latus rectum of parabola def, and eh
ac and cutting, at points g and h , the two sides about the right angle, ab and is drawn ordinate-wise to the same latus rectum from e , a point of parabola
b e , when they have been extended in the same direction, while ed has been def. Therefore, if a line drawn ordinate-wise from a parabola of a right-
extended in that same direction to/ . I say that gh -hf = fe-ed. And because angled right cone to the axis of this parabola falls etc., as above, which it
fg / ef = de I dc, while cd = fh by the corollary to the first element of this was necessary to demonstrate.
[booklet] (for on both sides there is a ratio of V2~/1), therefore, by Proposi­
tion VI. 16 of the Elements of Euclid (“ If four proportional straight lines The Sixth Conic Element
etc.”), gf-fh = fe-ed. Therefore, if a point is assigned in the base or sub­
tended side of an isosceles right triangle, from which etc., as above, which If a line drawn ordinate-wise from a parabola of a right-angled right cone
it was necessary to demonstrate. to the axis of this parabola falls beyond the latus rectum [positioned on the
axis], the square of the line drawn ordinate-wise will be equal to the rectan­
The Fifth Conic Element gle which arises from the latus rectum and that segment of the axis which
lies between the line drawn ordinate-wise and the vertex of the parabola.
If a line drawn ordinate-wise from a parabola of a right-angled right cone Therefore, in right-angled right cone abed [see Fig. 6.7] let there be a
to the axis of this parabola falls on the latus rectum [positioned on the axis] of parabola def, whose axis is fg , and from point d line dg is drawn ordinate-
the parabola, the square of the same line drawn ordinate-wise will be equal to wise and cuts the axis of the parabola at point g . And let the latus rectum of
the rectangle which arises from the latus rectum and that segment of it [on the same parabola be eh, which, by the sixth definition, is always equal to
the axis] taken between the vertex of the parabola and the line drawn the coterminal segment of the axis,/7i. Incidentally it should also be noted
ordinate-wise. here that among the ancient geometers each of these two lines eh andfh is
Therefore, in right-angled right cone abed [see Fig. 6.6], whose apex is found to have the name of latus rectum for the reason (I strongly believe)
b and base is acd, let there be parabola def with axis fg and vertex/; and let fhat they are equal. But, to return now from the digression to the purpose
fg be the latus rectum of the same parabola def. And so d g , drawn ordinate- at hand, let the cone’s axial triangle be a bc. And so I say that dg2 = gf-fh.
wise, is equal to f g . Let e be given in parabola def, the point from which eh And so let us understand that some plane through eh and parallel to the
is drawn ordinate-wise to latus rectumf g . I say that the square of eh, drawn base of cone abed cuts the same cone. Therefore section ihk common to
ordinate-wise, is equal to the rectangle which arises from g f andf h . There­ this plane and the cone’s axial triangle abc is parallel to the subtended side
fore let some plane cut cone abed through eh and parallel to the base acd·, ac of right triangle abc. Finally, by the fourth postulate, iek, the section
therefore, by the fourth postulate of this [booklet], iek, the section common common to the plane and the curved or conical surface, is a circle. There­
to the cutting plane and the conical surface, will be a circle. And let the fore, by Proposition III.31 of the Elements of Euclid, angle iek is a right
cone’s axial triangle be abc, which bisects circle ek; and ik, the common angle. And because, by the sixth definition, eh is at right angles to ihk,
section of the cone’s axial triangle and circle eik, is a diameter of the same therefore, by the corollary to Proposition VI.8 of the Elements of Euclid,
circle iek and it will of necessity go through point h. And because (1) adc is eh is the mean proportional between ih and hk. But, by the definition of
also a semicircle, and (2) by construction dg is at right angles to diameter ac latus rectum, eh is equal to fh. Therefore,fh is the mean proportional be­
of the same semicircle adc, and (3) dg is equal to fg by hypothesis and the tween ih and hk. Therefore, by the fourth conic element [of this booklet],
sixth definition of this [booklet], and because [4] by Proposition III.21 of ag-gc = gf-fh. But dg2 = ag-gc, for adc is the circumference of a semi­
296 ARCHIMEDES IN THE MIDDLE AGES WERNER’S ON CONIC ELEMENTS 297

circle and, by Proposition III.31 of the Elements of Euclid, angle adc is a The Ninth Conic Element
right angle, and, by construction, dg is at right angles to ac. Therefore
dg2 = gf'fh, g f being the axis of the parabola andfh the latus rectum. But With a parabola of a right-angled right cone given, and with its axis given,
dg, the line drawn ordinate-wise, by construction and hypothesis falls to to give the right-angled right cone.
the axis of the parabola beyond the latus rectum fh. Therefore, if a line Let parabola ab of a right-angled right cone be given [see Fig. 6.10], with
drawn ordinate-wise from a [point of a] parabola of a right-angled right cone its [given] axis ac. And, by the eighth conic element [of this booklet], let
to its axis falls beyond the latus rectum [positioned on the axis], the square the latus rectum of parabola ab be given, and let it be be. And, by Proposi­
etc., as above, which it was necessary to have demonstrated. tion 1.23 of the Elements of Euclid, let there be constructed, on axis ac at
point a , a right angle cad, the plane of which is at right angles to the plane of
the parabola abc. Also let straight line ad be equal to a c ; and, with cd con­
The Seventh Conic Element nected, let ad be bisected in points, and let ad be extended in the direction
The ratio of the squares of any two lines drawn ordinate-wise from a of a to /su c h that a f = ae. And let fg be drawn parallel to ac, and let cd,
parabola of a right-angled right cone is as the ratio of those segments of the which has been produced in the direction of c , fall upon fg at point g . And
axis that are coterminal with them. since by construction fg has been drawn [parallel] to ac, therefore, by
Therefore let there be a parabola abed of a right-angled right cone [see Proposition VI. 2 of the Elements of Euclid, dc leg = da I af.Butda = 2 af;
Fig. 6.8], the axis of the parabola being aefg, to which axis we let be led or­ therefore dc - l e g . And because, by Proposition 1.47 of the Elements of
dinate-wise [lines] be and dg. I say, therefore, that be2 / dg2 = ea / ag. Euclid, dc2 = 2 ac2, therefore dc2 / ac2 = dc / eg. Or since, by the corol­
Therefore, let a f be the latus rectum of parabola abed. And because, by the lary to Proposition VI.20 of the Elements of Euclid, “ similar polygons
fifth or sixth conic element, ( l ) be2 = fa ae, and similarly (2) dg2 = ga-af, [. . .] are in the duplicate ratio of [their] similar sides” , therefore dc / eg
but, by Proposition V I.1 of th^Elements of Euclid, (3)fa-ae / ga -af = ae / = (dc lea)2. Therefore, by the definition of duplicate ratio, ac, or its equal
a g , for they are under the same altitude af, [which is] the latus rectum of the be, is the mean proportional between dc and eg. And because by construc­
parabola by construction or by hypothesis: therefore (4) the squares of tion angle bed is a right angle, and, by the corollary to Proposition VI.8 of
be and dg equal to these rectangles have the [same] ratio as ae to a g . There­ the Elements of Euclid, angle dbg, constructed by imagining lines db and
fore, [the ratio of the squares of any two lines drawn ordinate-wise] from a bg connected, is a right angle, therefore, by Proposition III.31 of the
parabola of a right-angled right cone etc., as above, which it was necessary Elements of Euclid, with dg bisected in h, circle dbg, described on center
to demonstrate. h with radius dh, will go through point b . Then, withfh joined, and because
the angle a t /is a right angle and df is equal to fg, therefore the angles at h
The Eighth Conic Element are right angles, and each of the two angles which are onf g , the base of right
triangle fhg, is half of a right angle. Therefore, if fh is the fixed side of the
With a parabola produced in a right-angled right cone given, and with its partial triangle fhg and the same partial triangle fhg is rotated until it re­
axis given, to give the latus rectum of the parabola. turns to the place from which it had begun to be moved, by the first defini­
Therefore let abc be the given parabola of a right-angled right cone [see tion [of this booklet] it will describe a right-angled right cone whose base is
Fig. 6.9], with [given] axis ad and vertex a. By Proposition 1.33 of the circle dbg. Finally, this triangle fgh, by its rotation, little by little or suc­
Elements of Euclid, let there be constructed, on axis ad at given point a, cessively will cut the parabola ab in all of its points. Moreover, that the
rectilinear angle dac equal to half of a right angle; and let ac cut parabola circle dbg is the base of the cone produced by the rotation of triangle/g/i is
abc in point c, from which point we let line cd be drawn ordinate-wise to clear as follows. Since circle dbg has been erected [perpendicularly] to the
axis ad. And because angle adc in triangle acd is a right angle and cad is half plane of triangle dfg because by construction the plane abc was erected
of a right angle by hypothesis, therefore, by Proposition 1.32 of the Ele­ [perpendicularly] to the plane of triangle dfg, and by hypothesis angle acb
ments of Euclid, angle acd is also half of a right angle since the three in­ is a right angle, therefore straight line be is erected [perpendicularly] to the
terior angles of a rectilinear triangle are equal to two right angles. There­ plane of triangle dfg by the third definition of Book XI of the Elements of
fore triangle acd is isosceles by Proposition 1.6 of the Elements of Euclid. Euclid (“ A plane is at right angles to a plane etc.” ). But the plane of circle
And because cd, drawn ordinate-wise to axis ad, is equal to ad, therefore, dbg passes through straight line be. Therefore, by Proposition XI. 18 of the
by the sixth definition [of this booklet] regarding a latus rectum of a parab­ Elements of Euclid, the plane of circle dbg is at right angles to the plane of
ola, either cd or its equal ad is the latus rectum of the given parabola abc. triangle dfg. Therefore the same circle dbg is the base of the cone which
Therefore, with a parabola produced in a right-angled right cone given, partial triangle fgh has described by its rotation. Therefore, right-angled
etc., as above, which was necessary to effect. right conefdbg in whose conical surface the given parabola ab is described
298 ARCHIMEDES IN THE MIDDLE AGES WERNER’S ON CONIC ELEMENTS 299

be produced indefinitely in both directions from klmn. And let there be cut
is given. Therefore, with a parabola of a right-angled right cone given, also
off on the perpendicular drawn through / on both sides two lines, [each]
is given the cone in whose conical surface the same parabola is described,
equal to bf, and let these lines be lo and Ip. But also let there be taken two
which it was necessary to have demonstrated.
equal straight lines on the perpendicular drawn through m , on the one side
mq and on the other mr, so that each of the lines mq and mr is equal to per­
The Tenth Conic Element pendicular &/. Let this be done as often as one likes; and with these points
Given two lines drawn ordinate-wise from a parabola of a right-angled (like o , p , q , r ) joined on each side of klmn by straight lines, it will thus be
right cone and given the segment of the axis between them, the whole axis evident that the proposed parabola has been described. Therefore let
straight line klmn be equal to the given latus rectum ab. And because (1)
of the same parabola will be given.
Therefore let abc be the given parabola [see Fig. 6.11], and let the lines afe is a semicircle by construction and (2) angle afe is a right angle by
be and cd drawn ordinate-wise from it, as well as the segment de of the axis Proposition III.31 of the Elements of Euclid if two straight lines a f and/<?
between them, be given. I say that the whole axis of the parabola aed may are imagined as drawn, therefore, by the corollary to Proposition VI.8 of
[then] be given. And since, by the seventh conic element [of this booklet], the same Elements, (3) b f is the mean proportional between ab and b e . Now
cd2 / be2 = da / ae, therefore, by subtraction, (cd2 - be2) / be2 = de / ea. kn is equal to the given latus rectum a b , and lo is equal to perpendicular bf,
But, with three terms in this proportion given by hypothesis, the fourth and kl is equal to be. Therefore, each of these lines lo and Ip is the mean
term, ae, the segment of the axis, is given. Therefore, the whole axis aed proportional between kn and kl. Therefore, by Proposition VI. 17 of the
of parabola abc is given. Therefore, given two lines drawn ordinate-wise Elements, kn-kl = lo2 = Ip2. Then, by similar reasoning, kn-km = mq2
from a parabola of a right-angled right cone etc., as above, which it was = mr2. We shall demonstrate the same thing in no different way for the rest
of the perpendiculars drawn through the points of kn extended. There­
necessary to demonstrate.
fore, by the converse of the fifth or sixth conic element [of this booklet],
the curved line qokpr is the conic section which is called a parabola and
The Eleventh Conic Element
which falls in a right-angled right cone and which is the sort that the ninth
Given a straight line according to which lines drawn ordinate-wise in a conic element [of this booklet] teaches how to construct when the latus
parabola are squared, to describe this parabola in a given plane. By the rectum is given. Therefore, given straight line ab, or its equal kn, accord­
sixth definition, the straight line according to which lines drawn ordinate- ing to which lines lo, Ip, mq, mr, and the remaining lines, [all] drawn or­
wise are squared is called by another name the latus rectum of the parabola. dinate-wise, are squared, parabola qokpr has been described in a given
By the same reasoning the parabola there which it is proposed to describe plane, which it was necessary to have effected. It ought to be noted in pass­
is that which falls in a right-angled right cone. ing that by the amount the parts of latus rectum kn are narrower by such an
Therefore, let ab be the given line according to which the lines drawn amount will the proposed parabola be more truly drawn. In fact, any
ordinate-wise are squared [see Fig. 6.12]. And let this ab be cut into any straight line joining any two immediately neighboring points differs in a
number of equal parts: a c , cd, d b ; and let b e , equal to any one of them, be small degree and quite insensibly from the curved bit of the parabola which
added to ab in a straight line, and on point b let bf be erected at right angles is bounded by the same points. In addition, it should be noted that a seg­
to ae, which bf we let be extended indefinitely in the direction o f/; and, ment of the parabola that lies next to the apex is distant by a trifling and al­
with ae bisected at d, let semicircle afe be described on center d with radius most insensible space from the circumference of the circle which is de­
ad, the semicircle cutting perpendicular bf in point/. Again, let e g , equal to scribed on a center that is the bisection point of the latus rectum with a
be, be added; and let the whole line be aeg, which again we let be bisected radius equal to half of the same latus rectum, as is manifestly clear from the
at a [new] point h , and again let semicircle aig be described on h as a center accompanying description [see the superior diagram in Fig. 6.12].
with radius ah, the semicircle crossing the perpendicular b f at point i. And
in this way let there be added in the straight line as many lines equal to eg as The Twelfth Conic Element
we please and let there be described [just as many] semicircles cutting per­
pendicular b f in individual points beyond point / . Further, let a certain If from some point taken outside of a given circle two lines are drawn,
other line klmn, equal to a b , be taken in the subject plane. And let it be in­ one indeed to the center and the other tangent to the same circle, and from
definite in the direction of n , having parts kl , l m, mn, equal in number and the point of tangency a perpendicular is dropped on the line that was drawn
to the center, [then] the line drawn to the center, the radius of the circle,
magnitude to parts a c ,c d ,d b , and so on with any number of equal parts as­
and the segment of the line drawn to the center which lies between the cen­
sumed to infinity. And through points /, m, n, and the other points as­
ter and the perpendicular are continually proportional.
signed to straight line klmn, let perpendiculars be erected, which we also let
300 ARCHIMEDES IN THE MIDDLE AGES WERNER’S ON CONIC ELEMENTS 301

And so let ab be the given circle [see Fig. 6.13], whose center is c, and the two straight lines connected to d and e from any other point in addition
let the point taken anywhere outside of circle ab be point d, from which to /h a v e the same ratio as ad to ae. Therefore, if from some point taken
point let straight line cd be connected to center c , cutting circle abdita, and outside of a given circle etc., as above, which it was necessary to have
let line bd be drawn tangent to circle ab at point b . Further, let be be a per­ demonstrated.
pendicular from the point of tangency b to cd. I say that cd drawn to center
c , the radius ac of circle a b , and ce are continually proportional. Let b and Corollary
c be joined. And since, by Proposition III. 17 of the Elements, angle cbd is a From this it also becomes clear that conjoined line a f bisects angle efd, so
right angle, and the angles at point e are right angles by the definition of per­ that angle afd = angle afe. For, as it has been demonstrated that d f / ef
pendicular, therefore cd / be = be I ce. But ac = be, for each is a radius of = ad! ae, therefore, by the second part of Proposition VI. 3 of the Elements
circle ab. Therefore, the three straight lines: cd drawn to center c of circle of Euclid, straight line a f bisects angle efd. The same will be evident for
ab, radius ac of the same circle, and ec, are continually proportional. every angle produced by two straight lines drawn from some point in cir­
Therefore, if some point [taken] outside of a given circle etc., as above, cumference ab to points d and e. Therefore the corollary has become
which it was necessary to demonstrate. manifest.
Corollary
The Fourteenth Conic Element
From this it is also obvious that the straight lines cd, ac and ce are con­
In a given cone, to draw a plane tangent to the same cone through a
tinually proportional according to the ratio of ad to ae. For, by Proposition
straight line produced from the vertex of the cone to [its] base.
V. 19 of the Elements of Euclid, whole cd / whole ac = (cd - ac)/ (ac - ce). Therefore let abed be a cone whose base is circl eacd [see Fig. 6.15]. And
Therefore, remainder ad / remainder ae = whole cd / whole a c . Therefore,
let b be the vertex of cone abed, from which vertex let line bd be joined to
the three lines cd, ac, and ce are continually proportional according to the
the base acd at its circumference acd in point d, line bd being, by the first
ratio of ad to ae, and so the corollary is manifest.
postulate, a straight line lying in the conical surface of cone abed, and let it
be proposed that a plane passes through straight line bd tangent to the
The Thirteenth Conic Element
conical surface of cone abed on straight line bd. Therefore let line de be
If from some point taken outside of a given circle two straight lines are drawn through point d tangent to circumference acd at d. And so it has be­
drawn to the same circle, one to the center and the other tangent to the come clear that the plane in which the two straight lines bd and de are
circle, and [if] a perpendicular is drawn from the point of tangency to the drawn is tangent to the conical surface of cone abed on straight line bd, for
line drawn to the center, and [if] from a point assumed anywhere in the cir­ those two straight lines bd and be lie in the same plane by a proposition of
cumference of the same circle two straight lines are joined, one indeed to Book XI of the Elements of Euclid. Therefore, in the given cone, plane bde
the said point outside of the circle and the other to the terminus of the said is drawn through a straight line bd that proceeds from the vertex of the
perpendicular: the ratio of the same lines drawn from [the point in] the cir­ cone to the circumference of the base and the plane is tangent to the curved
cumference of the given circle is as the ratio of the [segment of the] straight surface of cone abed on straight line bd, which it was necessary to effect.
line drawn to the center of the circle which lies between the point assumed
outside and the same circle [i.e. circumference] to the straight line which is Corollary
included between the same circle [i.e. circumference] and the aforesaid
From this it is also clear that if some plane is tangent to the subject
perpendicular.
conical surface on a given straight line, that same plane is also tangent to
Keeping the same description and figuration as in the preceding element,
every curved line which is described in the conical surface and cuts the
let p o in t/b e taken anywhere in the circumference of circle ab [see Fig. said straight line, and the common section of the two lines— straight and
6.14]. Letdf ande f be connected from this point. I say thatd ftfe = ad / ae. curved— will be the point of tangency.
Therefore let cf be connected; and because the sides about the common
angle ecf in the two triangles cdf and cef are proportional, for, by the pre­ The Fifteenth Conic Element
ceding element dc / cf - cf I ce, therefore the two triangles cdf and cef are
equiangular by Proposition VI.6 of the Elements of Euclid (“and [will To draw a straight line tangent to the curved line of any given conic sec­
have] those angles equal which the corresponding sides subtend.”). There­ tion in a given point.
fore c f / ce = df / ef·, but cf / ce = ad / ae by the corollary to the preceding Therefore letfgh be a certain conic section of any kind in given cone abed
conic element. Therefore ad / ae = d f I ef. Similarly it will be proved that whose base circle is acd [see Fig. 6.16], and on the curved line of the same
 WERNER’S ON CONIC ELEMENTS 303
302 ARCHIMEDES IN THE MIDDLE AGES
of Euclid. Then, by hypothesis, angle edg is a right angle; therefore ci and
section fgh let the given point be h, through which point it is necessary to dg are parallel. Therefore, just as dg is erected [at right angles] to the plane
draw a straight line tangent to the same curved line fgh in point h . There­ of triangle cdh, so is ci. Therefore, by Proposition XI. 18 of the Elements
fore, from the vertex b of cone abed and through the same point h, let of Euclid, the plane tangent to cone fegh is erected [at right angles] to the
straight line bhd be produced which cuts the base of the cone in its circum­
plane of triangle cdh. And because the plane of the cone’s axial triangle
ference in point d, and let plane bde be applied through bhd to the conical fgh, as was evident, is also erected [at right angles] to the plane of triangle
surface of cone abed, so that it is tangent to the same conical surface on cdh, therefore section eh, common to the plane of the cone’s axial tri­
straight line bhd by the fourteenth conic element. And let hk be the section angle fgh and to plane ceh, is erected [at right angles] to the plane of the
common to the plane in which section fgh lies and to plane bde. And be­ same triangle cdh by Proposition XI. 19 of the Elements of Euclid. There­
cause line hk is the common section of these planes, it has only point h in fore ci and eh are parallel by Proposition XI.6 of the Elements of Euclid.
common with curved linef g h . Therefore, by definition and by the corollary And ci is also parallel to dg, as was evident; therefore eh is also parallel to
to the preceding conic element, straight line hk will be tangent to the curved dg. And since, by construction, ce is tangent to parabolaabc in pointc, and
line fgh at point h . Therefore a straight line has been drawn tangent to the point e is a point common to plane ech and to the plane of the cone’s axial
curved line of conic section fgh through given point h , which it was neces­ triangle f g h, therefore point e necessarily lies in the common section eh of
sary to effect.
these planes fgh and che. And, as was evident, eh is parallel to dg; there­
The Sixteenth Conic Element fore, by Proposition 1.29 of the Elements of Euclid, angle ahe is equal to
angle agd ; therefore it is also equal to angle ahd. Finally the angles at a are
If a straight line is tangent, in a given point, to a parabola which falls in a right angles, and side ah is common to the two triangles aeh and adh.
right-angled right cone, and [if] the same tangent and the axis of the parab­ Therefore the two triangles adh and aeh are equilateral by Proposition 1.26
ola are [both] extended in the same direction until they meet, the exterior of the Elements of Euclid. And because equal are the sides which are sub­
part of the extended axis which lies between the vertex of the parabola and tended by equal angles, therefore ad = ae. Therefore, if [a straight line is
the said point of intersection [of tangent and axis] is equal to that segment of tangent], in a given point, to a parabola which falls in a right-angled right
the axis which is interposed between the same vertex of the parabola and the cone etc., as above, which it was necessary to demonstrate.
line drawn ordinate-wise from the point of tangency. But if eg is an arc less than a quadrant [see Fig. 6.18], therefore the plane
Therefore let abc be a parabola of the sort proposed, with axis a d [see tangent to cone fegh on straight line he will meet with the diameterfdg ex­
Fig. 6.17]. And let straight line ec be tangent to parabolaabc at point c. And tended in the direction of g. And so let it meet at point /. And line hi con­
let tangent ce and axis a d [both] produced in the direction of a meet at point nected will necessarily go through point e. For straight line hi is the section
e , and let c d be drawn ordinate-wise from the point of tangency c to axis a d common to the said plane tangent to the cone on straight line he and to the
of the parabola. I say that a d = a e . And because by hypothesis the given plane of the axial trianglefg h . Therefore point e necessarily lies on straight
parabola ab c falls in a right-angled right cone, therefore the same cone is line hi. Then let cd, drawn ordinate-wise, be conjoined, which line cd, by
given by the ninth element. And so let it be given as cone f e g h with vertex hypothesis and construction, is at right angles to diameter fdg. And if the
h, axis dh, and base f c g . And let the cone’s axial triangle be f g h , cutting cone’s axial triangle fhg were laid in the same plane as circle fcg, point h
plane a c d at right angles through the axis a d of the parabola. And sof c g will will be on the circumferencefcg because angle fhg is a right angle and the
be the circumference of a semicircle. Then, with eh connected, it will be on subtended side fg is the diameter of circle f c g . Finally, let dg be connected.
the curved surface of cone f e g h . Now first let arc eg be one half of semi­ And because ci is tangent to circlefcg at c , and cd is drawn from c at right
circle f c g , or a quadrant of the whole circumference of the circle, as the angles to fgi, which was itself drawn through the center of the circle, there­
subject figure has [it]. Therefore from pointc the line led ordinate-wise falls fore, by the [thirteenth] conic element demonstrated above, the two angles
on d, the center of circle f c g . And so the plane of triangle cdh will be erected dhg and ghi are equal. And because, by construction, the angles at a are
at right angles to two planes, namely that of the cone’s axial trianglef g h and right angles and side ah is common to the two triangles ahd and ahe, there­
that of circle f c g by Proposition XI. 18 of the Elements of Euclid. And fore the two triangles adh and ahe are equiangular and equilateral by
through straight line ch let plane ceh be led, tangent to cone f e g h on straight Proposition 1.26 of the Elements of Euclid. Therefore ad —ae. Therefore,
line ch by the fourteenth conic element; finally, the section common to the if [a straight line is tangent,] in a given point, etc., as above.
plane of circle f c g and to plane ceh tangent to cone f eg h is ci, which is also Now if arc eg [see Fig. 6.19] exceeds a quadrant of a circle, therefore ci
erected at right angles to the plane of triangle c d h . F ord , by construction, will cut diameterfg extended in the direction off. And so again it becomes
is tangent to circle fcg in point c and cd has been drawn from center d toe,
clear that straight lines hi and d a , when extended in the directions of h and
the point of tangency to the parabola, by Proposition III. 18 of the Elements
304 ARCHIMEDES IN THE MIDDLE AGES WERNER’S ON CONIC ELEMENTS 305

a , meet one another at the same point e , unless someone wished to affirm gle produced by (1) the [segment of the] straight line [extending from the
that the one straight line may intersect the other straight line in two points, top terminus] to the first perpendicular and (2) its segment which lies be­
which is most absurd and very strange to an experienced geometer. For, tween the same perpendicular and the second side of the axial triangle is to
if thus, two straight lines would include a surface, which in no way can be the square of the same perpendicular as the ratio of the rectangle included
done, it being contrary to a common opinion of geometers. [Let the cone’s under (1) the [segment of the] straight line [extending from its top terminus]
axial trianglefhg be imagined as lying in the same plane as circle fcg.] And up to the second perpendicular and (2) its segment which is terminated [on
so let the center of circle fcg be taken and let it be k. And let h and k be con­ the one end] at the second perpendicular and [on the other end] at the
joined. And because, by construction, triangle fgh is isosceles, therefore [second] side of the axial triangle to the square of the same second per­
the angles at k are right angles. And because hk and kg are radii of circle f c g , pendicular.
hence they are equal. Therefore, in triangle hkg, each of the two angles on And so let abc be a cone [see Fig. 6.20], whose base circle is be and axial
base gh is half of a right angle. And through point h let hi be drawn parallel triangle is abc. And let side ab of this triangle be extended as far as one
to fg. And because gfi and hi are parallel, therefore, by Proposition 1.29 likes in the direction of apex a to point d, from which point let straight line
of the Elements, angle ehl is equal to angle fih. In addition, by the twelfth de be drawn to base be of the cone’s axial triangle and cut side ac of the
conic element [of this booklet], the three lines ik,kf, and dk are continually same triangle at/; and in straight line ef let any two points g and h be taken,
proportional. Then let dh be connected. Therefore ik / kh = kh / dk. And be­ from which points let there be erected to the plane of triangle abc two per­
cause right angle dkh is common to the two triangles hik and dkh and, as pendiculars gi and hk that proceed to the conical surface at points i and k. I
has been demonstrated, the sides about that same angle are proportional, say that dg- gf / gi2 = dh-hf I hk2. Therefore let two lines Igm and nho be
therefore, by VI.6 of the Elements of Euclid, the two triangles hik and drawn through points g and h parallel to bec and terminated at the sides ab
kdh are equiangular, and equal are the angles to which sides of the same and ac of triangle a bc. Then let there be understood two planes through
ratio are subtended. Therefore angle dhk - angle hik. But it has already Igm and nho that cut cone abc and are parallel to base be. Therefore, by
been demonstrated that angle ehl = angle hik. Therefore angle dhk = angle the fourth postulate, the sections common to these same planes and to the
ehl. Moreover angles Ihg andghk are equal, for each is equal to angle hgk as conical surface will be circles lim and nko, whose diameters are Im and no.
was evident and by Proposition 1.29 of the Elements of Euclid. Therefore, And so lg-gm = gi2 and nh-ho = hk2, by the corollary to Proposition VI.8
by axiom (‘‘If equals are added to equals etc. ” ), the two angles ehg andghd of the Elements of Euclid. For the [imagined] angles lim and nko are right
that are composed of equals are [themselves] equal. Now, in the two tri­ angles by Proposition III. 31 of the same Elements. And since, by Proposi­
angles aeh and ahd, the angles at a are right angles and side ah is common. tion VI. 23 of the Elements, rectangular parallelograms have the ratio to one
Therefore, by Proposition 1.26 of the Elements of Euclid, the two triangles another that is compounded of [their] sides, therefore dg -gf / Ig -gm = (dg /
ahe and ahd are equiangular and equilateral and sides ad and a e , which are gl)-(fg / gm). Similarly, dh-hf / nh-ho = {dh / hn)-(fh / ho). But dg / gl
subtended by equal angles, are equal. But, by hypothesis, straight line ec = dh / hn. Also fg / gm = fh I ho. Therefore dg- gf / Ig-gm = dg- gf / gi2
is tangent to parabola abc at point c . Therefore, if a straight line is tangent, = dh-hf / nh-ho = dh -hf I hk2, from the axiom: those ratios are the same
in a given point, to a parabola which falls in a right-angled right cone, and which are composed of the same ratios. [Then simplifying this by taking the
[if] the same tangent and the axis of the parabola are [both] extended in the second and the fourth of these equal ratios, dg- gf / gi2 = dh-hf / hk2.]
same direction until they meet one another, the exterior part of the ex­ Therefore, if one side of an axial triangle in a given cone is extended beyond
tended axis which lies between the vertex of the parabola and the said the apex of the cone etc., as above, which it was necessary to demonstrate.
[point of] intersection is equal to the segment of the axis which is interposed
between the same vertex and the line drawn ordinate-wise from the point Corollary
of tangency of the parabola, which it was necessary to have demon­
strated so far. From this it also becomes clear that hk > gi. For nh > gl and ho > gm.
Therefore nh-ho > lg-gm. But, as was evident, hk2 = nh-ho and gi2
The Seventeenth Conic Element = lg-gm. Therefore hk2 > gi2. And so the corollary is clear, namely that
hk > g i, from the axiom: those sides are greater whose squares are greater.
If one side of an axial triangle in a given cone is extended beyond the apex
of the cone, and [if] from the terminus of the same extended side a certain The Eighteenth Conic Element
straight line is drawn to the base of the axial triangle and cuts the other side
of the same triangle, and [if] in the same straight line, inside the cone, there Given some rectangle and a square and a straight line, there will [also]
are erected from any two points two lines perpendicular to the plane of the be given a square to which the square of the given straight line will be in the
axial triangle and proceeding to the conical surface, the ratio of the rectan­ same ratio as the given rectangle to the given square.
306 ARCHIMEDES IN THE MIDDLE AGES WERNER’S O N CONIC ELEMENTS 307

Therefore let the given rectangle be a [see Fig. 6.21], the given square b, And so let abc be a cone [see Fig. 6.23], having a base circle b ed; and let
and the given straight line c . Then, by the last proposition of Book II of the abc be the cone’s axial triangle, and let side ab of this triangle be extended
Elements of Euclid, there will [also] be given a straight lined which squared as far as one likes in the direction of a to e ; and from point e let straight line
equals the area of rectangle a. And, by Proposition VI. 12 of the Elements of ef be drawn inside of triangle abc and cut side ac at g and base be in point/;
Euclid, let the ratio of the given straight line c to straight line e become as and let straight line gh be erected at right angles to the plane of axial triangle
the ratio of straight line d to the side of square b . And since, by hypothesis, abc, and let a plane pass through eg and gh and cut [the plane of] triangle
d 2 = rect. a, and, by Proposition VI.22 of the same Elements, “ if four abc in e f and the conical surface in line gid. And so straight line gh will be
straight lines are proportional, the rectilineal figures similar and similarly tangent to section gid; therefore plane egh cuts axial triangle abc at right
described on them will also be proportional,” therefore the square of d, or angles. Then let it be understood that plane egh is produced indefinitely in
its equal area, namely the rectangle a , has the same ratio to square b as the the direction of gh and the cone similarly in the direction of base bed. And
square of c to the square of e. Therefore, given rectangle a and square b because the same plane egh does not go through the apex of the cone, there­
[and straight line c], given [also] is the square of e, to which the square of fore the straight line connecting g and i, or any other points of section gid,
c is related as rectangle a is to the square b, which it was necessary falls inside of the cone by the second postulate. Therefore line gid is curved.
to effect. With a point k taken on it (line gid), let line kl be drawn from k perpendicular
to the plane of triangle abc, which perpendicular necessarily falls on ef, the
The Nineteenth Conic Element section common to triangle abc and plane egh. In addition, let straight line
eg be bisected at m . And, by the eighth conic element, let el-lg / Ik2 = gm2/
If two given rectangles of unequal lengths are joined to the squares of
gh2. Finally, let straight line mh be connected. I say that straight line mh
their widths, and these two aggregates are equal to one another, the square
and curved line gid, if they are extended indefinitely, never meet, and the
[part] of the aggregate of greater length will be less than the square [part] of
more they are extended the closer they come to each other. But, if it be pos­
the aggregate of shorter length.
sible, let them proceed and meet at point/, from which point let perpendicu­
Therefore let two rectangles be given [see Fig. 6.22]: abc, whose length lar in be drawn to egf, which in necessarily falls on egf, the section com­
ab is the longer, and def, whose length de is the shorter; and to rectangle
mon to plane egh and triangle abc, and is perpendicular to the plane of the
abc let there be added directly the square bch of the width b e . Similarly, to
same triangle, or it is erected [at right angles]. Since plane egh is erected
rectangle def let there be joined directly the square efg, and the aggregate
[at right angles] to the plane of triangle abc, therefore en ng / in2 = gm2 /
ach is equal to the aggregate dfg. I say that ef, the width of rectangle def,
gh2. Hence also en-ng I in2 = mn2 / in2. And since, by Proposition V.9 of
is greater than b e , the width of rectangle abc. For if ef, the width of rectan­
the Elements of Euclid, “ magnitudes which have the same ratio to the same
gle def, which is the side of square efg, is equal to the width be of rectangle
[magnitude] are equal to each other,” therefore en-ng = mn2, for each has
abc, which width is evidently the side of square bch, then eg, the side of
the same ratio to the square of in as the square of gm has to the square of
square efg, will be equal to bh, the side of square cbh, and the whole straight
gh . This is impossible by Proposition II.6 of the Elements of Euclid. Since
line deg will be less than abh, from the axiom: if equals are added to equals
eg is bisected at m and gn is directly added to it, therefore, by the same
etc. Therefore the aggregate of rectangle def and square feg will be less
Proposition II.6 of the Elements of Euclid, mn2 = en-ng + gm2. And so
than the aggregate of rectangle abc and square cbh by Proposition VI. 1 of
the first part of the proposed conic element is evident. For, as has been
the Elements of Euclid, which is contrary to the hypothesis. For aggregates
demonstrated, the described lines— straight line mh and curved line gid—
ach and dfg are taken as equal to one another. Again, by the same kind of
never meet, no matter how far they are extended.
argument, we will prove that width ef, or the side of square feg , is not less
Then I say that by the amount that lines mh and gid are extended farther,
than width be, or the side of square cbh. Therefore the square feg, in ac­
by so much more do they come closer to one another. Therefore let line kl be
cordance with the shorter length a b , will be greater than square cbh, in ac­
extended and fall on mh at point o . And in gid, the hyperbola or conic sec­
cordance with the longer length ab. Therefore, if two given rectangles of
tion, let there be assumed, beyond k, a point/?; from which point letpq be
unequal lengths etc., as above, which it was necessary to demonstrate.
drawn perpendicular to egf, and, extended in the direction ofp , let it meet
mh, itself extended, at point r. And because ok2 + 2 ok-kl = pr2 + Ipr- pq
The Twentieth Conic Element
[as will be proved below], and because ml / lo = mq / qr, while mq > ml,
To extend two lines— one straight and the other curved in the form of a therefore qr > lo. And because, by the seventeenth conic element, pq
hyperbola— such that the farther they are extended the closer they ap­ > kl, therefore, by the nineteenth conic element, ko > pr. Therefore point
proach one another without ever meeting even if they are extended in­ p is closer to straight line mh extended than is point k . But each of these
definitely. points k and p is on the hyperbolic section gid. And since the same can be
308 ARCHIMEDES IN THE MIDDLE AGES WERNER’S ON CONIC ELEMENTS 309

demonstrated for any other point which is on the same curved line of hyper­ curate will be the hyperbola described through d that does not meet lines
bolic section gid, to infinity, therefore the farther that the straight line mh ab and b e . Let the first of these sections be e i , and through i let ikl be drawn
and the curved line of the hyperbolic section g i d are extended, the closer parallel to ede and cut be in point /, and let ik be cut off from il such that
they are, which it was necessary to demonstrate in the second place. ik2 = h i - i f And because, by Proposition II.6 of the Elements of Euclid, fh
is bisected in b and fi is directly added to f h , therefore bi2 = h i- i f + b f 2.
Lemma or Assumption And because ikl is equal to bfi, therefore//2 = h i- i f + b f 2. But, by hypothe­
sis, h i - i f = ik2. Therefore, by axiom (“ if equals are taken away from equals
Moreover, with the premissed hypotheses and construction standing, it
the remainders are equal” )/:/2 + 2 ik-kl = b f 2 = f g 2. By the same method
will be clear in what follows that k o 2 + 2 ok -kl = p r 2 + 2 r p - p q , how­
we will demonstrate that d e 2 - h e- ef , for be = ec by Proposition 1.6 of the
ever often the hyperbola is described with respect to its asymptote, as will
Elements of Euclid since in right triangle bee the angles which are on base
be clear. For el-lg I kl2 = g m 2 / g h 2 = Im 2 / l o2, by Proposition VI.4 of the
be are equal (for each of them is one half of a right angle), and because, by
Elements of Euclid. Therefore, by Proposition V.19 of the Elements of
Proposition II. 6 of the same E l e m e n t s , d e 2 = h e - e f + b f 2, and, by Proposi­
Euclid, withgra2 = Im2 - (el-lg), g m 2 1 (ok2 + 2 ko-kl) = g m 2 / g h 2. There­
tion II.4 of the same work, c e 2 = c d 2 + d e 2 + 2 c d - d e , therefore b f 2
fore, by the second part ofV.9 of the Elements of Euclid, g h2 = ko2 + l o k - k l ,
+ h e - e f = c d 2 + e d 2 + 2 c d - d e . But, by construction, b f 2 = f g 2 = c d 2
for gm2 has the same ratio to each of them. Similarly, it will also be demon­
+ 2 c d - d e . Therefore, by axiom (“ if equals [are taken away] etc.), h e - e f
strated that g h 2 = p r 2 + 2 rp -pq. But, by axiom, magnitudes equal to the
= e d 2. Then, if points of the same kind as d and k are established by a
same magnitude are equal to one another. Therefore, k o 2 + 2 ok-kl = r p 2
similar construction in the rest of the lines drawn parallel through the points
+ 2 rp -pq. Therefore the lemma or assumption is evident.
of the divisions of ef, and [if] these points [like d and k] are connected by
straight lines, a certain inflexed line will be fashioned that is scarcely dif­
The Twenty-first Conic Element
ferent from a hyperbola whose vertex is/an d axis is ef, by the converse of
To describe a hyperbola through a given point and between two given the lemma of the twentieth conic element. For the hyperbola that is de­
asymptotic straight lines that meet at a right angle. scribed through points d and/ and does not meet be will pass through the
Therefore let ab and be be the two straight lines that meet each other in points assigned in the said way in d e , i k , and the remaining parallels, by the
point b and include right angle a b c , and let d be the given point [see Fig. same lemma of the twentieth conic element. And just as one half of the
6.24]. The intention is to describe a hyperbola through point d that does not parabola d k f h a s been constructed in the direction of c, so the remaining
meet given straight lines a b and be. Therefore let right angle ab c be bi­ half will also be constructed in the direction of a. And curved lin e f k d and
sected by straight line be extended, by means of Proposition 1.9 of the straight line b e, when extended in the direction of c d, never meet, by those
Elements of Euclid. And let de be drawn from point d at right angles to be things which were demonstrated concerning the twentieth conic element
and fall on be at point e. Finally, let the same perpendicular de be ex­ and by its corollary. Therefore, given straight lines ab and b e, which touch
tended in both directions and cut ab in point a and be in point c . And let b f each other in point b and contain a right angle, hyperbola d k f has been de­
be cut off from be such that, by the last proposition of Book II of the Ele­ scribed through point d lying between these lines, a hyperbola that does not
ments , its square ( b f 2) is equal to a parallogramatic area or rectangle whose meet straight lines ab and be, even if the same hyperbola and the same
length is equal to de + ec and whose width is equal to de. And through point straight lines ab and be are extended indefinitely in the direction of a d c ,
/ let line f g be drawn parallel to ce and cut line be at g. And since, by con­ which it was necessary to demonstrate.
struction, angle bee is a right angle and f g is parallel to e de, therefore, by
Proposition 1.29 of the Elements of Euclid, angle bfg is a right angle. But,
Lemma or Assumption
by hypothesis, angle f b g is half a right angle; therefore, by Proposition 1.32
of the Elements of Euclid, angle b g f i s half of a right angle. And because the Now in order that the straight lines whose squares equal rectangles con­
two angles which are on the base bg of triangle bfg are equal, therefore, by tained by he and ef, hi and if, and the like [lines], may be found briefly and
Proposition 1.6 of the Elements, f g = bf, whose square ( b f 2), by construc­ approximately and may be given in a moment, the following short cut
tion, is equal to the rectangle having its length equal to ce + e d and its should be used.
width equal to cd. Therefore/?2 = (ce + ed)-cd, so that/#2 = c d 2 + 2 c d-d e And so let straight line mn be taken equal to he and e f placed together in a
[since ce = e d + c d ] . Then let be be extended in the direction of b to point straight line [see Fig. 6.25]. And let no be equal to e f and m o equal to he.
h , such that bh = bf. Finally, let e f be cut into any number of sections and Then from point o let op be erected at right angles to m n . And let mn be bi­
through the points of the sections let parallels to ac be drawn and cut given sected in q , and with q as the center and mq as the radius let semicircle mpn
lines ab and be, and the narrower are the sections of e f taken the more ac­ be drawn, which semicircle cuts perpendicular op inpointp. And because,
310 ARCHIMEDES IN THE MIDDLE AGES

by Proposition III.31 of the Elements of Euclid, [imagined] angle mpn is a CHAPTER 7

right angle and op is drawn from it perpendicular to mn, therefore per­
pendicular op is the mean proportional between mo and on. And, by
hypothesis, mn —he and no = ef. And, as it has been demonstrated that
de is the mean proportional between he and e f by Proposition VI.6 of the
Elements of Euclid, therefore op = de. Again let no be divided into as
many equal parts as there are sections of ef, with nr = ei. And with q as the
center and qr as the radius let semicircle rst be drawn and cut perpendicular The L ast Stages o f the M edieval
op at 5 and mq ati. And so I say that os = ik. For no - ef and nr = ei. But,
by axiom, if equals are taken from equals the remainders are equal; there­ Traditions o f Conic Sections
fore or = f i. By similar reasoning ot = hi. Therefore the whole rot = hi
+ if, the latter lines having been joined together in a straight line. There­
fore os = ik, for os2 and ik2 are each equal to hi-if. Finally, by the same
method are given other lines whose squares are equal to like rectangles. In the preceding chapter we saw European scholars poised to make use of
Therefore the proposed lemma or assumption is clear. the Greek text of the Conics of Apollonius, without, however, doing so in
any thorough or coherent way. I pointed to the limited use of the Conics by
The Twenty-second Conic Element Giorgio Valla in the preparation of his De expetendis et fugiendis rebus
opus. I also noted the use of the text by Andreas Coner (1508 or thereafter),
Straight lines drawn from a hyperbola to the asymptotes form equal the corrector and owner of William of Moerbeke’s autograph of his transla­
rectangular areas. tion of the works of Archimedes, and by Francesco Maurolico in his tracts
Therefore let straight lines ab and be be the asymptotes to hyperbola def of 1534. During the next generation the promise began to be fulfilled. Three
[see Fig. 6.26]. And from points e an d / of hyperbola def let lines eg, eh, fl , Latin versions of the Conics were produced. The first of these was that pre­
andfle be drawn to asymptotes ab and ac,fl parallel to eg andfk parallel to pared by Giambattista Memmo (Ioannes Baptista Memus) and published
eh. I say that rect. ageh = rect. alfk. Therefore letgm, equal to eg, be cut after his death by his nephew Giovanni Maria Memmo: Apollonii Pergei
off from eg. And let me, connected and extended in the direction of e, cut philosophi, mathematicique excellentissimi Opera per doctissimum
ab at n. And again let le be equal tofl; and let ef, extended in the direction of philosophum Ioannem Baptistam Memum Patritium Venetum, mathe-
/ , cut straight line ab in b and hyperbola def in d. And since each of the maticharumque (!) artium in Urbe Veneta lectorem publicum de Graeco in
angles at points m and n is, by construction, equal to half of a right angle, Latinum traducta et noviter impressa (Venice, 1537).1 This translation
therefore me / eg = en / eh. For on each side the ratio is that of a diagonal
to the side of the square. And we shall also demonstrate in the same way 1 As samples of Memmo’s style I quote the following propositions (with the beginnings and
that the same ratio exists between cf andfl. And since the two triangles bfk endings of their proofs); 7r: [Liber I] “ Propositio Undecima. Si conus piano secetur per
and ehn are equiangular, therefore, by Proposition VI.4 of the Elements of axim, secetur autem et altero plano secanti basim coni per rectam ad rectos existentem basi
trianguli per axim, et diameter sectionis equidistans sit uni laterum trianguli per axim, quae
Euclid, bf / en = fk I eh. And because, by a corollary to the twentieth conic
a sectione coni equidistans ducta est communi sectioni secantis plani et a basi (/ basis) coni
element, b f f c = me-en, therefore, by the second part of Proposition VI.6 usque ad diametrum sectionis, poterit contentum sub recepta sub ipsa a diametro ad sumita-
of those same Elements, b f / e n = em / c f = eg / fl. But it was just now tem sectionis, et alius cuiusdam lineae, vel rationem habet ad eam quae inter coni angulum
demonstrated that fk / eh = b f / en. Therefore fk / eh —eg / fl, by Proposi­ est et sumitatem sectionis, quam quadratum quod a basi trianguli per axim ad contentum
tion V. 11 of the Elements of Euclid (“ratios which are [the same] with one sub reliquis duabus (!) trianguli lateribus. Vocatur autem talis sectio Parabole. Sit conus
cuius sumitas punctum a [Fig. 7.1], basis autem bg circulus, et secetur plano per axim et
and the same [ratio] etc.” ). But the sides about equal anglesgeh andlfk are
faciat sectionem triangulum abg, secetur autem et altero plano secanti basim coni per rec­
proportional; therefore, by Proposition VI.4 of those same Elements, rect. tam de ad rectos existentem bg et faciat sectionem in superficie coni dze, diameter autem sec­
ageh — rect. alfk. Therefore, the straight lines drawn from a hyperbola to tionis zh equidistans sit uni laterum trianguli per axem, scilicet ag, et a puncto z ducatur zt
the asymptotes form equal rectangular areas, which it was necessary to linea ad rectos zh et faciat ut quod ab bg ad id quod sub ag (! bag) sic zt ad z a , et relictum sit
have demonstrated. quodam (/ quoddam) punctum in sectione contingens et sit c, et per c ducatur cl equidistans
de. Dico quod id quod fit ab cl equale est illo quod sub tzl. . . ■ (7v). . . . Vocetur autem
Here ends the Booklet of Twenty-two Conic Elements.
talis sectio parabole, tz autem ad quam possunt ductae ordinatae ad zh diametrum. Voce-
turque et recta. . . . (Sample calculations omitted).
“ Propositio XII. Si conus plano secetur per axim, secetur autem et altero plano secanti
basim coni per rectam ad rectos existentem basi trianguli per axim, et diameter sectionis

311
312 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 313

productus (/) coincidat uni laterum trianguli per axim extra coni sumitatem, quae a sectione was, of course, of the first four books only, since these were the only books
ducta est equidistans communi sectioni secantis plani et basi[sj coni usque ad diametrum
extant in the Greek manuscripts that were available (and indeed they re­
sectionis, poterit id quod superficies adiacens ad quandam lineam ad quam rationem habet
in linea manens diametro sectionis tangensque angulum extra triangulum quam quadratum main today the only books extant in Greek). As Heiberg has pointed out,
quod fit adducta (/ a ducta) a sumitate coni ad diametrum sectionis usque ad basim trianguli Memmo made a number of corrections and additions but still left serious
ad contentum sub basi[s] sectionibus quas fecit ducta, latitudo (/ latitudinem) habens recep­ errors, for he was not expert in Greek.2 Though I do not know for sure which
tam sub ipsa a diametro ad excedens sumitatem sectionis specie simili et similiter iacens Greek manuscript he used for his translation, I assume that Heiberg is cor­
contento sub tangente extra angulum trianguli et illius ad quam possunt ductae. Vocetur
rect in his suggestion that since Memmo taught at Venice it was Bessarion’s
autem haec sectio hyperbole. Sit Conus cuius sumitas punctum a [Fig. 7.2], basis autem
circulus bg, et secetur plano per axim et faciat sectionem triangulum abg, secetur autem et manuscript that he used (see Chap. 6 above, η. I).3 Memmo’s translation
altero plano secanti basim coni ad rectam de ad rectos existentem bg basi trianguli abg et was certainly the version used by his friend Niccolo Tartaglia, both in the
faciat sectionem in superficie con idze lineam. Diameter autem sectionis zh productus (/) coin­ few citations he made to William of Moerbeke’s translation of On the
cidat uni laterum trianguli abg, scilicet ag, extra sumitatem coni ad t, et per a ducatur ac Equilibrium of Planes, which he published along with other of the Moer-
equidistans diametro sectionis, et ab z ducatur zl ad rectos [zh] et faciat ut quod ab ca ad id
beke versions in 1543,4 and in the few passages on conic sections found in
quod sub bcg sic tz ad zl, et relictum sit quodam (/ quoddam) punctum contingens in sectione,
scilicet m , et per m ducatur mti equidistans de, et per n ducatur nx equidistans zl, et iuncta tl pro­
his General Trattato.5 It will be noticed that in the latter work, by using
ducatur usque ad x , et per l et x ducantur lo et xp equidistans zn. Dico quod mn potest zx quod defettione as an alternate term for an ellipse, he is adopting Memmo’s term
adiacet ad zl, latitudo (/ latitudinem) habens zn, excedens specie lx existenti [simili] plano (/ {defectio). It is also evident from the General Trattato that Tartaglia was
rectangulo) sub tzl. ■ . . (8r). . . . vocetur autem talis sectio hyperbole, dz ( / lz) autem ad acquainted with one of the principal sources of the medieval treatment of
quam potest ductae ad zh ordinate, voceturque eadem et recta, transversa autem zt. . . . conics, since he remarks that Witelo in his Perspectiva in part demon­
(Sample calculations omitted).
“ Propositio Decimatertia. Si Conus plano secetur per axim, secetur autem et altero plano strated the properties of the parabola and the hyperbola.6 Memmo’s work
coincidenti utrique laterum trianguli per axim, nec ducto [equidistanti] ad basim coni, was also known to the second interpreter of the Conics, the eminent mathe­
neque subcontrarie, planum (8v) autem in quo est basis coni et secans planum coincidat ad matician Francesco Maurolico, who was highly critical of Memmo’s trans­
rectam ad rectos existentem vel basi trianguli per axem vel in rectam ipsi, quae a sectione lation of the Conics in a letter written to Pietro Bembo in 1540 and pub­
coni equidistans ducta est communi sectioni planorum usque ad diametrum sectionis,
lished in Maurolico’s Cosmographia in 1543:7
poterit quod superficies adiacens ad quandam lineam ad quam habet rationem diameter
sectionis quam quadratum quod fit a ducta a sumitate Coni preter diametrum (/ equidistans
diametro) sectionis usque ad basim trianguli [hic om. Mem.\ ad contentum sub lineis ab ipsa the letters that designate points and magnitudes. I have changed vocal “ v” to “ u” and to
ad trianguli lineas receptis], latitudinem habens receptam sub ipsa a diametro ad sumitatem “ ae” . I have given only one diagram for each proposition instead of the pair of diagrams
sectionis deficiens specie simili et similiter posito contento sub diametro et illa ad quam pos­ given by Memmo. Notice that Memmo designates an ellipse as defectio and asymptotes as
sunt ductae ordinate. Vocetur autem talis sectio defectio. Sit conus cuius sumitas punctum non tangentes. Compare Valla’s use of defectus for an ellipse (see above, Chap. 6, n. 6). I
a [Fig. 7.3], basis autem bg circulus, et secetur plano per axim et faciat sectionem triangulum have indicated in parentheses or brackets Memmo’s errors of commission and omission.
abg, secetur autem et altero plano coincidenti utrique laterum trianguli per axim, nec equi­ Memmo’s translation is quite literal as collation with Heiberg’s text demonstrates.
distanti basi coni nec subcontrarie ducto, et faciat sectionem in superficie Coni die lineam, 2 Gr 2, p. LXXXI: “ in mathematicis Memus non pauca, maxime in ordine litterarum,
communis autem sectio secantis plani et eius in quo est basis coni sit zh ad rectos existens computatione recte deducta feliciter correxit et suppleuit, sed graviora reliquit; et Graecae
bg, diameter autem sectionis sit ed, et ab e ducatur et ad rectos ed, et per a ducatur ac equi­ linguae, ut erat οψιμαθή?, non peritissimus erat; uelut uocabulum πορίζβιν non nouit, cuius
distans ed et faciat ut quod ab ac ad id quod sub bcg sic de ad et, et relictum sit quodam (/ loco lacunam reliquit fol. 24u (I p. 150, 2, 6) et fol. 25u (I p. 154, 23, 26); idem fecit eadem de
quoddam) punctum in sectione et sit /, et per / ducatur Im equidistans zh. Dico quod Im potest causa in διβλόι^τι (I p. 62, 26; 94, 13; 116, 28) fol. 10u, 15u, 197, in ε’ίδη (I p. 122, 18) fol. 20r, in
quod superficies quae adiacet ad et, latitudinem habens em, deficiens specie simili eo quod &P ληφθϊ) (I p. 118, 9; 120, 14) fol. 19u, in χαταχθήσοντοη (I p. 172, 21) fol. 27u cet.” The
sub dez. . . . vocetur autem talis sectio defectio, et et ad quam possunt ductae ad de or­ references in parentheses are to Heiberg’s own edition. Cf. my Archimedes in the Middle
dinate, ipsa autem et recta, transversa autem de. . . . (Sample Calculations omitted). Ages, Vol. 3 (Philadelphia, 1978), pp. 563, n. 30; 564, n. 32.
“ (32v) [Lib. II.] Propositio Quarta. Duabus datis lineis angulum continentibus et puncto 3 Gr 2, pp. LXXXI-LXXXII.
intra angulum, describere per punctum Coni sectionem vocatam hyperbole, quare non 4 Clagett, Archimedes in the Middle Ages, Vol. 3, pp. 563 -64.
tangentes ipsas esse datas lineas. Sint duo lineae ag t t a b contingentem angulum continentes 5 N. Tartaglia, La quinta parte del General trattato d e ’ numeri et misure (Venice, 1560),
eum qui est ad a [Fig. 7.4], et datum sit quodam (/ quoddam) punctum sit d, et decens sit per 12r-v, 46r-v, 82r-v. Note that Tartaglia gives as an alternate term for an ellipse the Italian
d, gab describere in non tangentes hyperbole. Coniungatur ad et producatur ad e, et ponatur rendering of Memmo’s defectio, namely defettione (12r: “ . . . quella vera conica settione,
ae equalis da, et per d ducatur dz equalis (/ equidistans) ab, et ponatur zg equalis az, et che da nostri antichi mathematici e chiamata Elipsis (/), et da Apollonio Pergeo difettione. . . .
coniunctagd producatur ad b , et fiat quod est sub d e , h equale ei quod fit ex g b , et producta ad II modo di fare la Elipsis, over Defettione, speculativamente nel dimostra copertamente
describatur circa ipsam per d hyperbole, quare ductas posse id quod ea que sunt ad h ex­ Apollonio Pergeo . . .” ). I have changed the consonantal « tov.
cedentia specie simili ei quod est sub de, h . Quoniam igitur dz equidistans est ba et gz equalis 6 Ibid., 12v: “ . . . vero e che ci sarebbe da dar regola da designar le altre due specie di
est za, igitur et gd equalis db\ quare quod fit ex gb quadruplum est ei quod fit ex gd. settioni conice, delle quali da greci l’una e detta parabola, et l’altra e chiamata hiperbole, la
Et est quod fit ex gb equale ei quod est sub de, z (/, h). [Hic om. Mem.: Utrumque igitur proprieta delle quali in parte lo dimostra Vitellione nella sua perspettiva, ma perche la
quorum que fiunt ex bd, dg quarta pars est speciei que est sub de, h.] Igitur ab et ag non scienza di tai figure e quasi deperdita, me ne passo con silentio. . . .”
tangentes sunt descriptam hyperbolem.” 7 M. Clagett, “The Works of Francesco Maurolico,” Physis, Anno XVI (1974), p. 175, n.
As usual, I have altered the punctuation and capitalization somewhat, and I have italicized 9 (item [3]).
314 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 315

the very famous Conics of Apollonius of Perga, where it will be manifest that rectam de ad rectos ipsi bg, et faciente in conica superficie sectionem dze, cuius diameter zh
Ioannes Baptista Memmius (/), deceived principally by [his] ignorance of aequedistet uni laterum trianguli, ut puta ag\ et zt ad rectos ipsi zh, ita ut zt ad za sit sicut
mathematics, introduced puerile errors into its translation. quadratum bg ad rectangulum bag·, et a quolibet puncto in peripheria sectionis ut c agatur
penes ipsam de ad diametrum zh recta lineae/. Dico quod quadratum cl aequale est rectangulo
Still, Memmo’s translation was the basis for Maurolico’s version, in tzl. . . . Vocetur autem talis sectio Parabole, ipsa autem tz ad quam possunt ductae ordinate
ad zh diametrum vocetur et recta; et manifestum est quod cl semper media proportionalis est
spite of all of the latter’s corrections and additions (see end of note 11 be­
inter lz, zt.
low). Maurolico was a far more skilled mathematician than Memmo, and “ XII. Si conus plano secetur per axim, secetur autem et altero plano secante basim Coni
the mathematician of Messina achieved the most profound knowledge of per rectam ad rectos existentem basi trianguli per axim, et diameter sectionis producta coin-
conic sections of any mathematician in the first half of the sixteenth cen­ cidat uni laterum trianguli per axim extra coni verticem, quae a sectione ducta est aeque­
tury. As we saw in the preceding chapter (above note 9), Maurolico knew distans communi sectioni secantis plani et basis coni usque ad diametrum sectionis, poterit
id quod superficies adiacens ad quamdam lineam ad quam linea in rectum manens diametro
some form of the text of the Conics as early as 1534.1 presume that he must sectionis, occurrensque exterius lateri trianguli, eam rationem habet quam quadratum quod
have seen (and perhaps made notes on) some Greek manuscript of the text fit a ducta a summitate Coni, penes diametrum sectionis in basim trianguli, ad contentum sub
since it is quite certain that he did not see Walther’s supposed translation, basis segmentis ab ea factis et latitudinem habens receptam sub ipsa diametro ad verticem
for, addicted as he was to citing and summarizing the works he had read, he sectionis, excedens specie simili et similiter iacente contento sub coincidente latere tri­
would certainly have somewhere mentioned such a translation had he seen anguli et sub illa ad quam possunt ductae. Vocetur autem haec sectio Hyperbole. (14) Conus
cuius vertex a [Fig. 7.2], basisque circulus b g , secetur plano per axim faciens triangulum abg,
it. Yet it was only Memmo’s translation that he mentioned in his letter of
secetur et altero plano secante basim Coni per rectam de ad rectos ipsi bg et faciente in Conica
1540.8 It was in 1547 that Maurolico executed his version of the Conics, superficie sectionem dze, cuius diameter zh coincidat uni laterum quod fit ga apud t . Item ac
basing it mainly on Memmo’s earlier translation. He completed through penes zh , et zl ad rectos ipsi zh, ita ut tz ad zl sit sicut quadratum ac ad rectangulum bcg, et a
the third book by 2 June (and not 29 June, as given in the edition), and the contingenti puncto sectionis ut pote m ducatur mn penes d e , et connexa tl compleatur rectan­
fourth book by 24 June.9 To these he added a reconstruction of Books V gulum zlon, et collapsis in unum tl, no apud x , compleatur rectangulum Ipxo. Dico quod iam
mn potest rectangulum zx , adiacens scilicet ad ipsam zl, sub latitudine zn, et excedens specie
and VI, presumably on the basis of the brief indication of their contents in Ix, simili rectangulo tzl. ■ ■ ■ Vocetur autem talis sectio Hyperbole, ipsa autem lz ad quam
Apollonius’ proemium to Eudemus.10 Maurolico’s version of the first four possunt ductae ad zh ordinate, voceturque eadem et recta, transversa autem z t , ipsum autem
books exists in manuscript (Real Biblioteca del Escorial, J.III.31), while rectangulum tzl species sectionis. Corollarium. Et manifestum est quod sicut tz ad zl sic est
the whole work was not published until a century later: Francisci Mauro- rectangulum tnz ad rectangulum znx, hoc est ad quadratum mn, namque eadem ratio fuit quae
lyci Messanensis Emendatio et restitutio Conicorum Apollonii Pergaei tn ad nx.
“ XIII. Si conus plano secetur per axim, secetur autem et altero plano coincidente utri que
(Messina, 1654).11 As in the case of Maurolico’s version of Archimedes, laterum trianguli per axim, neque ducto penes basim coni neque subcontrarie, planum
autem in quo est basis coni et secans planum coincidat ad rectam ad angulos rectos existen­
8 In a letter to Pietro Bembo of 1536 (Ibid., p. 172, n. 2) Maurolico simply mentions the tem, vel basi trianguli per axim vel in directum ipsi, quae a sectione coni aequedistans ducitur
work of Apollonius without any further remarks on its translation. communi sectioni planorum usque ad diametrum sectionis, poterit spacium quod adiacet ad
9 Ibid., p. 157. quamdam lineam ad quam habet rationem diameter sectionis quam quadratum quod fit a
10 F. Maurolico, Emendatio et restitutio Conicorum Apollonii Pergaei (Messina, 1654), 5r: ducta a vertice coni penes diametrum (15) sectionis usque ad occursum basis trianguli ad
“ Est enim quintus quidem de minimis et maximis, ut plurimum. Sextus de aequalibus et contentum sub tota linea (quae constat ex basi et adiuncta occurrente) et sub ipsa occurrente,
similibus coni sectionibus.” Earlier (4v) Maurolico explains in a letter to the reader his in­ et latitudinem habens receptam (MS, recepta ed) ex diametro sub ipsa ad summitatem sec­
terest in these two books: ” Est enim quintus ut Apollonius innuit, ut plurimum de maximis tionis, deficiens rectangulo simili et similiter posito contento sub diametro et illa ad quam
et minimis. Sextus de similibus et aequalibus sectionibus. . . . Sed ex his quintus et sex­ possunt ductae ordinate. Vocetur autem talis sectio Ellipsis sive defectio. Conus cuius vertex
tus magis quam reliqui a nobis desiderabantur. Nam quamvis tota horum quatuor libellorum a [Fig. 7.3], basisque circulus bg, secetur plano per axem, sitque sectio per z triangulum
materia pendeat a doctrina priorum, tamen de maximis et minimis aut de similibus et aequalibus abg, secetur et altero plano secante basim coni per rectam zh ad rectos ipsi bg vel eidem in
sectionibus nulla fit in illis mentio.” rectum productae, et coincidente lateribus trianguli abg, et nec aequedistante basi coni nec
11 I present from this edition the same propositions that I extracted from Memmo’s ver­ subcontrarie ducto, et faciente in conica superficie sectionem die, per 9. Item agatur as penes
sion in note 1 above: pp. 12-15, 55 (cf. MS Escorial J.III.3 1 ,6r-8v, 41r): ” [Lib. I, Prop.JXI. ed, coincidens ipsi bg apud s , sitque et ad rectos ipsi ed, ita ut de ad et sit sicut quadratum
Si conus piano secetur per axim, secetur autem et altero plano secante basim Coni per as ad rectangulum bsg, et a contingenti puncto sectionis / agatur ad diametrum Im penes zh,
rectam ad rectos existentem basi trianguli per axim, et diameter sectionis aequedistans sit uni et connexa dt compleantur rectangula etnm, xnto. Dico iam quod Im potest rectangulum
laterum trianguli per axim, quae a sectione Coni ducitur aequedistans communi sectioni secan­ ex, adiacens ad et sub latitudine em, et deficiens specie xt simili rectangulo det. . . . Vocetur
tis plani et basis coni usque ad dia( 13)metrum sectionis, poterit contentum sub recepta ab ipsa autem talis sectio Ellipsis sive defectio, et ipsa et ad quam possunt ductae ad de ordinate, et
ex diametro ad summitatem sectionis et sub alia quadam linea quae rationem habet ad eam eadem recta; transversa autem de. Item rectangulum det species sectionis. Corollarium.
quae inter Coni angulum est et summitatem sectionis eandem quam habet quadratum quod a Et manifestum est quod rectangulum dme ad rectangulum emx, hoc est aed (! ad) quadratum
basi trianguli per axim ad comprehensum sub reliquis duobus trianguli lateribus. Vocetur Im, est sicut dm ad mx vel de ad e t ; videlicet sicut transversum ad rectum speciei latus. Item
autem talis sectio Parabole. Conus cuius vertex a [Fig. 7.1]: basis bg circulus, secetur plano patet quod secunda diametros ellipsis est media proportionalis inter transversum et rectum
per axim, sitque sectio per z triangulum abg, secetur et altero plano secante basim Coni per speciei latus, hoc est inter ipsas de, et. Nam dimidium secundae diametri potest rectangulum
316 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 317

Maurolico made no effort to translate literally. (Indeed I suspect that he Greek works, paid scrupulous attention to the extant text, adding his own
had only Memmo’s version at hand when he undertook the preparation of views on the content and the state of the text in accompanying commen­
his own version.) His chief concern in all of his versions of Greek works taries.13 Commandino’s version included not only the Lemmata of Pappus
was to produce an effective mathematical work.12 For example, the reader to the Conics (preceding each of the first three books of the Conics), but
will see above in note 11 in the text of Proposition 1.11 (at the end of the also the commentary of Eutocius interspersed throughout the text and his
proof), that Maurolico added the alternate form of expressing the ordinate own comments similarly dispersed. To these works he added the two books
of a parabola as the mean proportional between the line that equals the axial of Serenus (De sectione cylindri and De sectione coni). Commandino s
intercept (lz) and the latus rectum of the parabola (zt), a form of the basic published work bears the title: Apollonii Pergaei Conicorum libri quattuor
property of the parabola given by Witelo in Proposition IX.40 (Chap. 3 una cum Pappi Alexandrini lemmatibus et commentariis Eutocii As-
above, n. 41, end of proof) and by the author of the Speculi almukefi com­ calonitae, Sereni Antinsensis philosophi libri duo, nunc primum in lucem
positio in the second part of Conclusion 4 (see Chap. 4, Text A). Mauro­ editi. Quae omnia nuper Federicus Commandinus Urbinas mendis quam-
lico’s mastery of the content of the Conics is not only evident from his ver­ pluribus expurgata e Graeco convertit, et commentariis illustravit (Bologna,
sion of the text but also from a beautiful little tract De lineis horariis liber 1566).14 It is fair to say that Commandino’s superb version of the Conics
tertius, which I shall discuss below.
13 Ibid., pp. 619-20. Also consult Commandino’s version of the Moerbeke translation of
A third version of the Conics was produced by that other great student of
Archimedes’ On Floating Bodies published there.
Greek mathematicians writing around the middle of the century: Federigo 14 From this edition (13v - 14r, 14v- 15v, 44v) I give the same propositions which I presented
Commandino. Unlike Maurolico, Commandino, in his various versions of from Memmo’s translation (note 1) and from Maurolico’s version (note 11): “ (13v) Theorema
XI. Propositio XI. Si conus piano per axem secetur, secetur autem et altero plano secante
quod sub dimidio transversi et dimidio recti; quare et tota poterit rectangulum sub totis con­ basim coni secundum rectam lineam, quae ad basim trianguli per axem sit perpendicularis,
tentum. . . . et sit diameter sectionis uni laterum trianguli per axem aequidistans, recta linea quae a sec­
“ (55) [Lib. II. Prop.]IV. Duabus datis lineis angulum continentibus et puncto intra an­ tione coni ducitur aequidistans communi sectioni plani secantis et basis coni usque ad
gulum, describere per punctum coni sectionem vocatam Hyperbolen, cuius non tangentes sectionis diametrum, poterit spatium aequale contento linea quae ex diametro abscissa inter
sint ipsae datae lineae. Sint duae lineae ba, ag contingentem angulum bag continentes [Fig. ipsam et verticem sectionis interficitur et alia quadam quae ad lineam inter coni angulum et
7.4], et datum sit quodam (/ quoddam) punctum, quod sit d; oportet Hyperbolem describere verticem sectionis interiectam, eam proportionem habeat quam quadratum basis trianguli per
quae per d punctum eat et cuius non tangentes sint ba, ag. Coniungatur ad et producatur, axem ad id quod reliquis duobus trianguli lateribus continetur. Dicatur autem huiusmodi
sitque ipsi da aequalis a e , et per d ducatur dz aequidistans ipsi a b , et ponatur ipsi az aequalis sectio parabole. Sit conus cuius vertex punctum a [Fig. 7.5], basis bc circulus, seceturque
zg, et coniunctagd producatur ad b, et fiat rectangulum deh aequale quadrato bg, et producta plano per axem, quod sectionem faciat triangulum a b e, et secetur altero plano secante basim
ad describatur circa ipsam Hyperbole, per 53 praecedentis libri, cuius transversa diametros coni secundum rectam lineam d e , quae ad bc sit perpendicularis, et faciat sectionem in super­
sit ed, recta vero h, ut ordinate ductae possint rectangula adiacentia ad ipsam h, cum latitu­ ficie coni dfe lineam, diameter autem sectionis fg aequidistans sit uni laterum trianguli per
dine sub ipsis ad summitatem receptis, et excedentia specie simili ei quod sub de, h rectan- axem, videlicet ipsi ac·, atque a puncto / lineae fg ad rectos angulos ducatur fh, et fiat ut
gulo; quae Hyperbole sit tdk, transiens iam per datum punctum d. Cumque dz sit ipsi ab quadratum bc ad rectangulum bac ita linea /i/a d /a , sumatur praeterea in sectione quodlibet
parallelus, etgz, za aequales erunt, per 2. sexti Eucl., et bd, dg aequales, quare quadratum punctum A:, et per k ducatur kl ipsi de aequidistans. Dico quadratum kl rectangulo hfl aequale
bg quadruplum erit ad quadratum gd, et ideo quadratum gd quadrans speciei sectionis, quae esse. . . . (14r). . . . Vocetur autem huiusmodi sectio parabole, et linea hf iuxta quam
scilicet subde, h lateribus continetur. Igitur per primam huius, ipsae ba,ag sunt non tangentes possunt quae ad fg diametrum ordinatim applicantur, quae quidem etiam recta appel­
ipsius tdk descriptae Hyperboles. Et perinde factum est quod faciendum proponebatur.” labitur. . . .
When these passages are compared with those in note 1 above, it is evident that Maurolico “ (14v) Theorema XII. Propositio XII. Si conus plano per axem secetur, secetur autem
has kept fairly close to Memmo’s text when presenting the enunciations but rather freely et altero plano secante basim coni secundum rectam lineam quae ad basim trianguli per axem
changes the proofs, adding what seemed to him to make the proofs clearer. The figures for sit perpendicularis, et sectionis diameter producta cum uno latere trianguli per axem extra
Maurolico’s text are those of Memmo’s text. I have made the usual punctuation and capitaliza­ verticem coni conveniat, recta linea quae a sectione ducitur aequidistans communi sectioni
tion changes in Maurolico’s text, converted vocal “ v” to “ u” , and have once quoted a read­ plani secantis et basis coni usque ad sectionis diametrum, poterit spatium adiacens lineae
ing of the manuscript in the Escorial. Needless to say, the corollaries added to the ends of ad quam ea quae in directum constituitur diametro sectionis, subtenditurque angulo extra
Propositions 1.12 and 1.13 are Maurolico’s corollaries and not Apollonius’ and appear in triangulum, eandem proportionem habet quam quadratum lineae quae diametro aequidistans
italic type in the edition. Indeed throughout the whole text Maurolico has made a great many a vertice sectionis usque ad basim trianguli ducitur ad rectangulum basis partibus quae ab ea
mathematical additions. After studying the specimens given here and numerous other fiunt contentum, latitudinem habens lineam quae ex diametro abscinditur inter ipsam et
propositions, I think it doubtful that Maurolico had a Greek manuscript at hand when he verticem sectionis interiectam, excedensque figura simili et similiter posita ei quae continetur
produced his version. The many changes he made in Memmo’s text seem to have been dic­ linea angulo extra triangulum subtensa et ea iuxta quam possunt quae ad diametrum ap­
tated by mathematical considerations rather than philological ones. plicantur. Vocetur autem huiusmodi sectio hyperbole. Sit conus cuius vertex a punctum
12 Compare my remarks on Maurolico’s versions of the works of Archimedes in Archi­ [Fig. 7.6], basis circulus bc, et secetur plano per axem, quod sectionem faciat triangulum
medes in the Middle Ages, Vol. 3, pp. 749, 788-812 (passim). All of the texts of Maurolico abe, secetur autem et altero plano secante basim coni secundum rectam lineam de ad bc
which I have published in that volume demonstrate a style determined by mathematical basim trianguli abe perpendicularem, faciatque sectionem in superficie coni lineam dfe, et
rather than textual considerations. sectionis diameter/)? producta cum ipso ac latere trianguli abe extra coni verticem conveniat
318 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 319

(based on Greek manuscripts, but with some attention to Memmo’s trans­ Commandino’s version): Apollonii Pergaei Conicorum libri octo et Sereni
lation) became the standard version until the appearance of Edmund Hal­ Antissensis De sectione cylindri et coni libri duo (Oxford, 1710). (While the
ley’s edition of the Greek text (with Latin translations greatly influenced by Greek text prepared by Halley has been supplanted by Heiberg’s version,
Halley’s work remains an exceedingly useful version, containing as it does
in puncto h , deinde pero ducatur lineaak diametro aequidistans, quae secet b c , et ab/ducatur a Latin translation of the Arabic text of Books V -V II [not produced in a
/ ad rectos angulos ipsifg, fiatque ut quadratum ka ad rectangulum bkc ita hf linea ad lineam modern edition except for a short piece published by L. M. L. Nix; see
/ · Sumatur autem in sectione quodlibet punctum m, et per m ducatur mn aequidistans de, above, Chap. 1, end of note 5] and a reconstruction of Book VIII). It will be
et per n ip s i/ aequidistans ducatur nox\ postremo iuncta hl et ad λ: producta, per Ix ipsi fn noted that I have included sample propositions from all three of the six­
aequidistantes ducantur Ιο,χρ. Dico lineam mn posse (15r) spatiumfx, quod quidem adiacet
lin e a e /, latitudinem habens fn, exceditque figura Ix simili ei quae hfl continetur. . . .
teenth-century versions of the Conics in notes 1,11 and 14, though it is not
Dicatur autem huiusmodi sectio hyperbole, et linea//iuxta quam possunt quae adfg ordinatim my intent to study the later fortunes of Apollonius’ text. I have done this so
applicantur, quae quidem etiam recta appellabitur, transversa vero hf. . . . that the reader will be able to follow the varying vocabulary and techniques
“ Theorema XIII. Propositio XIII. Si conus plano per axem secetur, et secetur altero that existed in the Renaissance treatment of conic sections. Incidentally, at
plano conveniente utroque latere trianguli per axem, quod neque basi coni aequidistet the time of the appearance of Commandino’s version of the Conics Fran­
neque subcontrarie ponatur, planum autem in quo est basis coni et secans planum conveniant
secundum rectam lineam quae sit perpendicularis vel ad basim trianguli per axem vel ad eam
cesco Barozzi mentions that he had found Memmo’s version defective and
quae in directum ipsi constituitur, recta linea quae a sectione coni ducitur aequidistans com­ that he had gone through it to correct it before Commandino’s version had
muni sectioni planorum usque ad diametrum sectionis, pote(15v)rit spatium adiacens appeared and that he used the same Greek manuscript as had Commandino
lineae ad quam sectionis diameter eam proportionem habeat quam quadratum lineae diametro (see below, Text B, Extract IV, marginal page number 92). Be that as it
aequidistantis a vertice coni usque ad trianguli basim ductae habet ad rectangulum conten­ may, Barozzi never published his corrected version, though, as we shall
tum basis partibus quae inter ipsam et rectas trianguli lineas interficiuntur, latitudinem habens
lineam quae ex diametro ab ipsa abscinditur ad verticem sectionis, deficiensque figura simili
see, he made considerable use of the Conics in Admirandam illud geo­
et similiter posita ei quae diametro et linea iuxta quam possunt continetur. Dicatur autem metricum problema.
huiusmodi sectio ellipsis. Sit conus cuius vertex a punctum [Fig. 7.7], basis circulus bc, Despite the publication of Memmo’s translation in 1537 the medieval tra­
et secetur plano per axem, quod sectionem faciat triangulum abc, secetur autem et altero ditions of conic sections continued to exert their influence until the 1560’s.
plano conveniente cum utroque latere trianguli per axem, neque basi coni aequidistante neque Indeed two years before the appearance of the Memmo translation, the
subcontrarie posito, quod faciat sectionem in superficie coni lineam de, et communis sectio
plani secantis atque eius in quo est basis coni sit fg perpendicularis ad bc, diameter autem
first edition of W itelo’s Perspectiva was published, thus making available
sectionis e d , et ab e ducatur eh ad ed perpendicularis, perque a ducta ak ipsi ed aequidistante, the material we have examined in detail in the third chapter above: Vitel-
fiat ut quadratum ak ad rectangulum bkc ita de ad eh, sumatur praeterea in sectione punctum lionis mathematici . . . πβρϊ οπτικής, id est de natura, ratione, et projec­
/, et per / ipsi fg aequidistans ducatur Im. Dico Im posse spatium quod lineae eh adiacet, tione radiorum visus, luminum, colorum atque formarum, quam vulgo
latitudinem habens em, deficiensque figura simili ei quae deh continetur. . . . Vocetur perspectivam vocant, librix (Nuremberg, 1535). (This edition— and its re­
autem huismodi sectio ellipsis, et linea eh iuxta quam possunt quae ad diametrum de ordinatim
applicantur, quae quidem et recta vocabitur, ed vero transversa. . . . print of 1551— was later displaced by the Risner 1572 edition of the works
“ (44v) [Lib. II.] Problema I. Propositio IIII. Datis duabus rectis lineis angulum con­ of Alhazen and Witelo which I have so often cited in Chapter 3 above
tinentibus et puncto intra angulum dato, describere per punctum coni sectionem quae hyper­ [see note 5].) Then in 1548 Antonius Gogava Graviensis published more of
bole appellatur, ita ut datae lineae ipsius asymptoti sint. Sint duae rectae lineae ab, ac [Fig. the medieval tradition of the conics in his work: Cl. Ptolemaei Pelusiensis
7.8] angulum ad a continentes, sitque datum punctum d, et oporteat per d circa asymptotos mathematici operis quadripartiti in Latinum sermonem traductio: Adiectis
bac hyperbolen describere. Iungatur ad, et ad e producatur, ita ut da sit aequalis ae, et per d
ipsi ab aequidistans ducatur df, ponaturque q/ aequalis f c , iuncta vero cd producatur ad b, et
libris posterioribus Antonio Gogava Graviens. interprete. . . . Item, De
quadrato cb aequale fiat rectangulum ex de et g. Deinde producta ad circa ipsam per d hyper­ sectione conica orthogona, quae parabola dicitur; Deque speculo ustorio,
bole describatur, ita ut applicatae ad diametrum possint rectangula adiacentia lineae g, ex- libelli duo, hactenus desiderati, restituti ab Antonio Gogava Graviensi
cedentiaque figura ipsi deg simili. Quoniam igitur aequidistans est rf/ipsi ba et cf aequalisfa, (Lovanii, 1548). The first of the added texts (the De sectione conica or­
erit cd ipsi db aequalis. Ergo quadratum cb quadruplum est quadrati cd, atque est quadratum thogona) was a slightly modified edition of Regiomontanus’ version of the
cb aequale rectangulo de, g. (45r) Utrumque igitur quadratorum bd, dc quarta pars est
figurae quae lineis de, g continetur; quare ba, ac descriptae hyperbolae asymptoti sunt.”
Speculi almukefi compositio, and I have given the variant readings from
Again I have altered the punctuation (most often the colons to commas), italicized the let­ Gogava’s edition in my text of the Regiomontanus version in Chapter 5
ters marking points and magnitudes, and converted consonantal “ u” to “ v ” . I have omitted above (see Text C).
the marginal references to the Elements of Euclid and to prior propositions of the Conics. I The second of the added texts (the De speculo ustorio) was a shortened
have also left out the marginal majuscule that links the text of Prop. I.XI with the following
version of Alhazen’s De speculis comburentibus, comprising the first four
comment of Eutocius. Incidentally, I have omitted the comments of Eutocius to Propositions
I. 11 and 1.13 and Commandino’s own comments and additions to Propositions 1.12,1.13 and propositions of the medieval text. It deserves no detailed study; but I
II. 4. should note that in Proposition 1, the only proposition of interest to my
320 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 321

geometrical study of conic sections, the editor confuses the lettering in the The culminating point of the Renaissance texts that stayed strictly
proof of the third case of the proposition, substituting the letter / (already within the medieval Latin traditions of conic sections was the appearance
used by him and by Alhazen for the latus rectum) for the letter m found in of Oronce Fine’s version of the Speculi almukefi compositio (with sup­
the original tract.15 There are other errors and omissions. Gogava every­ plementary material from Alhazen’s De speculis comburentibus and
where employs parabola for the original tract’s mukefi. Witelo’s Perspectiva), published with the title; De speculo ustorio, ignem
ad propositam distantiam generante, liber unicus. Ex quo duarum lin­
earum semper appropinquantium et nunquam concurrentium colligitur
15 Ed. cit. above in text, sign. S 2 verso— T 1 recto (following the figures 1.1-1.3 con­
structed for Chap. 1, n. 33): “ In omni sectione parabola cuius protrahatur sagitta et separatur demonstratio (Paris, 1551). I have given the complete text below (Text A),
ex extremitate sagittae quantum est quarta lateris eius erecti, omnis linea protracta aeque- but without an English translation, since I have already translated the
distans sagittae et proveniens ad sectionem et alia conversa ad punctum quod separat quartam original version in Chapter 4 above. I shall confine myself to an analysis of
continent cum linea contingente sectionem super illud punctum duos angulos aequales. Sit this work. Before doing so, I alert the reader to remarks I made in Volume
sectio parabola abg [Fig. 1.1], eius sagitta ad, latus erectum /, et secabo ex ad lineam ae
Three of Archimedes in the Middle A g es, pp. 1174, 1209-22, concern­
aequalem quartae lineae /, et producam cb invicem ipsi d a , et coniungam b e , ducamque kbh
contingentem. Dico angulum cbk aequalem esse angulo ebh. Sit primum angulus beh acutus ing Fine’s deficiencies as a geometer. He fares somewhat better in the
[Fig. 1.1]. Pono quod angulus cbk sit aequalis angulo beh; et quia cb et da sunt parallelae, De speculo ustorio, partly because he scarcely goes beyond his sources
erit angulus cbk aequalis angulo bhe. Sed angulus cbk est aequalis angulo hbe, ergo angulus in proposing demonstrations and partly, I suppose, because the geometry
ebh angulo bhe, ergo linea be lineae eh, et quadratum be quadrato eh; ductaque bz per­ of the parabola is relatively simple. Nevertheless Fine does commit some
pendiculari sagittae erunt 2 quadrata e z , zb aequa quadrato eh. Sed quadratum bz est aequale
geometrical errors in his Proposition VIII that are similar to errors found in
ductui az in /, sicut probavit Apollonius; ergo quadratum ez et multiplicatio az in / sunt aequa
quadrato ea (/ eh). Sed ea est quarta /; ergo quod fit sub za et ae [quater] aequum ei est quod Conclusion 10 of the Speculi almukefi compositio, as we shall see. I also
sub za et l. Ergo per secundam secundi quod sub za et ae cum quadrato ez aequum quadrato should note here that in this account I have followed Poulle in adopting the
eh; ergo ah est aequalis za. Sed est bh contingens et bz secundum ordinem. Dico rursus an­ spelling Fine instead of Fine.16
gulum cbk aequalem angulo ebh. Ducta enim bz secundum ordinem, quoniam bh est con­
tingens, erit az aequalis a h ; ergo ea in az quater cum quadrato ez aequatur quadrato eh. Sed munis; manebit ze aequalis Ih (mh); sed ea est aequalis al (! am)·, ergo za ipsi ah. Sed bh est
ea est quarta /; ergo multiplicatio / in za cum quadrato ez est aequalis quadrato eh. Verum contingens et bz secundum ordinem. Dico ergo rursus angulum cbk aequalem angulo ebh.
ductus / in za est quadratum bz, quia est se[Tl rec/o]cundum ordinem; ergo quadratum bz et Ducta enim bz secundum ordinem, et quia bh contingit, erit za aequalis ah, sitque al (! am)
quadratum bez (/ ez) sunt aequalia quadrato eh. Sed duo quadrata bz et ez sunt quadratum aequalis ae; manet ergo ze aequalis Ih (/ mh)·, ergo zl (f zm) aequalis est eh, et quadratum
e b ; ergo quadratum eb aequale est quadrato eh, et eb aequale eh, et angulus ebh angulo ehb. unius quadrato alterius. Sed quod ex za in ae quater cum quadrato ze est aequale [quadrato
Et quia cb aeque distat d a , angulus cbk aequalis est angulo ebh; ergo angulus ebh angulo cbk zm . Ergo quod ex za in ae quater cum quadrato ze est aequale] quadrato eh. Sed ductus za in
aequalis est, quod voluimus. ae quater est ductus za in /, cum ea sit quarta; ergo quod ex za in / cum quadrato ze est aequale
“ Sit angulus beh rectus, ut in secunda figura [Fig. 1.2]. Ponam igitur duos angulos aequales; quadrato eh. Sed quod ex za in / est quadratum bz cum bz sit secundum ordinem; ergo
et quia cb aeque distat ab (/ ad), est angulus cbq (/ cbk) aequalis angulo ehb. Sed angulus quadratum bz et quadratum ze sunt aequalia quadrato eh. Sed quadratum bz et quadratum
cbk per positionem aequatur ebh; ergo angulus ebh aequalis ehb; ergo linea be lineae bh,{! ze sunt quadratum be; ergo quadratum be est aequale quadrato eh; ergo be linea est aequalis
eh), et quadratum eb quadrato eh, sed quadratum eb ei quod fit ex ea in /, ergo multiplicatio eh lineae, et ergo angulus ebh aequalis est angulo ehb. Sed angulus ehb ipsi cbk aequalis est,
ea in / quadrato eh. Sed quod ex eh in / est quarta quadrati /, quoniam ea est 4 (/ 4a) /; [ergo cum linia (!) cb aequidistet liniae (/) dh; ergo angulus ebh est aequalis ipsi cbk. Idem contingit
quadratum eh est quarta quadrati /; ergo linea eh est medietas lineae /; sed ea est quarta/;] ergo in omni linea quae in sectione protrahitur invicem sagittae eius, quod cum conversa ad
linea ae aequalis lineae ah. Sed bh est contingens et be secundum ordinem. Dico ergo rursus punctum e continet cum linia (/) contingente angulos duos aequales. Atque illud est quod
angulum cbk aequari angulo ebh. Cum enim linea bh sit contingens et be secundum ordinem, proposuimus.”
erit [linea ae] aequalis ah; cumque ae sit quarta/, est eh semis/; ergo quadratum eh est quarta Once again, I have changed the punctuation somewhat (particularly many commas to semi­
quadrati /; sed quod fit ex ea in / est quarta quadrati /, quia ea est quarta /; ergo quod ex ea in / colons) and also rarely the capitalization. Furthermore, I have italicized the letters marking
aequatur quadrato e[T l verso]h sed ductus ea in / aequat[ur] quadrato eh; [ergo quadratum points and magnitudes. The diagrams in the edition are out of order (as the editor himself
eb est aequale quadrato eh;] ergo linea be lineae eh, et angulus ebh ipsi ehb; et quia cb et dh notes): the first is the third and the third the first. The reader should note that in the diagrams
sunt parallelae, est angulus cbk aequalis ehb; sed ebh angulus ipsi ehb; ergo angulus cbk ipsi to which I have referred him (i.e. Figs. 1.1-1.3) the letter t occupies the position of the letter
ebh aequalis est. c in Gogava’s edition. In the text of the third case, the editor has erroneously converted the
“ Sit tandem angulus beh obtusus [Fig. 1.3]. Ponam ebk aequalem ipsi ebh, per analysin letter m everywhere it appears in the original tract to the letter /, thus making letter / do double
videlicet; et ob cb parallelam dh, erit angulus cbk aequalis ipsi ehb', sed cbk per positionem duty both as a point and as a designation for the latus rectum. Furthermore, the editor has not
aequus est ipsi ebh; ergo ebh ipsi ehb, et linea eb lineae eh, et quadratum be quadrato eh, marked on the diagram his new letter /. It should be given as /' and put in the position of m on
ductaque bz secundum ordinem quadratum bz et quadratum ze sunt aequalia quadrato eh. Fig. 1.3. Needless to say, this omission would have greatly confused the reader who did not
Sed quadratum bz aequale est ei quod fit ex az in l·, ergo quod ex az in / cum quadrato ze est have a copy of the medieval text at hand. Finally, notice that, except for the analytic proof
aequale quadrato eh; sed ea quartae /; ergo quod ex za in ae quartae eius quod ex za in /; ergo of the third case, the editor has omitted all of Alhazen’s sentences that label each of the ana­
quod ex za in ae quater cum quadrato ze aequale est quadrato eh; et ponam al (! am) aequalem lytic and synthetic proofs (see Chap. 1 above, n. 33).
a e ; ergo quod ex za in ae quater cum quadrato ez aequatur quadrato zl (/ zm );ergo quadratum 16 See E. Poulle, “ Fine, Oronce,” Dictionary of Scientific Biography, Vol. 15 (New York,
zl (/ zm) quadrato eh aequale est quadrato (/ ergo zm aequalis est) eh; sitque el (/ em) com- 1978), p. 156, η. 1.
322 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 323

One might first suppose that Fine based his text on Gogava’s edition of Title, Preface, and Poem o f Tribute to the Author. By the time of the six­
the Regiomontanus version of the Speculi almukefi compositio published teenth century the commonplace expression for a burning mirror was
only three years earlier. But this does not appear to be so. For in the first speculum ustorium in place of the medieval speculum comburens. For
place, in his Proposition IX, Fine includes a description of the conditions example, this change was evident in Gogava’s edition of Alhazen’s De
of good steel that depends on the Appendix to Conclusion 10 of the original speculis comburentibus under the title of De speculo ustorio. It is also
version of the Speculi almukefi compositio, an appendix not included in the evident in Fine’s title. Fine begins his preface with the dedication of the
Gogava edition. In the second place, Fine nowhere mentions in his Prop­ work to the prominent English statesman and diplomat Sir John Mason
osition X that the curve whose asymptotic property is being proved is a (1503-66), who had been appointed King’s Scholar in Paris not long after
hyperbola, while the edition of Gogava so labels that curve. In fact, in en­ 1524-25 (see the article on Mason in the Dictionary o f National Biography).
titling that proposition as “ Propositio X. ex supradictis corollaria” Fine Fine, like Alhazen before him, stresses in the preface that it is the mirror
seems to imply that the desired curve is a parabola (a false inference he hollowed out in the form of the section of a right-angled right cone called a
could have drawn from certain of the manuscripts of the Speculi almukefi parabola which produces convergence of reflected rays at a single point and
compositio — see Chap. 4 above, variant readings to lines 104-109 and hence produces fire more quickly and intensely. Fine goes on to say that,
line 107 of Conclusion 10). Then in the second paragraph Fine leaves the some twelve years earlier (1539), out of the last demonstrations of the ninth
same impression when he says that “ while we were constructing the book of Witelo’s Perspectiva and out of the Conics of Apollonius, to­
above-said demonstrations of a parabolic section of a right-angled right cone gether with the Elements of Euclid, he (Fine) put together and demon­
there occurred to us a certain conceit (imaginatio) that ought not to be strated in a single book a device for describing a parabola and constructing
omitted, and which was once treated by certain people, namely one con­ a parabolic mirror that will burn at a certain distance.17 We are not to infer
cerning two lines such that, when constructed as well in the same plane as from this remark that Fine actually used the text of Apollonius’ Conics (for
in diverse planes, the more they will be produced the closer to each other he did not) but only that he used Witelo’s citation of Apollonius’ work. Fine
they will come, without however ever meeting even if they are produced in­ further says that Witelo has demonstrated “ in his manner” many things
definitely.” The implication seems to be that the parabola, with the proper­ concerning the parabola but has not taught ‘‘how the inflexed or curved line
ties already proved, would somehow provide the required curve and of the same parabolic section (on which it is clear the whole matter de­
straight line. Whether or not Fine made that false implication, he was cer­ pends) is to be described.” He continues (3r) by saying that he has seen
tainly silent concerning the name of the curve. only one author of uncertain name who treats of this matter in a work
If, then, it is apparent that Fine did use a manuscript of the Speculi al­ “ translated from the Arabic tongue in such a confused and involuted
mukefi compositio, we should inquire as to which one it was. While no way . . . that we could scarcely elicit any sense out of the text or discern
certain answer can be found, it seems likely that he had access to the now the single diagram which would correspond to the text.” This is no doubt a
lost manuscript that Fusoris consulted at the Sorbonne (see Chap. 5 above, reference to the Speculi almukefi compositio (which he is continually
Text A, Cap. 3). Before analysing the content of Fine’s version, I should paraphrasing) and particularly to Conclusion 10 of that work with its Fig.
note certain changes I have made in the edition of 1551 in presenting it as 4.12, which was presented in an erroneous way in some of the manuscripts.
Text A below. As in the case of Werner’s text in Chapter 6 ,1 present to the If indeed this is a reference to the Speculi almukefi compositio, then Fine
reader a text that reflects the meaning and intent of the author. To ac­ has erred (as did Heiberg and Weidemann some three and one half cen­
complish this, I have made some minor changes. I have eliminated the ac­ turies later) in assuming that the work was translated from the Arabic. It is
cents from a, e, and 6 when used in adverbs and prepositions (e.g., I give merely the Arabic terminology for conic sections adopted by the author
“ a” instead of “ a” for the preposition, “longe” instead of “longe,” from Gerard of Cremona (and from Witelo) that gives the tract its Arabic
“porro” instead of “porro,” etc.). Furthermore I have expanded “ g” to look. If the reader examines the preface of the Speculi almukefi com­
“ ae” , changed consonantal “ u” to “ v” and vocal “v” to “ u” . I have al­ positio (Chap. 4 above, Text) and my discussion of it, he will see that the
tered the punctuation which might confuse the reader, generally removing author was in fact some Latin scholar who depended on the early material
commas that stand between subject and verb and those that precede de­
pendent clauses. I have eliminated some decorative “boxes” in which 17 He speaks of the distance as being of such a length as the radius of some assumed circle.
proportions verbally stated in the text are repeated with lines drawn to con­ This would be the radius of a circle whose diameter is the whole axial segment between the
nect the terms of the proportions. These boxes add nothing to the under­ vertex of a parabola formed from a right-angled right cone and the intersection o f the parab­
standing of the text. Finally, I have corrected within parentheses Fine’s or ola’s axis with the axis of the cone. But Fine does not appear to be saying this. In Proposition
IX below Fine makes an erroneous identification of the aforesaid axial segment with the radius
the printer’s errors and I have occasionally added bracketed material that of the base of the right-angled cone from which the parabola is generated. That radius is
helps to make the author’s intent clear. actually V2~· radius of the cone's base. The whole statement here (2v) is puzzling.
324 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 325

translated from the Arabic. Incidentally, it is evident that Fine has freed tions (see Fig. 7.9) is a right-angled right cone is obvious and was not taken
himself from the Arabic terms, replacing linea munani by linea inflexa and from the earlier authors. Definition 6 describes the section of a right-
sectio mukefi by parabola. In his concluding remarks, he says that he angled right cone that is called a parabola as a plane surface contained by an
would have left the publication of his work to some better time “ if the “ inflexed line” (the term used by Valla and Werner) that lies in the surface
authority of certain worthy and zealous men had not finally ordered that it of the cone and by a diameter of the base circle of the cone, which plane
be published.” He leaves to the reader the judgment as to whether he has surface is at right angles to the plane of the axial triangle of the cone. Defini­
surpassed all others in this matter, a judgment, I regret to say, that must be tion 7 adds that the arrow (sagitta) or diameter (dimetiens) of the para­
a negative one, particularly when we compare Fine’s work with Regio­ bolic section is the straight line common to the plane of the parabolic sec­
montanus’ version of the Speculi almukefi compositio and with Werner’s tion and the plane of the axial triangle; it cuts one side of this triangle and is
Libellus. Following the preface is printed (3v) a poem in tribute to Fine by parallel to the other. Unlike the similar definitions in the Speculi almukefi
the physician and astrologer Antoine Mizauld (ca. 1520-78), who had compositio and Werner’s Libellus, Fine limits his definition to a parabola
studied with Fine.18 Mizauld celebrates Fine’s demonstration of the and does not include definitions of a hyperbola and an ellipse (presumably
production, use and matter of the burning mirror “ that produces fires because the principal objective of his work is to treat of the parabolic mir­
without fires and incinerates all things without tinder.” The worthy Fine, ror). If he had included here the definition of a hyperbola, it would have per­
we are told, goes Prometheus one better. While the latter dared to transport haps led him to identify the curved line in his last proposition as a hyper­
a flame from heaven to earth, the former draws out of the earth itself bola. Note finally that in Definition 7 Fine has used both sagitta and dimitiens
heavenly flames, “ which no one could do unless he be of the race of to name the axis of the parabola, while the author of the Speculi almukefi
the gods.” compositio had used axis and sagitta and Werner axis and diameter. The
Definitions and Postulates. Immediately preceding the definitions, Fine vertex of the parabolic section is described in Definition 8 as the “ highest
gives a short introduction not unlike that given by Alhazen in the De specu­ point of the arrow or diameter of this section.” The author of the Speculi
lis comburentibus (see Chap. 1 above, ns. 2 and 32). In this introduction almukefi compositio had called it the caput sectionis. In Definition 9 Fine
Fine says that in addition to using the Elements of Euclid, he supposes calls the base of the parabolic section its latus rectum (which he later
some things from the Conic Elements (though in fact he takes nothing distinguishes from the parameter latus erectum, as we shall see). Thus for
directly from Apollonius’ work, as I have already said). A similar statement him the latus rectum is the diameter of the base of the cone that terminates
was made by Witelo in his Perspectiva (see Chap. 3 above, n. 5). The first a parabolic segment. Other segments of the parabola are said to be “ in­
definition is essentially the Euclidian definition of a cone. It is similar to the creased” (auctae) or “ diminished” (diminutae) with respect to the orig­
definition given in Prop. 1.89 of the Perspectiva of Witelo (Chap. 3, n. 8), to inal segment (see Definition 10). It will be the “ diminished” or “ truncated”
Description 1 of the Speculi almukefi compositio (Chap. 4, Text), and to segments that he will be interested in when he describes his instrument
Diff. 1 of Werner’s Libellus (Chap. 6, Text). It differs from the others in for making a parabolic mirror (see Proposition VIII below). His use of the
limiting the definition to the right-angled right cone produced by the rota­ expression sectiones diminutae for smaller segments of the parabola
tion of an isosceles right triangle about one of the sides including the right might have been confusing to someone who knew only the “ Arabic”
angle. Like Werner (and Valla) he uses the term conus, which he identifies vocabulary for conic sections preserved by Gerard of Cremona, for we
with the expression rotunda pyramis (the customary medieval expression remember that Gerard had translated the Arabic expression for an ellipse
used by Witelo and the author of the Speculi almukefi compositio). Then as sectio diminuta (see Chap. 1 above, end of n. 21), and indeed the author
follow Defs. 2 -4 of the axis, the base, surface, and the apex of this cone, of the Speculi almukefi compositio had presented this expression for an el­
definitions not copied verbatim from the earlier authors but clearly de­ lipse in his fourth description. But Fine saved himself from this confusion
pendent on them. The assignment of the numbers 2 -4 to these definitions by omitting the definitions of the hyperbola and ellipse, as I have already
he seems to have taken from Werner’s work. Definition 5 is of the side of a said. Definition 11 defines lines of order (lineae ordinis) in the manner of
cone, which side (like Witelo) he calls the length (longitudo) of a cone, the Speculi almukefi compositio. But he also uses the expression lineae . . .
though he also calls it the side (latus), as had Werner. Recall that the author ordinatim extensae, which he probably took from Werner. Definition 12
of the Speculi almukefi compositio had called it the hypotenuse (ypotenusa) concerns the latus rectum or parameter of the parabola, which, as I have al­
of a cone. Fine’s explanation as to why the cone described in these defini­ ready said, Fine calls the latus erectum and distinguishes from the base
of the parabolic segment, which he called latus rectum. He was led to
adopting this distinction from the confusion mentioned by the author of
18 Biographie universelle (Michaud), new ed. (s.d.), Vol. 28, pp. 427-28; Enciclopedia
universal, Vol. 35 (Madrid, 1958), p. 1141; and J. H. Zedler, ed., Grosses vollstandiges
the Speculi almukefi compositio, who, we recall (see Chap. 4, Text, De­
Universal-Lexikon, Vol. 21 (repr. Graz, 1961), c. 654. scriptio 4a), noted that any line of order (that is to say, double line of order
326 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 327

or chord) could be called the latus rectum, but that this was not what Apol­ connects two points in the surface of the cone and which when produced
lonius meant by the expression. The Apollonian latus rectum was that does not pass through the apex must fall inside of the cone. Note once more
(double) line of order that passed through the midpoint of the axis (taken that the cone here described is a right-angled right cone, while in the earlier
from the vertex of the section to the intersection of the axis of the sec­ text, the type of cone is not specified.
tion with the axis of the right-angled right cone from which the parabolic Proposition II. This is equivalent to Conclusion 3 of the Speculi almukefi
section was produced). In describing the Apollonian latus rectum the compositio, with a basically similar proof that is considerably more prolix.
earlier author used both expressions latus rectum and latus erectum. But It shows that if a plane is drawn parallel to the base of the cone, the common
Fine thought that he might preserve the distinct meanings of latus rectum section of it with the surface of the cone is the circumference of a circle
by limiting latus rectum to the rectilinear base of the parabolic segment and whose center is on the axis of the cone. Fine adds a corollary to the effect
latus erectum to the Apollonian parameter. Like the author of the Speculi that the plane cutting the cone and parallel to its base produces a cone
almukefi compositio, Fine notes that the latus erectum is the same for any similar to the original cone. Again, Fine specifies that the original cone in
segment of the parabola. Definitions 6-12 are then illustrated with respect this proposition is a right-angled right cone, while the author of the Speculi
to Fig. 7.10. These definitions are followed by a lemma or assumptum to almukefi compositio specified it only as a right cone.
the effect that the common section of the surface of a right-angled right Proposition III. This is the first part of Conclusion 4 of the Speculi al­
cone and a plane through the apex and axis of the cone is a right isosceles mukefi compositio. It shows that the latus rectum of the parabola formed
triangle. This is equivalent to Conclusion 1 of the Speculi almukefi com­ from a right-angled right cone is double the axis, i.e., double the axial seg­
positio and to the third postulate of Werner’s Libellus, except that Fine ment between the vertex of the parabola and the intersection of the axis of
specifically limits the lemma to a right-angled right cone. the parabola with the axis of the cone. Though Fine’s proof is essentially
To these definitions and the lemma Fine adds four postulates taken from like that found in the Speculi almukefi compositio, since it too is based on a
optics, and in doing so he is following the procedure of the De speculis fundamental property of the circle (namely that a perpendicular drawn
comburentibus of Alhazen.19 The first underlies all of geometrical catop­ from the circumference to the diameter is the mean proportional between
trics and declares (5v-6r) that “ solar rays incident to any given surface of a the segments of the diameter produced by the perpendicular), still Fine
mirror are as certain straight lines and accordingly in geometrical demon­ does not proceed to a proof of the axial property of the parabola (namely
strations they have the same force that mathematical lines have to each that the squares of any two lines of order are related as the axial intercepts
other.” 20 The second is the law of reflection for a plane mirror,21 while the of these lines of order). For Fine’s proof it was essential to prove that mh
third extends the law to any convex or concave mirror.22 The fourth merely = 2 hn (see Fig. 7.4), which demonstration he produced as a lemma follow­
says that if rays are reflected from any surface of a mirror so that they coin­ ing the proof. Note that Fine has omitted the exceedingly useful alternate
cide in one point, it is possible that fire be generated in this point alone. A statement with which the author of the Speculi almukefi compositio con­
corollary adds that when the solar rays incident to the surface of any con­ cluded his proof at the first part of the proposition, namely that the latus
cave mirror are reflected to one certain point from all directions, then it fol­ rectum is four times the semiaxis (i.e. four times its axial intercept). Once
lows that such a mirror, among all burning mirrors, produces the quickest more we see that the medieval proof is more economical than Fine’s
and most intense combustion. Such a mirror is the parabolic mirror that he demonstration.
will treat. Now Fine says that he will pass on to some properties of a para­ Proposition IV. This is the second part of Conclusion 4 of the Speculi al­
bolic section that will be useful for the mathematical understanding of the mukefi compositio, namely that a line of order is the mean proportional be­
proposed mirror. tween the latus rectum and the line of order’s axial intercept. Again the
Proposition I. This is equivalent to Conclusion 2 of the Speculi almukefi basic similarity between the proofs of the two works is obvious, though
compositio. Though considerably paraphrased from the earlier text, it fol­ once more Fine’s is unnecessarily prolix. As Corollary I Fine gives the
lows both lines of proof suggested there to show that a straight line which Apollonian form of the proposition, namely that the square of the ordinate
is equal to the rectangle composed of sides equal to the latus rectum and
19 J. L. Heiberg and E. Wiedemann, “ Ibn al Haitams Schrift iiber parabolische Hohl- the line of order’s axial intercept. This form of the proposition was given by
spiegel,” Bibliotheca mathematica, 3. Folge, Vol. 10 (1910-11), p. 220. Witelo as Proposition IX.40 of his Perspectiva. In the course of his com­
20 Ibid., radix 1. Cf. Witelo’s Perspectiva, ed. of F. Risner (Basel, 1572), pp. 61, 63 - 64. ment Witelo had made the identification of this form of the proposition with
21 Alhazen, De speculis comburentibus, ed. cit. in n. 19 above, p. 220, rad. 2. Cf. Witelo, the mean proportional form (see Chap. 3 above, n. 41).
Perspectiva, ed. cit. in n. 20 above, p. 195, Prop. 10.
Fine’s Corollary II to Proposition IV is equivalent to the first part of Con­
22 Alhazen, De speculis comburentibus, ed. cit., p. 220, rad. 3; see also radices 4 -7 that
detail the geometry of these angles. Cf. Witelo, Perspectiva, ed. cit., p. 200, Prop. 20. clusion 5 of the Speculi almukefi compositio, with an appeal to the same
328 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 329

propositions of Euclid’s Elements, It shows that if a circle is drawn through angled right cone. Fine has chosen his form of the statement to bring it in
the vertex of a parabola and the extremities of a line of order (that is, a line with his Proposition III. Fine’s Corollary III merely reiterates the
double line of order or chord), the center of the circle lies on the axis of the point made in his preface, that the parabolic mirror would produce stronger
parabola. Fine’s Corollary III is equivalent to the second part of Conclu­ and quicker combustion, for no other mirror, he says, would reflect all of
sion 5 of the Speculi almukefi compositio. It asserts that that segment of the the solar rays to a single point.
diameter of the circle drawn in Corollary II which is intercepted between Propositions VIII-IX. These propositions embrace the content of Con­
the terminus of the line of order and the circumference of the circle in the clusion 10 of the Speculi almukefi compositio and its appendix. That is,
direction of the base of the section is equal to the latus rectum of the section. they concern the making of an instrument by which the parabolic mirror
Proposition V. This is Conclusio 6 of the Speculi almukefi compositio, can be produced and the making of the mirror itself. Before starting with
with the same proof that depends on Fine’s Proposition IV, for hm is not the material of Conclusion 10, Fine outlines briefly the geometry involved
only the mean proportional between mn and ml by the corollary to Prop. in describing the parabolic section and its “ diminished” segment that will
VI.8 of the Elements but is also, by Fine’s Proposition IV, the mean pro­ produce burning at a given distance (see Fig. 7.25). This distance is doubled
portional between the latus rectum efg and ma (see Fig. 7.17). Thus efg / to produce the axis (ac) of the desired section and also (he says falsely)
mn = Im / ma, as the proposition asserts. the radius of the base of the right-angled right cone from which the section
Proposition VI. This is the converse of Conclusion 7 of the Speculi al­ and its diminished segment may be produced. (This latter is false because
mukefi compositio (and accordingly also the converse of 1.35 of the Conics the radius is in fact V 2-ac.) A perpendicular de is constructed to ac such
of Apollonius), for it says that if from a point in the parabola a straight line that da and ae are each equal to a c . Then dc and ce are drawn to complete
is drawn perpendicular to the axis (see e f in Fig. 7.18) and another line the isosceles right triangle dee, with angle dee the right angle. When tri­
(eg) is drawn from that point until it meets the axis in such a point that angle dee is rotated about dc a right-angled right cone is produced from
a f - ag, then line eg is tangent to the parabola at e. Fine presented the which the desired parabola with axis abc is to be produced. Then dc is the
converse of Conclusion 7 because he had noted that the proof of Conclu­ axis of the cone and ce the radius of the cone’s base. Then we draw fg
sion 7 shows that the proposition is true because its converse is true. At any through point b at right angles to a b c ,fg being equal to de, and b f and bg
rate, the proof is drawn directly from the Speculi almukefi compositio, with each being equal to abc. Line hi is also drawn at right angles to abc with
the usual changes of letters, style and length of proof. This having been each of its segments he and ci equal to dc (and thus to ce). Henc e fg is the
proved, the actual proposition equivalent to Conclusion 7 of the earlier text latus rectum of the parabola and hi is the base of the parabola hfagi. So
is then presented by Fine as a corollary and its proof is said to be by con­ much for the whole section and its segment fa g , either of which could be
versa demonstrandi ratione. used to make the mirror. But perhaps the mirror constructed from the
Proposition VII. This is Conclusion 8 of the Speculi almukefi com­ whole parabola hfagi, or even from its segmentfa g , would be too large for
positio (and thus is equivalent to Proposition 1 of the De speculis com­ practical construction. Hence Fine wishes to describe a smaller segment of
burentibus of Alhazen). Fine’s proofs of the three cases come directly from the parabola, a segment he calls the sectio parabola diminuta, that will
the Speculi almukefi compositio, and hence they do not contain Alhazen’s still reflect the rays to focus b . This will be the smaller segment with axis
proofs by analysis, which had been omitted by the author of the Speculi al­ ak. (He errs by saying that the base of that diminished segment of the parab­
mukefi compositio. I hardly need say once more that Fine’s style is more ola is the chord Im of the circular segment aim which is described by line
expansive than that of the medieval author. abc. It should rather be the chord of circular segment whose radius is one
To this proposition Fine adds as Corollary I a proposition equivalent to half of a line equal to ak + 2ac.) With such a construction, axis ak could be
Proposition 3 of the De speculis comburentibus of Alhazen and to Proposi­ only a foot or a foot and a half [if we assumed like the author of the Speculi
tion IX.42 of Witelo’s Perspectiva (see Chap. 3 above, n. 43). But Fine re­ almukefi compositio that the desired burning distance is to be 20 feet or
writes these accounts in terms of his own tract. The same can be said for some other large distance] and thus we would have a mirror of somewhat
Corollary II, which is equivalent to Alhazen’s Proposition 4 and to Witelo’s more manageable size. We may explain this by noting that if the required
Proposition IX.43 (ibid.). This proposition is also equivalent to the rather focal distance was 20 feet and ak was 1 foot, then we would have to hollow
truncated Conclusion 9 of the Speculi almukefi compositio. Fine says that out a mirror to the depth of 1 foot instead of one to the depth of 40 feet were
the focal distance of the parabolic section according to which the mirror we to use the whole section. Even so, numbers like this are merely used as
has been formed is equal to half of the axis of the parabolic section, while examples. For if the paraboloid constituting the surface of the mirror had a
Alhazen and Witelo had said that the focal distance from the vertex is one height of 1 foot and a focal distance of 20 feet, the radius of the base of the
quarter of the latus rectum of the parabolic section. Needless to say, these paraboloid would still have to be almost 9 feet (i.e., r2 = depth Tatus rec­
statements are equivalent when the section has been formed from a right­ tum = 1-80)!
330 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 331

At this point Fine begins his paraphrase of the account given in Conclu­ certainly Witelo’s Perspectiva continue to be cited in the context of the
sion 10 of the earlier text concerning the construction of the tool by which discussion of parabolic mirrors. For example, we should note that Giam­
the mirror may be made. Then at the end of Proposition VIII, under the battista della Porta used Fine’s tract (and possibly also either the Gogava
rubric Idem aliter, Fine presents another way to describe his diminished edition of the Speculi almukefi compositio or Witelo’s Perspectiva, or
parabolic segment. This is precisely the method used in Element XI of both) in his short treatment of the parabolic mirror in his first work: Magiae
Werner’s Libellus (see Chap. 6 above, Text), which, as I have already naturalis sive de miraculis rerum naturalium libri ΠΠ (Naples, 1558).23
noted, was essentially the application of the converse of Conclusion 5 of But the spreading knowledge of Apollonius’ Conics began to alter the treat­
the Speculi almukefi compositio. It deserves no extended comment. I ments of conic sections so that little was left of the medieval traditions in
merely note that, like Werner, Fine says in his corollary that the more such treatments. A case in point is Francesco Maurolico’s De lineis horariis
points taken on the axial segment be (see Fig. 7.29), the less erroneous liber tertius, which he completed in his monastery of S. Maria a Partu on 19
will be the parabola drawn. Proposition IX completes the discussion of the
July, 1553.24 I have spoken earlier of his version of the Conics of Apol­
instrument that will hollow out the paraboloidal surface and adds a para­ lonius. Though Maurolico had knowledge of the medieval traditions, at
phrase of the “ Conditions of a Good Steel’’ that was appended to Conclu­
least in the form of Alhazen’sD e speculis comburentibus, which in 1540 he
sion 10 of the Speculi almukefi compositio. Fine includes another tech­
lists as a work of Archimedes,25 and later in his Index lucubrationum
nique of preparing a mirror by fashioning a paraboloidal mold (typus) from
which a metallic mirror may be cast and he comments on how the surface
of the mirror is to be polished. In Appendix I to Proposition IX Fine re­ 23 Bk. IV, p. 151: “ De Parabolae segmento, et de speculis aliis ustoriis Caput XV. . . .
marks, as indeed had Alhazen in the De speculis comburentibus (see Scito Paraboleam sectionem inter caeteras maxime perure, radios enim violentius congregat,
et in unum coarctat. Sectio haec rectangula vel parabola dicitur. Ut vere cupientibus pateat
Chap. 3 above, n. 43, end of Prop. 5), that the parabolic section from which
fabrica, speculum tale conficiendi cape normam: Sit distantia nota, ad quam radios emit­
the mirror is to be made may not include the vertex of the section, i.e., an tere et comburere quaeris, eam dupla, et tali diametro rectangula rotunda pyramis erigatur:
annular mirror may be produced (see Fig. 7.33). Finally, in Appendix II, sed commoditatis gratia e cera, vel creta fiat, inde plana superficie axi (/ lateri rotundae
Fine observes that “ out of metallic and fusible material, and a scarcely dif­ pyramidis) equidistans ( / equidistanti) pyramidis abscindatur portio, et parabolica dicetur: si
ferent procedure of polishing, could be made any other given mirrors, versus fastigium in amblygonio hiperbolem, versus basim in oxigonio ellipsim habebis, nos
parabolen quaerimus. Sit tamen portio minor, ut facilius excavetur speculum, et mirabilior
whether plane or curved and hollowed out.” Thus ends the main part of
res videatur: abscissae portionis superficiem, plana tabula, vel ferri lamina ut inferius
the treatise. docebitur, figuram imitaberis, inde per fastigium, et basis medium axem, inducito fixam,
Proposition X. Like the author of the Speculi almukefi compositio, Fine et materiam aliquam excavabis chalybeam, vel ferream, vel (si mavis) mixturam aliquam,
could not refrain from adding the proposition on the two lines that approach qua parabis archetypum: tale sic efformatum speculum parabola dicitur, et secundum datam
each other as they are extended but never meet, even if extended in­ distantiam violenter Soli appositum, ut ipsius axis sit contra Solem, comburet: incidentes
enim radii speculo, et perpendiculariter ferientes, simul in unum reflectuntur et coincidunt,
definitely. He takes this proposition directly from the medieval text, giving
ut omnes unius radii vicem gerant comburentis multum: ideo nulla est inter sectiones alia,
two of the three cases presented there, namely the first one where the quae longius, et valentius ignem immittat quam parabola.” The principal source is evidently
curved line and the straight line are on the surface of a cone but are not Fine’s De speculo ustorio, Prop. 9. Note, for example, that della Porta has copied Fine’s
in the same plane and the second case where the two lines are in the same error of taking the focal distance (i.e., the burning distance from the vertex of the parabola)
plane. Omitted is the case where the two lines are both hyperbolas ap­ and doubling it to produce the radius of the right-angled cone from which the parabola is
produced. There is another rather curious error in the passage I have quoted above, when
proaching the same asymptote. The reader will notice that I have made
the author speaks of fashioning a right-angled cone of wax or chalk and cutting it by a plane
some minor corrections of and additions to Fine’s text of the first case. parallel to the axis of the cone to produce a parabola. It should say that the plane is to be
I have already commented on the fact that Fine does not identify the curved parallel to the side of the cone, for the author goes on to say that if it meets such a side toward
line in these cases as a hyperbola. This is surprising in view of the fact that the apex of the cone a hyperbola is produced or if it meets it toward the base of the cone an
he appears to have read Werner’s Libellus. ellipse is produced. The generation of the three conic sections in this manner he could have
drawn from the Gogava edition of the Speculi almukefi compositio (descriptio 3a) or from
It should be evident from the foregoing summary that Fine’s text is not
the latter’s source, namely Witelo’s Perspectiva, Prop. 1.98, or both. That he read the
particularly original, at least so far as the mathematical treatment of the Perspectiva seems likely since he uses Witelo’s expression for the parabola, sectio rectan­
parabola is concerned. It is indeed but a thinly veiled paraphrase of the gula, at the beginning of the passage. Note also that della Porta uses the medieval expres­
Speculi almukefi compositio, with additions from Alhazen’s De speculis sion for a cone, rotunda pyramis. A much longer account of the parabolic mirror was in­
comburentibus and citations to Witelo’s Perspectiva. We have seen that cluded in the greatly augmented edition of the work: Magiae naturalis libri xx (Naples,
1589), Bk. XVII, Chap, xiv-xvii, which shows the influence of Fine’s work in even greater
Fine tends to embroider the older texts and to add additional references to detail.
Euclid’s Elements. With Fine’s text, the medieval tradition of the parabolic 24 M. Clagett, “ The Works of Francesco Maurolico,” p. 159.
mirror is substantially completed, though of course Fine’s work itself and 25 Ibid., p. 176, item [21],
332 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 333

Maurolici as a work of Ptolemy,26 the content of Alhazen’s work is not is equivalent to that of Proposition 1.35 of the Conics of Apollonius and to
reflected in his De lineis horariis liber tertius. In the latter work he remarks that of Conclusion 7 of the Speculi almukefi compositio, Maurolico’s proof
that he “ will put together some things that are partly excerpted from our is a direct proof rather than an indirect one, as were those of Apollonius and
Apollonius and partly demonstrated by us.” 27 F. Amodeo has lucidly sum­ the medieval author. In fact, I think that Maurolico’s proof may have been
marized this tract and stressed its originality.281 must reluctantly not treat influenced by the earliest direct proof of which I know, namely that of Ele­
this tract in detail, since most of it has no connection with the medieval ment XVI of Werner’s Libellus (see Chap. 6 above, Text). I think it likely
traditions of conic sections. However, Maurolico’s fifth rule for the parab­ that Maurolico’s proof reflects Werner’s proof of the first case (see Fig.
ola gives a description of the parabola like that of Werner’s Element IX, 6.17), where arc eg is a quarter of the circumference, though the later proof
a description which was based on the converse of Conclusion 5 of the differs from the earlier one at its end. In one respect Maurolico’s proof is
Speculi almukefi compositio .29 One further proposition of Maurolico’s may more general since it concerns a parabola generated from any cone, and I
have been influenced by the tradition that we have so far studied. This is give it here in toto'fi1
Proposition XII of Chap. 4, which is one of several of Maurolico’s proofs
that G. A. Borelli so admired.30 Though the enunciation of this proposition XI. If from any point in a diameter extended outside of a parabola a line is
drawn that is tangent to the curve, and from the point of contact an ordinate is
28 Ibid., p. 179, line 19. Cf. Maurolico’s Praeparatio ad Archimedis opera, completed in drawn to the diameter, then the intercepts of the diameter between the vertex
1550, in which he says that some attribute the work to Archimedes and others to Ptolemy (see of the section and the exterior point and [between the vertex and] the ordinate
my Archimedes in the Middle Ages, Vol. 3, p. 814, line 73). are equal.
27 F. Maurolico, Opuscula mathematica (Venice, 1575), p. 263: “ Quoniam igitur in super­ This is demonstrated as follows. In any cone let triangle hkt be the axial tri­
ioribus libellis vix species flexarum, vix pauca circa earum axes, ac delineationes tetigerimus,
angle [see Fig. 7.38]. Let the transverse diameter of the parabola be da and an
lineis tantum horariis describendis intentis, in hoc postremo libello percurremus aliqua circa
ordinate be dg, which by the lemma just premissed will be as a diameter of
conicarum sectionum diametros, ac proprietates, et contactus, necnon circa contrapositarum
Non tangentes, ut si quid remansit obscurum, hic apertius notescat. Perstringemus igitur
some circle or ellipse,32which circle or ellipse we let be hgt, and we let straight
aliqua ex Apollonio nostro partim decerpta, partim per nos demonstrata: ut quam facillime line gx be tangent to the periphery of hgt at point g. Then straight line gx, by
fieri poterit, flexarum huiusmodi diffinitiones, accidentia et proprietates praecipuae aperian­
tur: ab ipsis diffinitionum elementis exordium capientes.”
28 F. Amodeo, “ II trattato delle coniche di Francesco Maurolico,” Bibliotheca mathe­ Fr. Maurolicus Messanensis primo nitidissime libros 4 conicorum Apollonii exposuit, quin­
matica, 3. Folge, Vol. 9 (1908-09), pp. 123-38. The reader should notice Maurolico’s tum et sextum libros proprio marte adinventos adiunxit anno 1547. Deinde lib. 2 ( / 3) de lineis
original derivation of the latus rectum (ed. of 1575, pp. 265-66), which he calls the recta horariis anno 1553 breviarium conicorum composuit, in quo egregias demonstrationes
diametros. Referring to Fig. 7.36, we should observe that Maurolico shows that the latus rec­ excogitavit linearum tangentium sectiones conicas et asymptotarum, quas ob eorum sum­
tum is the fourth proportional eh in the following proportion: ed / dc = ad I eh. The key mam prestantiam duo viri praeclari Midorgius anno 1639, et Gregorius a S. Vincentio anno
axial property is given in the corollary to this rule, namely that the squares of the ordinates 1647 amplexati sunt, et iis sua opera exornarunt.” I have omitted accents and have changed
are to each other as the axial intercepts of the ordinates. consonantal u to v.
29 Ed. of 1575, p. 267: “ REGULA 5a. Ex datis denique paraboles diametris duntaxat pos­ 31 Ed. of 1575, pp. 275-76: “ XI. Si a puncto quopiam in diametro extra Parabolen ducta
sum delineare periferiam: ut exempli gratia, sit recta diametros parabolae ab [Fig. 7.37]. periferiam tangat, et a puncto tactus ordinata ducatur ad diametrum: tunc receptae a vertice
Axis vero, seu diametros, ac. Volo circa diametrum ac lineare parabolen cuius recta dia­ sectionis ex diametro ad punctum exterum et ad ordinatam sunt aequales. Quod sic demon­
metros sit ab. Ponam in rectam unam utranque ca,ab; et secta ac per sextum cap. praemissi stratur. Sit in cono quopiam triangulum per axem hkt [Fig. 7.38], in quo diameter transversa
libri in aliquot aequas portiones, utputa quatuor in puncti[s] d, e, f , describam super bc, bd, parabolae sit da, ordinata dg, quae, per immediate praemissum lemma, erit pro diametro
[be,]bfdiametros singulas singulos semicirculos bmc,bmd (! bnd),boe,bpf. Deinde a puncto alicuius circuli vel Ellipseos: qui circulus sive ellipsis sit hgt, cuius periferiam in puncto g
a excitabo ad rectos ipsi bc lineam am secans periferias in punctis m , n , o , p . In axe autem tangat recta linea gx, quae per primam harum conclusionum ordinata erit ad diametrum d g , et
seu diametro parabolae ac ut dictum est divisa in punctis d, c,/d u ca m ordinatas: Ic ipsi am perinde aequidistans diametro ht. Ducatur et in plano trianguli hkt per verticem k ipsi ht
aequalem, dk ipsi an aequalem, eh ipsi oa aequalem, et fg ipsi ap aequalem, singulas ad aliam aequidistans linea ke coincidensque diametro da apud e. Sic enim fiet ut ipsae ke, gx sint
partem ultra diametrum ac tantundem productas. Nam delineanda periferia parabolae ibit per aequidistantes et in plano positae, quod conicam superficiem tangit apud latus kg. Sola enim
a, g, h, k, l puncta et eorum correlativa ultima diametrum. Quae puncta quo plura fuerint, kg recta communis erit conicae (p. 276) superficiei, planoque praedicto in quo ke, hx (! gx).
eo certius delineabo per ea ductam periferiam leni fluxu et angulosae fracturae expertem. Quo fit ut recta linea in eodem plano connectens puncta e , g , et ulterius producta tangat in solo
Cuius operationis demonstratio est quod in semicirculis rectae am, an, ao, ap singulae puncto g conicam superficiem, et in eodem ipso puncto paraboles ag periferiam tangat in eius
possunt singula rectangula, quae possunt singulae ordinatae cl,dk, eh,fg in parabola, quibus plano iacens. Aequalis autem cum sit hd ipsi dt et ipsae hk, da aequidistantes: iam aequales
sunt aequales singulae singulis. COROLL. Unde manifestum est quod in delineatione para­ erunt ka, at, cumque ipsae ke, dt aequidistent, erunt et da, ae invicem aequales, receptae
bolae semicirculi ex quibus ordinatae eliciuntur sese contingunt apud extremum rectae scilicet a vertice sectionis a ad terminum tangentis e et ad ordinatam dg. Sicut fuerat demon­
diametri.” I have altered the punctuation and consequently some capitalization and made strandum. Et haec propositio est 35a primi conicorum.” I have made slight changes in the text
the usual changes of consonantal “ u” and vocal “ v” . I have omitted grave accents from in the same manner as in note 29 above.
some prepositions and adverbs, and I have italicized the letters marking points or magnitudes. 32 This lemma appears on p. 274 of the ed. of 1575: “Secundum lemma. Alterum lemma
30 G. A. Borelli, Elementa conica Apollonii Paergei (!) et Archimedis opera nova et erit hoc: Omnis ordinata in sectione conica est vel circuli vel ellipsis cuiuspiam diameter.”
breviori methodo demonstrata a loanne Alphonso Borellio (Rome, 1679), p. 2: “ Post Arabes Actually dg is shown in Fig. 7.38 as a semidiameter and hgt as a semicircle or semiellipse.
334 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 335

the first of these conclusions,™ will be perpendicular to diameter d# and thus Ghetaldi,36 Johann Kepler,37 and Bonaventura Cavalieri.38 One last topic
parallel to diameter ht. And in the plane of triangle hkt let there be drawn remains to be considered before concluding our account of the medieval
through vertex k line ke parallel to ht and meeting diameter da [extended] at e. traditions of conic sections. This concerns the flurry of activity generated
For let it be done in this way so that ke and gx are parallel and placed in a plane
which is tangent to the conical surface at side kg. For only straight line kg will
be common to the conical surface and to the aforesaid plane in which ke and gx 36 Marino Ghetaldi, Nonnullae propositiones de parabola (Rome, 1603). The principal
purpose of this small treatise is to show that any parabola formed from a right-angled right
lie. This is done so that a straight line in the same plane that connects points e cone (as used by Fine and Witelo) may be equally formed from any oblique cone, so that a
and# and is produced further is tangent to the conical surface in point# alone, burning mirror may be fashioned from a parabola produced from any cone. See p. 16: “ Col­
and in the very same point let it touch the curve of parabola ag, lying in its ligitur etiam omnes parabolas ad construenda comburentia specula esse idoneas. Demon-
plane. But since hd is equal to dt, and hk and da are parallel, hence ka and at stratum enim est ab Orontio et a Vitellione parabolas coni recti rectanguli ad constructionem
will be equal. And since ke and dt are parallel, hence da and ae will be equal to speculorum comburentium esse idoneas, sed parabola cuiuscunque coni eadem est quae coni
one another, that is, the intercepts from the vertex a of the section to (1) the recti rectanguli, ut prop. 4 demonstravimus, ergo omnes parabolae ad construenda com­
terminus e of the tangent and to (2) the ordinate dg [are equal], just as it was to burentia specula sunt ideoneae.” While the author in his propositions shows a thorough
be demonstrated. And this proposition is 1.35 of the Conics [of Apollonius]. knowledge of the medieval traditions, in the first five propositions (where he establishes the
equivalence of parabolas formed from a right-angled right cone and an oblique cone), he cites
the Conics of Apollonius plentifully in the margins.
None of the other propositions in which Maurolico diverges from Apol­ 37 Kepler appreciated the importance of the role of the focus in conic sections, and in
lonius seem related to the medieval traditions, either directly or through fact he introduced the term focus (with an original discussion of a second, blind focus in the
Werner, though to be sure Maurolico (like the author of the De duabus parabola) in his Ad Vitellionem paralipomena quibus Astronomiae pars optica traditur (Frank­
furt, 1604), pp. 92-96. In this passage he mentions Apollonius and the commentary of
lineis) develops a method of drawing asymptotes to hyperbolas (which he
Eutocius. His treatment is quite original and distinct from the treatment of conic sections
calls non tangentes contrapositarum) in relationship to the cones from found in medieval optical works. See also his Nova stereometria doliorum. . . . Accessit
which the hyperbolas are generated,3334 and he proves in a general way the Stereometriae Archimedeae Supplementum (Lincii, 1615), sign. [D 4 verso]-E [1 recto], for
same conclusion as Element XXII of Werner’s work but for any hyperbola similar material. The great emphasis put on foci in the modem treatment of conic sections
and not just for a right-angled hyperbola.35 is evident, for example, in P. de La Hire’s Sectiones conicae, (Paris, 1685), where the author
devotes the eighth book to the origins and properties of foci (p. 177: “ In hoc libro focorum
A considerable number of other authors after this time were well aware
ortus, et proprietates explanantur, ac demonstrantur. . . . MONITUM. De hoc puncto in
of the sixteenth-century tracts that reflected the medieval traditions. But Parabola tacuit Apollonius, cum huius sectionis axis determinatus non sit, neque figura, et
in their own treatises on conics or on the burning mirror they tended to rely simile punctum in Ellipsi, et hyperbola PUNCTUM EX COMPARATIONE FACTUM
to a significant extent on Apollonius. Examples of such authors are Marino dixerit, ex rectanguli quartae parti figurae aequalis ad axem applicatione, quadrato de­
ficientis in Ellipsi et excedentis in hyperbola; Neoterici geometrae UMBILICUM, ET
FOCUM tale punctum vocitaverunt.” ).
33 Ed. of 1575, p. 273: “ omnis recta linea tangens conicam sectionem apud extremum 38 See Bonaventura Cavalieri’s Lo specchio ustorio (Bologna, 1632). He mentions the
diametri ordinata est ad talem diametrum.” early literature on the subject in his preface (sig. A r-v): “ Et benche io sappi finalmente,
34 Ed. of 1575, p. 280: “ Quoniam Apollonius omnia fere conicorum demonstrata conatus che alcuni scorgendo a prima vista il titolo di SPECCHIO USTORIO, diranno questa esser
est in planum redigere, antiquioribus ingeniosior, neglecta conorum descriptione, et aliunde materia, della quale se ne sa hormai, quanto se ne puo sapere, havendo trattato pure de’
quaerens argumenta cogitur persaepe obscurius et indirecte demonstrare id quod con­ Specchi Ustorii Vitellione, Rogerio Bacconi [=Speculi almukefi compositio!], Orontio, il
templando solidae figurae sectionem apertius et brevius demonstratur. Id nos fecimus in Cardano, il G[h]etaldo, il Porta, il P. Gruemberger, il P. Biancano, che ne tocca un poco nella
4or postremis praecedentis capiti conclusionibus. Idem nunc de Non tangentibus contra­ sua Echometria, e finalmente il Magini, et altri, che con la loro esquisitezza di dottrina ci
positarum locuturi faciemus. Sunt enim Non tangentes duae rectae lineae se invicem in hanno insegnato tanto, che non lasciano luogo di poter dir piu cosa nuova intorno a simil sog-
centro contrapositarum hyperbolarum secantes et utrinque semper magis ac magis in in­ getto. A questo pero non rispondero altra, se non che si compiaccino questi tali di veder’ un
finitum periferiis approximantes, nunquam vero coincidentes. Et ob id Non tangentes sive poco tutto il Trattato prima, epoi, che giudichino, se lacosastacosi, come dal titolo gli paredi
Non coincidentes appellatae: de quibus Apollonius in 2° conicorum locutus est. Nos igitur poter’ a prima fronte congetturare.” The properties of the three conic sections as pertain­
huiusmodi linearum proprietates demonstraturi hoc praefabimur, duas hyperbolas in duobus ing to reflection occupy the first 70 pages. Then there is a great variety of subjects treated
rectis conis factas ac similibus triangulis per vertices conorum ductis aequidistantes similes from pages 71-224, including a long speculative account of the nature of the burning mirror
esse, ut in 6° conicorum lib. ostensum est. Unde omni propositae hyperbole similis ac etiam that Archimedes was supposed to have devised for the defense of Syracuse, an analysis of pro­
similis et aequalis collocari potest in aliquo recto cono, ut ibidem traditur. Ostendemus jectile motion (whose path he declares to be a curved line “ differing insensibly from a parab­
igitur hic lineas Non tangentes incedere per centrum contrapositarum, et complecti angulum ola,” pp. 164-65), various methods of describing the conic sections, etc. Notice that there is
aequalem angulo verticali trianguli cui planum hyperboles aequidistat.” one description (pp. 201-04) of the parabola in which we see Cavalieri employing Werner’s
35 Ed. of 1575, p. 283: “ Quod parallelogramma inter Non tangentes et periferiam locata technique of laying the latus rectum along the axis (starting from the vertex) and producing
sunt invicem aequalia, quodque tam tangentis sectionem a tactu quam secantis eandem a thereby the axial terminus of an ordinate equal to the latus rectum. The whole procedure is
periferia ad Non tangentes, recepta segmenta sunt aequalia. Caput 7 .” interesting (see Fig. 7.39). He takes AX as an indefinitely prolonged axis of the parabola to be
336 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 337

by Moses Maimonides’ mention of the asymptotic relationship of a hyper­ property is a case of something that has been demonstrated by the intel­
bola and its asymptote in Book I, Chap. 73 of his Guide o f the Perplexed, of lect, i.e. in Apollonius’ work, but it does not fall within the scope of the
which a medieval Latin translation became available in print under the imagination. The difficulty with the early translation noted above lies in the
title Dux seu Director dubitantium aut perplexorum (Paris, 1520). Follow­ fact that the Conics of Apollonius was obscurely rendered as De mensis
ing this translation literally we may render the pertinent passage as fol­ indisciplinabilibus (or De mathematicis disciplinis). Hence anyone de­
lows:39 pending on this translation, without a prior knowledge of Apollonius’
work, would not have had any inkling as to what book the author was
It has also been demonstrated in the second part De mensis indisciplina-
bilibus (mg. Mathematicis disciplinis; in fact Book II of the Conics of Apol­ referring.
lonius) that there are two lines between which there is a certain distance in the Just two years before the publication of this Latin text of Maimonides’
beginning, and the more they are extended [the more] the distance is di­ work the well-known humanist of Ferrara, Celo Calcagnini (1479-1541),
minished and they come nearer to each other, without however ever meeting, in a letter of 1518 to the German astronomer Jakob Ziegler mentioned this
even if they are extended indefinitely and each may approach the other. But problem and a demonstration of it that had been referred to in a letter to
this does not fall in the scope of the cogitation (/ imagination). And of these Celo’s prince (Ippolito d’Este).40 Calcagnini is puzzled and indicates that
two lines one is a straight line and the other is a curve, as will be explained Euclid appears to deny that it is possible, thus implying (and indeed Cal­
there [in the aforementioned work]. cagnini specifically states) that both the desired lines are straight lines (and
The context of this passage (which is obscured in the Latin translation) the text of Calcagnini’s letter is accompanied by a diagram showing both
concerns the difference between imagination and intellect. The asymptotic lines as straight lines). Hence Calcagnini is writing to Ziegler on behalf of
his prince to ask him what he thinks about it. We do not know who the
drawn. A will be the vertex, B the focus. Mark C such that BC = BA. On C draw a circle prince’s original correspondent was or to what proof he was referring. It
AOFZ with radius CA. It will cut axis AK in point F. Draw NFG as an indefinitely long per­
could be that the unknown correspondent was referring to the proof pres­
pendicular to AK through F. Mark N and G on this perpendicular so that NF and FG are each
equal to the latus rectum AF. Hence N and G are on the parabola. Here is where Werner’s ent in either the De duabus lineis or the Speculi almukefi compositio, or
method of laying the latus rectum on the axis enters into play. Then take any point E between that he was referring to his own proof which he devised after reading a
A and F and draw an indefinitely long line through E and parallel to NG. This cuts the circum­ manuscript of Maimonides’ work. At any rate, it is clear that neither Cal­
ference of circle AOFZ in points O andZ. Connect AO and AZ and transfer them to line MH cagnini nor his prince had any knowledge of the earlier medieval proofs or
so that ME = AO and EH = AZ. Then M and H will lie on the parabola. For (1) OA2 = OE2
of that in the Conics of Apollonius.
+ AE2, (2) OE2 = E F E A , (3) EA EF + EA2 = EA ■(EF + EA) = EA FA, and hence (4)
OA2 = ME2 = EA FA. So, by the basic property of the parabola, namely that the square of an There was some interest in the problem among Jewish commentators on
ordinate is equal to the rectangle contained by the axial intercept of the ordinate and the latus the Guide of the Perplexed.41 I shall limit myself to treating those com­
rectum, point M (and similarly point H ) will lie on the parabola whose latus rectum is AF. ments available to the main traditions of European mathematics. The first
Further, if we take any point K on the axis beyond F and construct a semicircle on AK as a
diameter, then its semicircle will cutJVG at point T. Draw A T and transfer it to an indefinitely 40 C. Calcagnini, Opera aliquot (Basel, 1544), p. 55: “ Mox quom de te ac studiis mathe­
long perpendicular through K so that LK = AT - KQ. Hence points L and Q will lie on the maticis rursus verba facerem, idem princeps doctissimi viri atque in illis gravioribus studiis
parabola since L K 2 = KA AF and KQ2 = KA AF, which equations satisfy the property of exercitati literas ostendit; quibus ille sibi gratulatur rationem probationemque nuperrime
the parabola which we noted above. With the same procedures other points on the parabola invenisse, qua sine fuco et reiecta omni sophistica contentione demonstrare possit: datis
may be determined. duobus lineis in plano, quae in infinitum portendantur (/), quarum altera recta sit, altera non
39 Rabi Mossei Aegyptii Dux seu Director dubitantium aut perplexorum (Paris, 1520), longo intervallo a priore dissita itidem, sed quae vergat inclinetque ad alteram: eam tem­
35v, Bk. I, Chap. LXXII (/, LXXIII; there are two chapters with the designation Chap. periem ac moderationem servari posse, quamvis utraque sine termino provehatur, ut nun­
LVIII): “ Demonstratum est etiam in secunda parte de +mensis indisciplinabilibus (marg.: quam invicem colliment, nec posterior priorem secet. Quod sane mihi admirabile videri in re
+Mathematicis disciplinis), quod duae lineae sunt in principio suo aliquantum distantes: et Geometrica affirmavi. Tum princeps iussit, ut quid tu de ea re arbitrareris, per literas ex­
quanto magis protrahantur diminuitur distantia, et appropinquant sibi, et tamen nunquam quirerem. Quam ut facilius intelligas, binas lineas in calce huius paginae subliciam, quales
concurrent licet protrahantur in infinitum, et utraque appropinquet alteri. Istud vero non cadit ille sibi effingit. Tu rem diligenter pensa, et rescribe. Videor profecto apud Euclidem nescio
in illam cogitationem, et illarum duarum linearum altera est recta: altera vero curva, sicut quid hac de re legisse: qui mihi videtur negare id fieri posse, quod doctus ille vir pollicetur. Sed
ibi explanabitur.” Cf. Moses Maimonides, The Guide o f the Perplexed, transi, of S. Pines tu aliquid certius compertiusque significa.” The letter is dated (p. 55): “ Quartodecimo
(Chicago, 1963), p. 210. Cf. F. Barozzi, Admirandum illud geometricum problema tredecim Calendas Novemb. XVIII.”
modis demonstratum (Venice, 1586), p. 253 (not extracted in Text B below): “ Similiter 41 It would be a worthwhile project for an hebraist to examine the various commentaries on
demonstratum est in libro secundo de Conicis, quod possunt in eodem plano exire duae the Guide of the Perplexed in order to trace how the allusion to the asymptotic relationship in
lineae, quae in principio sunt aliquantulum distantes ab invicem, et quanto magis protrahun­ Chap. 73 was treated. I have found reference to four authors who treat the problem: (1)
tur, diminuiter distantia, et appropinquant sibi: nec tamen invicem coniungantur, licet in Efodi ( = Issac ben Moses ha-Levi = Profiat Duran), second half of the fourteenth century, in
infinitum producantur, alteraque alteri appropinquet. Istud autem non potest excogitari, his comments on the Guide given in the Hebrew edition of the Guide (Sabionetta, 1553)— see
neque in Imaginationem cadit. Earum duarum linearum altera est recta, et altera curva, sicut Encylopaedia Judaica, Vol. 6 (New York, 1971), cc. 299- 300. This treatment of the problem
ibi declaratum est.” was translated and discussed by Francesco Barozzi, who mistakenly assigned it to Rabbi
338 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 339

treatment of the problem that became available to mathematicians without different volumes of the Revue des etudes juives: Vol. 27 (1893), pp. 91-105; Vol. 28 (1894).
a knowledge of Hebrew was that of Moise Provenzale (inappropriately 228-46; Vol. 29 (1894), 111-26. He discusses the tract on asymptotes in Vol. 28, 228-32,
called Moses of Narbonne by Cardano and Barozzi, as we shall see), a and gives his French translation of it in Vol. 29, 119-26. As I said above in the text, Proposi­
tions 4 and 5 of the De duabus lineis almost certainly provided the major source for Motot’s
Rabbi active at Mantua in the middle of the sixteenth century.42 His work work. Motot’s tract contains the following elements:
on the asymptotic problem appears to have been composed in 1549.43 At (1) p. 119. A statement of the asymptotic relationship to be shown. This statement is taken
any rate, it was translated into Italian by Giosepho da Padova (Joseph literally from Chap. 73 of Bk. I of the Guide o f the Perplexed of Maimonides.
Schalit ben Jakob) and published under the title: Opera nova di giometria (2) pp. 119-20. The demonstration of a proposition auxiliary to Proposition 4 of the De
in dimostrare come possano uscire due linee sopra una superficie le quali duabus lineis. It concerns [see Fig. 7.39a(A)] a right-angled right cone (and not just the right
cone of the De duabus lineis) cut by a plane perpendicular to the axial triangle ABC of the cone
proccedendo si accostino sempre ne poseno incontrarsi mai, cosa mera-
and parallel to the axis AD. The plane produces in the surface of the cone a curved line ERV
vigliosa e prodotta per verace e certa damo Mose d ’Egitto Heb. nel suo (which is in fact a hyperbola but which is not so designated by the author). The author demon­
Directio dubitantium a cap. 73 del primo composta per Magistro Moise strates that, if we erect on line EZ (the common line of the plane through HZ and the triangle
Provenzale Hebreo (Mantua, 1550). This translation was addressed to ABC) a perpendicular in the plane through H Z , which perpendicular extends from any point K
Don Diego Hurtado de Mendoza.44 Provenzale’s work consists of (1) a to point / on the surface of the cone, then, by (a) the property of a circle drawn through points
M , l , and L , namely the property IK 2 = MK KL, and (b) the equality of lines KL and K H , and
principal proof developed in the course of eighteen propositions and (2) a (c) the equality of lines MK and EK [(b) and (c) following from the fact that triangle ABC is a
brief final discourse with an alternate proof. The introductory paragraph right triangle with the right angle at A], H K E K = IK2. The same relationship can be shown
before the first proposition tells us that the problem was presented by in regard to any perpendicular erected on EZ. In this and the next two proofs the hyperbola is
“Rabbi Moses of Egypt in his famous work entitled Guide o f the Perplexed, generated as in the Speculi almukefi compositio, Conclusion 11, rather than as in Proposi­
in chapter 73 of the first [book].45 Provenzale does not, however, mention tions 4 and 5 of the De duabus lineis.
(3) pp. 120-21. The demonstration of a proposition equivalent to Proposition 4. With
by name the source of his demonstrations. But it is clear that the principal reference to the same diagram [Fig. 7.39a(A)], it shows that HK EK / IK 2 = HS ES / RS2,
proof in his tract is essentially like that of Propositions 4 and 5 of the De IK and RS being any two perpendiculars erected on line EZ in the manner of the preceding
duabus lineis (see Chap. 2 above). Provenzale could have read that tract. proposition. Note that before the proof Motot gives a literal translation of the enunciation of
But what is more likely is that he read the small treatise written in Hebrew Proposition 4 of the De duabus lineis. But the proof itself is tied to a hyperbola generated
on the subject by Simeon Motot, which itself follows the basic proof of the from a right-angled right cone as in (2). The proposition follows immediately from (2).
(4) pp. 121-24. The demonstration of a proposition equivalent to Proposition 5 of the De
earlier De duabus lineis.46 But Provenzale’s work is much more diffuse and duabus lineis. Its proof is once again related to the right-angled right cone and its hyperbola
produced as in (2) and (3). Like Proposition 5 of the De duabus lineis, it shows that curved
Samtou (=Ibn Shem-Tob, fl. ca. 1461-89), perhaps because the commentaries of the two line EV and straight line TS never meet [see Fig. 7.39a(B)]. It also shows that ID < YC
men are together in the Sabionetta edition of 1553 and Barozzi mistook the one for the other < FB < RA and consequently that the minimal distances are similarly related to each other,
(see G. Sacerdote, “ Le livre de l’Algebre et la probleme des asymptotes de Simon Motot i.e., that IZ < YK < EM < RL.
(suite),” Revue des etudes juives, Vol. 28 (1894), p. 229, n. 3.). (2) Simeon Motot, fifteenth (5) pp. 124-26. A fairly close, literal translation of Proposition 5 of the De duabus lineis.
century, who wrote a small tract on the asymptotic problem (see below, n. 46). (3) Solomon In this proof the cone is not specified as a right-angled cone and the hyperbola is generated as
ben Isaac (date?), who treated the problem in a brief work contained in a Hebrew manuscript in Proposition 5 of the De duabus lineis. That is, axial triangle ABC is assumed (see Fig.
in the British Library (Almanzi MS no. 213 = MS Add. 27, 107, ff. 81a-83a). See G. Margol- 7.39a(C)); sideAB is extended to some point D; line DZ is drawn to intersect AC; an indefinite
iouth, Catalogue o f the Hebrew and Samaritan Manuscripts in the British Museum, Part perpendicular (EH) is drawn to the plane of triangle ABC at point E; plane DHE is passed
III (London, 1915), p. 343. I have not identified this work and I leave any discussion of it to through DZ and EH and is thus perpendicular to the axial triangle. Plane DHE cuts the surface
someone more familiar with Hebrew texts. (4) Moses Abraham Provencal ( = Moise Proven­ of the cone in curved line ER (which is a hyperbola but which is not so designated by the
zale, 1503-75), who composed the tract which I now examine in detail. author). DE is bisected at Q. A specific length EH is taken such that DN NE / N M2 = QE2 /
42 Encylopaedia Judaica, Vol. 13 (New York, 1971), cc. 1258-59. EH2. A straight line is drawn from Q t o H and then extended indefinitely. This straight line
43 Margoliouth.op. cit. in n. 41, p. 343, in describing the MS cited in n. 41, mentions Proven­ QV is then the asymptote of the curved line ER. The proof of this proceeds as in Proposition
zale’s tract as contained on folios 84a-89a. It bears the date 1549 at Mantua. This may be an 5 of the De duabus lineis.
autograph. The Hebrew text was printed separately (Sabionetta, 1553?) and together with Since Sacerdote had no knowledge of the De duabus lineis (nor any of the Speculi almukefi
Maimonides’ Guide (Sabionetta, 1553), p. 1895, according to M. Steinschneider, Catalogus compositio), it is not surprising that he overestimated the originality of Motot’s work. I should
librorum Hebraeorum in Bibliotheca Bodleiana, 2nd ed. (Berlin, 1931), c. 1983. also add that though it is likely that Provenzale based his work on that of Motot (at least so
44 See my discussion of the activity of Don Diego Hurtado de Mendoza in my Archimedes far as Provenzale’s principal demonstration is concerned), we cannot be sure. In the Hebrew
in the Middle Ages, Vol. 3, pp. 549-50. version of Provenzale’s treatise he speaks of having studied a very short work, whose author
45 Ed. cit., 4r: “ . . . . cosa prodotta dal divo Rabi Mose d’Egitto nella famosa opra (!) he does not name (see Sacerdote’s study, Vol. 28, p. 230). If the work was not Motot’s, it
sua intitolata directio dubitantium, a cap. 73 del primo. . . .” could have been the De duabus lineis or even the tract of Solomon ben Isaac which appeared
46 For a French translation of Motot’s work, see the study by G. Sacerdote mentioned in n. in the same manuscript as Provenzale’s work (see above, n. 41). It is scarcely likely that the
41. In fact, Sacerdote’s long article on Motot’s two mathematical works appears in three confused comments of Profiat Duran influenced Provenzale.
340 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 341

geometrically unsatisfactory than either the De duabus lineis or Simeon steps (7) and (8) in Proposition 4 of the De duabus lineis. One would have
Motot’s work. Only occasionally are the enunciations of the propositions naturally expected step (9) at this point, namely that kn > hm. But this was
stated in general form. Usually they are given in terms of the magnitudes not specifically stated. Barozzi was later to add this conclusion as a third
specified in the figures. Furthermore there are virtually no real geometrical part of the proposition in his reconstruction of it (see Text B below, marg.
proofs, for most of the time the propositions are merely illustrated by the pg. no. 221). Proposition 12 (see Fig. 7.42) then asserts that if we draw al
use of numbers. (These defects led Francesco Barozzi to rework the whole perpendicular to the plane of the axial triangle and parallel to lines hm and
treatise, as we shall see later.) Provenzale’s first nine propositions are kn (drawn as in the previous proposition), and we draw a straight line from
ancillary and, for the most part, have their origin in Euclid’s E lem ents41 t, the midpoint of ta, to /, and we extend the line to /, and if we extend hm
I need not present them here, though the reader can discern their sub­ and hn to line ti, meeting it in points r and s, then al2 / at2 = hr2 / ht2 = ks2 /
stance by consulting Barozzi’s work below (see Text B, Extract VII, marg. kt2.50 This proposition is not in the De duabus lineis; it is here presented as
pg. nos. 211-18). Proposition 10 (see Fig. 7.40) first asserts that hi hr / a kind of auxiliary proposition. Proposition 13 (see Fig. 7.43) asserts that if
hd-ha = kl-ks / kd-da.4748 This is the inverse of step (6) of Proposition 4 of we assume al such that al2 / at2 = hm21 hd-ha = kn21 kd-ka, then by us­
th&De duabus lineis (see Chap. 2, Text A). Provenzale adds to this proposi­ ing Propositions 11 and 12 we arrive at the conclusion that hr2 / ht2 = ks2 /
tion the statement that the products of the segments of the transverse lines kt2 = hm2 / hd ·ha = kn21kd -ka ,51 which is similar to the equation of step
(such as ri, si, etc.) are greater the closer to the base such lines are taken.
Proposition 11 (see Fig. 7.41) has two parts: (l)h m 2/hd-ha = kn2lkd-ka\ plicati di quel traversi con li moltiplicati della commune al modo sopradetto sara concessa
(2) since the products that are the denominators in the previous proportion parimente alii quadrati di quelle collonne. Si che’l quadrato della collonna hm hara propor­
are greater as the transversals are greater and thus the products are greater cione col moltiplicato delYhd per la ha come havera il quadrato del la collonna kn col molti­
plicato del la kd per la ka. (9v) Poniamo in essempio la colonna hm di braza 2 il cui quadrato da
as the points h and k are closer to the base, hence the squares of the per­
4, facciamo poi la linea hd di braza 8 moltiplicando per la ha et sia braza: 4; dara 4 vie 8, 32,
pendiculars hm and kn equal to these products are likewise greater (or to che otiplica il quadrato delYhm-, poniamo poi la lin eal di dl (/) braza 12, la quale moltiplicata
put the matter more simply, kn2 > hm2).49 These parts are equivalent to per la ka di braza 8; dara 8 vie 12, 96, di cui 1’aottova (/) parte e 12, et tanto sara della colonna
kn il quadrato et essa sara di braza 3 · Vim circa, ch’e di 12 la radice. Et si come li sopradetti
47 They occupy folios 4v-8r of the ed. of 1550. At the end of the ninth proposition (8r) the traversi ac(10r)costandosi al diametro della basa cresceranno il moltiplicati di quelli cosi le
author writes: “ Sin qui furno dimostracioni conferenti al quisito. Da qui in[n]anzi sono dette colonne; ch’i quadrati loro sono a quelli moltiplicati eguali appressandosi alia basa ver­
proprie.” I should note further that this ninth proposition merely indicates that if a cone is ranno maggiori.”
bisected [through its axis] the section is triangular. 50 Ibid., lOr-v: Dimostratione xii. Se dal ponto a della commune db [Fig. 7.42], dove essa
48//?/</., 8r-9r: “ Dimostratione x. Se del detto triangolo un lato, come la gc [Fig. 7.40] linea entra nella piramide, uscira una collonnella propendicolare (/) come la al equidistante
uscita rettamente, poniamo al punto d et d’indi tirata una linea, come ladb entri nella piramide alle prefate colonne hm et kn prosuposta levata come quella in aere, mu tutta pero fuori
al ponto a et partendo li sopradetti traversi equidistanti nelli punti h et k postase sopra il della piramide, sopra la quale si tiri una linea retta trata dal proprio mezzo di quella parte,
diametro della basa nel punto b formerannosi per quella nel mezzo della piramide due tri- ch’avanza della commune fore della piramide, come la tl, dal ponto t imaginata, anch’essa
angoli congiunti, come lo abc et lo dbe con la linea dab commune ad amendue la onde come tirata in aere, et vadane, retamente deve si voglia, poniamo al ponto i certo e ch’ella sara
nella quarta s’e dimostrato, li moltiplicati d’una parte per l’altra di detti traversi, et de ogni sempre all’incentro propendicolarmente (/) dalla commune d b , et per lei si formera il triangolo
altro, chi’in tal guisa vi fusse saranno propertionati (/): Si che’l moltiplicato dell’/» per (8v) tbi, la onde se si faranno aggiungere le collonne hm et kn alia linea tl, come nelli punti r et j , li
Yhr havra proportione col moltiplicato delYhd per la ha come havra il moltiplicato della kl quadrati di quelle tre collonne equidistanti, et d’goni (/ d’ogni) altra, ch’in tal gui(10v)sa vi
per la ks col moltiplicato della kd per la ka; et secondo saranno propinqui li traversi al diametro fusse, saranno proportionati, come nella seconda e dimostrato. Si chel quadrato della colonna
della basa, cosi li moltiplicati di quelli una parte per l’altera verranno crescendo: perche al hara proportione col quadrato del Yat come havra il quadrato dell ’/zr col quadrato dell’/it,
ciascuna d’esse parti diverra sempre maggiore, per esser nel piu largo del suo triangolo, et anchora come havra il quaurato (/ quadrato) dell as (/ ks) col quadreto (!) della k t.”
come’chiaro per la presente si po vedere, et maggiormente discernere che procedendo la 51 Ibid., 1 0 v -llv : “ Dimostratione xiii. Se la colonella al [Fig. 7.43] sara di qualita (/
commune db con la pira(9r)mide in infinito si trovera sempre tra li lati ge et g c .” I have quantita) che suo quadrato havra proporcione col qua(llr)drato della at come hanno li
italicized letters representing points and magnitudes and eliminated the periods by which quadrati delle colonne km (! kn) et hn (/ hm) con li moltiplicati della commune al modo detto
these are set out in the text. I have, as usual, changed consonantal “ u” to “ v ” and vocal “ v” nella xi. Eglie noto per la procedente (/), che tal proporcione haranno anchora li quadrati delle
to “ u” . I have also made a few changes in punctuation. All of these kinds of changes are also colonne hr et hs (! ks) con li quadrati della commune pur’anche al modo predetto. Si chel
found in the texts of the succeeding footnotes. The printing of this text is miserable, as quadrato di tutta la hr hara proporcione col quadrato dell’/u, et cosi il quadrato di tutta la ks
will be evident to the reader of these notes, which follow the text in all of its imperfections. hara proporcione col quadrato della k t, come ha il quadrato della parte hm col moltiplicato dell
w Ibid., 9r-10r: “ Dimostratione xi. Se della commune db [Fig. 7.41] nella parte di quella hd per la ha et come ha il quadrato della parte kn col moltiplicato della kd per la ka. Seguitando
dentro la piramide, come dalli punti h et k, usciranno collone, et propendicolari assenderanno adunque l’essempio nostro nella undecima esendo la at di braza 2, il cui quadrato da 4, sara il
alia superficie della piramide ciascuna al circolo del suo diametro, come la hm et la kn quadrato della colonnellaa/ braza meza, ch’e di quatro l’ottava parte, et essa sara da tre quarti
prosumendo quelle colonne, et circoli levat’ in aere per quello si e proposto nella quinta; dara di braza in circa, ch’e di braza una meza la radice, et essendo poi la/if di braza 6, ilcui quadrato
tanto il quadrato di ciascuna colonna, quanto da il moltiplicato d’una parte, per l’altra del suo da 36; sara il quadrato della colonna hr braza 4 et meza ch’e di 36 l’ottava parte, et essa sara di
traverso; percio ch’egli e diametro del suo circolo; per tanto la proporcione, ceduta alii molti- braza: 2 et una ottava in circa, ch’e di 4 e una meza la radice. Essendo anchora la kt di braza
342 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 343

(5) in Proposition 5 of the De duabus lineis, except that it includes mem­ general enunciation whose two parts may be specified as follows: (\) hr2
bers for the two perpendiculars hmr and kns, while step (5) includes mem­ —hm2 — 2 hm-mr + mr2 and ks2 - kn2 — 2 kn-ns + ns2, which is
bers for only one perpendicular (with the later assumption that it is valid for equivalent to the last equation of step (6) of Proposition 5 of the De duabus
any other perpendicular). Proposition 14 then shows by the computations lineis·, (2) hence ns < mr, because (from Proposition 16) the differences of
made in the preceding proposition (again see Fig. 7.43) that the curve amnv the squares are equal and hence 2 hm -mr mr2 = 2 kn-ns + ns2.55 This
-I-

and the straight line tlrsi always come closer as the two lines are extended.52 second part is equivalent to steps (7)-(9) in Proposition 5 of the De duabus
But he adds that he will proceed to other demonstrations without making lineis. Finally, in Proposition 18 (see Fig. 7.44) the author demonstrates
use of computation. Proposition 15 (once more see Fig. 7.43), assuming the that the curve and the straight line do not meet.56 The demonstration is like
results of Proposition 13, asserts that (hr2 - hm2) / (ht2 - hd-da) = (ks2 li moltiplicati di quella sera quanto importa il quadrato d’essa at \ cosi lo avanzo del quadrato
—kn2) / (kt2 —kd-ka) = hr2 / ht2 = ks2 / kt2 = hm2 / hd-da —kn2 / kd-ka.53 dell ht sopra’l moltiplicato dtWhd per Yha come l’avanzo del quadrato della kt sopra 1’molti­
This series of equal ratios is similar to the proportions established in the plicato della kd per la ka e qual’altro avanzo simile per sortirne altre colonne vi concoresse
first two equations of step (6) of Proposition 5 of the De duabus lineis. [Fig. 7.43]” .
Proposition 16 (still consult Fig. 7.43) has three parts: (1) ht2 - dh-da = at2 “ Havendo adunque li sopradetti avanzi (14r) delli quadrati delle colonne magiori sopra li
qnadrati (/) delle minori una medesima proportione ciascun di loro col detto avanzo, che’la
and kt2 - kd-ka = at2, if we reduce the partially verbal statement to the quantita del quadrato della at et quella proportione istessa, che nella XIII s’e prosupposto
specific magnitudes, this is equivalent to the second equation of step (6) of haver secho anchora il quadrato della colonnella al, egli e necessario (14v) concedere
Proposition 5 of the De duabus lineis', (2) hr2 - hm2 = al2 and ks2 - kn2 quegli avanzi et quel quadrato contenere in se ciascun di loro una medesima quantita. Si che
= al2 (and similarly for the squares of any other perpendiculars erected l’avanzo del quadrato della colonna hr sopra l’quadrato della colonna hm sara in quantita
on dt); therefore, (3) hr2 - hm2 = ks2 - kn2 (and similarly for any other quanto e l’avanzo del quadrato della colonna ks sopra’l quadrato della colonna kn et l’uno
avanzo come l’altro sara quanto e ’lquadrato della colonnella al et tanto sarebbe ogni altro
perpendicular).54 Proposition 17 (see Fig. 7.43) is given a completely avanzo simile, che per altre colonne uscirne a quel modo ve accadesse, percio che tutti
sarebbono come essi proportionati.”
10, il cui quadrato da 100; sara il quadrato del( 11 v)la collonna ks braza 12 e meza ch’e l’ot- 55Ibid., 14v-15v: “ Dimostratione xvii. Attento che nelle sopradette colone per la dimo­
tava di 100, et essa sara di braze 3. et 9 sedesene in circa, ch’e di 12 e meza la radice.” stratione fata nell’ottava sara lavanzo (/ l’avanzo) del maggior quadrato sopra la minore
52Ibid., llv -1 2 v : “ Dimostratione xiiii. Se dal ponto a dove entra la commune nella quanto dal moltiplicato della colonna minore per l’avanzo della maggiore due volte, et il
piramide sara tirata una linea nella superficie d’essa piramide all’incontro propendicolarmente quadrato d’esso avanzo, egli e forza cedere quell’avanzo nelle colonne venir calando in ver
della commune ab so(12r)pra le colonne minori hm et hn (/ kn) et scorri per sino al circolo la bassa. Dove (come e dimostrato per ΓΧΙ.) vengono le colonne piu grandi; poscia che tutti
della bassa come la av [Fig. 7.43], imaginata anch’essa tirata in aere, egli e noto per il con- gli avanzi di (15r) quei quadrati cosi delle piu propinque alia basa’ come delle piu distanti da
puto fatto delle colonne minori nella undecima, et nell’antecedente delle maggiori: che le due quella’ contengono in se (come dianzi fu detto) una medesima quantita. Si che essendo
linee, cio e la curva et superficiale av et la retta li, si vanno sempre accostando; percio che l’avanzo del quadrato dell’/ir [Fig. 7.43] sopra’l quadrato della hm quanto da il moltiplicato
la prima distanza tra loro su la al di tre quarti di braza in circa; la seconda computata fu la delYhm per la mr due volte, et il quadrato di essa mr, et similmente l’avanzo del quadrato
mr (12v) d’uno ottavo in circa, ch’e l’avanzo dalla maggiore hr sopra la minore, hm, la terza della ks sopra’l quadrato della kn quanto da il moltiplicato della hn (! kn) (15v) per la ns due
computata similmente fu la ns, d’una sedesena in circa, ch’e l’avanzo della maggiore ks sopra volte, et il quadrato d’essa ns, forza e cedere, che Ians sia minore della mr, poi che ciascuno
la minore kn, et in tal modo procedendo si po far la descretione di qual’altra distanza si delli moltiplicati col quadrato a quel modo asscendono ad una somma, et e pur la kn maggior
ricerchi, et troverassi, che le distanze piu propinque alia basa veranno sempre calando, se che la hm il medesimo si po comprendere in cadun’ altro avanzo simile d’ogni altra colonna,
procedesse anchora tal figura in infinito, come dicevamo. che in tal guisa vi fusse, percio che come cresceranno le colonne in altezza, cosi l’avanzo di
“ Ma perche intendiamo provar tal effetto, in modo che senza far computo si discerna, ne quelle, ch’e la distanza tra la superficiale av et la retta li verra calando. Dunque esse linee si
verremo ad alcune altre dimostratione.” accosteranno sempre, anchor procedessero a quel modo in infinito.”
53Ibid., 12v-13v: “ Dimostratione. xv. Attento, che li quadrati delle colonne maggiori 56 Ibid., 15v-17v: “ Dimostratione. xviii. Che le sopra dette due linee non si congiungano
hanno proporcione con li quadrati della commune com’hanno li quadrati delle colone (/) mai, agevolmente per la presente si potra vedere. Concio sia, che se quelle si congiugnessero
minori con li moltiplicati di quella al modo detto nella XIII, l’avanzo anchora delli quadrati in qual si voglia punto della linea retta, prosupponiamo nel ponto o, et d’indi si gettasse sopra
delle colonne maggiori sopra li quadrati delle minori havra l’stessa proportione con l’avanzo la comune una propendicolare equidistante all’altre colonne maggiori et minori, come la
delli (13r) quadrati della commune sopra li moltiplicati di quella, come per la settima s’e op (16r) ellapur sarebbe maggiore et minore, una divenuta; poi che la superficiale au et la retta
demostrato. Si che l’avanzo del quadratio della colonna hr [Fig. 7.43] sopra’l quadrato della li sono in quel punto congiunte insieme la onde converrebbe chel quadrato di quellapo per la
colonna hm hara proportione con l’avanzo del quadrato della ht sopra il moltiplicato (13v) equidistanza con le colonne maggiori havesse proportione col quadrato della pt della com­
della hd per la ha, et Cossi l’avanzo del quadrato della colonna ks sopra l’quadrato della mune, come hanno li quadrati dell’altre (16v) collonne maggiori con li quadrati di quella al
colonna kn havra proportione con l’vanzo (/ l’avanzo) del quadrato della kt sopra l’molti- modo detto nella XII. Dall’altra parte per la equidistanza con le colonne minori gli converebbe
plicato della kd per la ka, come hanno li quadrati della hr et dela ks con li quadrati detti della a quel quadrato della po haver proportione col moltiplicato dellapd per la pa della commune,
commune, et come hanno anchora li quadrati dell’Am et dell’&n con li moltiplicati detti di come hanno li quadrati delle colonne minori con li moltiplieati (/) di quella, al modo detto
quella.” nella XI; et perche tal’e la proportione che hanno li quadrati delle colonne maggiori con li
54 Ibid., 13v-14v: “ Dimostratione xvi. Per esser la at come e prosuposta nella XII la quadrati della commune, qual’e la proportione che hanno li quadrati delle colonne mag­
meta della ad. Egli e noto per la sesta, chel’avanzo di detti quadrati della commune sopra giori, con li quadrati delle commune; qual’e la proportione che hanno li quadrati delle
344 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 345

that of steps (2)-(4) in Proposition 5 of the De duabus lineis, since both of its base (see Fig. 7.45), thus producing concentric circles. In them he
authors assume that the lines do meet in some point o and then show that draws two straight lines, one (da) from the center of the circle (represent­
this assumption is impossible. So ends Provenzale’s principal demonstra­ ing the apex of the cone before its projection) to the circumference of the
tion of the existence of lines that approach each other but never meet. I outer circle and the other (fr) tangent to the interior circle and parallel to
have stressed in my brief account that the propositions show similarity with the first line. Then two sets of parallel lines are drawn, the first one being
the various steps of Propositions 4 and 5 of the De duabus lineis. It is the perpendiculars ag,bh , and ci, and the second one being the longer non­
equally obvious that they are also similar to the equivalent steps in perpendicular lines a r,b s, and ct. Then it is clear that in the case of either
Elements XVII and XX in Werner’s Libellus (see Chap. 6, Text). And set, the distance between a circumference and its parallel grows less as we
while I doubt that Werner’s text was Provenzale’s source (since take larger circles. Hence gk < hi < if and rk < si < tf. Now if we con­
Provenzale would have probably designated the curved line as a hyperbola sider the cone as reconstructed and draw a line through rf parallel to an
had he read Werner’s work), it is of passing interest that Provenzale, like axial plane through ad, then we would have two lines formed in that plane,
Werner but unlike Simeon Motot and the author of the De duabus lineis, one a straight line in which r, s and t would lie and the other a curved line (in
omits (where the medieval authors include) the final step of declaring that fact a hyperbola) on the surface of the cone, a line to which rst is an asymp­
the perpendicular distance from a point on the curve to the asymptote tote. While the proof is very truncated, obscure, and open to the serious ob­
decreases the more the lines are extended. That is, Provenzale should have jections proposed by Barozzi (see Text B, marg. pg. nos. 242-48), who
added this step after showing that ns < mr (see Fig. 7.43). notes that the various minimal distances that are being sought cannnot be
As I noted above, Provenzale has added a brief discor so at the end of the aptly shown in a plane, it should be evident to the reader that Provenzale’s
tract.57 In it he projects all the circular elements of a right cone on the plane projection technique reflects what is essentially the same procedure as
that presented in Conclusion 11 of the Speculi almukefi compositio (see
colonne minori con li moltiplicati di quella, com’e detto nella XIII; conseguentemente con-
verebbe al quadrato della po che havesse quella medesima proportione col quadrato dellapt
Chap. 4 above, Text, and particularly the drawings added by the editor to
qual’egli havesse col’moltiplicato della pd per la pa, et questo non e possibile; percio chel Fig. 4.13). Hence it seems probable that he had seen the earlier work.
quadrato della pt e ’ pur’in quantita piu del moltiplicato della pd per la pa di tanto, quanto Provenzale’s tract was quickly read and mentioned by Girolamo Car­
importa il quadrato della at, come nella XVI s’e dimostrato: et come potra una sola quantita, dano, not in the first edition of his De subtilitate (Nuremberg, 1550) but in
ch’e il quadrato della po (17r) haver’una medesima proportione con due quantita l’una mag-
the edition of Basel, 1554, and those editions that succeeded it. In his first
gior dell’altra; forza e ( 17v) adunque, che tutta la po sia quella che’l suo quadrato havra la detta
proportione col quadrato della pt, ch’e la quantita magiore (/), et una parte di lei sia quella edition of 1550 Cardano’s treatment of conic sections is quite brief.58 While
chel suo quadrato havra le proportione detta col moltiplicato della pd per la pa ch’e la
quantita minore. Hor lo avanzo di lei sara la distanza tra le due linee, cio e la superficiale et come le sopradette, che sempre s’accostano, et non s’incon[t]rano mai benche di sopra forse
curva αν et la retta li. per notitificar la ptoportione (/), et quantita delli sopradetti avanzi si siano per altra via
“Tant’e che necessariamente sara sempre distanza tra loro, anchor procedessero in dimostrate.
infinito. FINIS.” “ Questo ne par prova a sofficienza in dimostrar, che la montuosa rotondita sia quella,
57 Ibid. , 18r-20r: “ Discorso. La causa principiale, onde procede tal’ effetto dalle predette che tal’effetto nella piramide parturisce.
linee sopra la piramide rotonda parmi veder che stia nella montuosa rottondita di quella; im­ “ E se in questo secondo modo voleste por quelle due linee in figura, poreste piglia(19v)re
pero che quando fusse tal piramide spianata, et tutta posta in sn (/ su) la basa, ella pur una rapetta, o altra materia, et da tale forma di piramide, tirale sopra una linea retta dalla
diverrebbe in tanti circoli l’uno dentro all’altro, e initi d’intorno ad un punto fatto centro, sommita alia basa, come la da et discosto in guisa di paralello allei dar’un taglio, et porvi
ch’era gia la sommita d’essa piramide, come nella presente figura [Fig. 7.45] per tal dimostra- dentro una lista di carta, che fuori n’esca per superficie, et tirarle sopra una linea proprio ac­
tione l’habbiamo posta in piano, et discrittovi di lei tre circoli, ak, bl, cf, d’intorno al centro canto il taglio, come la kf per la curva signar per giusta et ugual misura dalla prima da all’uno
d dal qual tirata la da, et all’incontrola rf paralelle, tirarem tra quelle dalli punti a, b, c tre et l’altro capo della carta sia la misura propendicolare o piu larga dalla piramide che la pro-
traversi equidistanti; siano quelli propendicolari como lo ag, lo bh, et lo ci o siano piu larghi pendicolare non sarebbe, et tiratavi una linea retta come la ri ritroveransi quelle due linee
dalli circoli, che li propendicolari non sarebbono, come lo ar, lo b s, et lo c t . Hor si come nella sopra la carta giustamente riportate.
presente figura piana all’uno et l’altro modo di troverai evidentemente si discerne (18v) la “ Hor siamo venuti nella dimostratione di tale tratato a compimento gratia a colui, a cui per
distanza tra quelli et li circoli per rispetto della volta circolare venir calando nelli circoli mag- essaltarlo tutte le lingue s ’appressano, ne intoppano pero gia mai in una (20r) minima parte di
giori che meno voltano; percio che la gk e pur minore dellTi/, et la hi e minore della if, et sua eccellenza; ma per quello che sappiamo, et possiamo dire, sia egli lo dato et essaltato
similmente e la rk minore della si et las/ minore della//, ne sara pero maiche non vi sia per lo sempre. Amen. FINIS. Stampato in Mantova ad instantia de Iacomo di Ruffinelli M .D.L.”
voltar del circolo distanza alquanto, cosi prosumendo quella figura montuosa ritornata in for- 58 G. Cardano, De subtilitate libri XXI (Nuremberg, 1550), Bk. XVI, pp. 303-05: “ [1]
(19r)ma piramidale, et la ad ascendente sopra qaella (/) dalla basa alia sommita, ritroverassi Quae vero circulo, hyperboli, et defectioni communes sunt, hae sunt. Ducta ex contingente
la paralellar/ con li detti travesi levata in aere discosta da quelli tre circoli, la istessa distanza perpendicularis super diametrum in directo puncti ex quo contingens ducta iacentem partes
calante in ver la basa al modo detto, restandovi sempre per la detta cagione qualche distanza. diametri sub eadem proportione dividit, sub qua tota linea ex puncto e quo contingent (/;
” Hor se sotto quella nella superficie della piramide tireremo una linea, come la kf et sara contingens, ed 1560) producta est ad centrum circuli veniens usque ad alteram circum­
curva, per che fala sommita, non havrem noi formate su la piramide in tal modo due linee, ferentiae partem, ad par(304)tem exteriorem se habet. Semidimetiens quoque proportionalis
346 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 347

not mentioning either Maimonides or Provenzale, it shows the influence of Libellus ,60 and Archimedes’ On the Quadrature o f the Parabola.61 In the
Apollonius’ Conics (in the form of Memmo’s translation),59 of Werner’s edition of 1554 Cardano expands his treatment of conic sections, giving the
generation of the conic sections from the cone, and adding a proof of the
(/; proportione media, ed. 1560) est inter eam quae a centro ad punctum exterius et eam finding of a curve (hyperbola) and straight line (asymptote) that have an
quae a centro ad locum ubi cadit perpendicularis ex loco contingentis super eandem dimetien­
asymptotic relationship.62 He remarks that although Apollonius has given
tem. Cum vero a terminis dimetientis duae contingentes ducuntur, ab eisdem vero punctis
per idem punctum circumferentiae mutuo ad alteram contingentem, erit quod sub partibus
contingentium his postremis lineis terminatarum rectangulum continetur aequale quadrato 80 The most solid piece of evidence indicating Werner’s influence is in paragraph [4] of note
dimetientis. [2] Defectionis duo sunt privilegia: primum quod proportio eius ad circuli super­ 58, where Cardano notes that when the ordinate of a parabola is equal to its axial intercept the
ficiem est velut rectanguli dimetientium defectionis ad rectangulum dimetientium circuli, ordinate is called the latus rectum of the parabola. This is straight from the sixth definition of
quod est quadratum. Secundum ex hoc ducit originem, quod proportio defectionis ad Werner’s Libellus (see Chap. 6 above, Text). Notice further that Cardano calls the perpen­
defectionem est velut rectangulorum sub dimetientibus earum propriis contentorum. [3] diculars to the axis of the parabola ordinatae and thus is one of the first to designate the "lines
Hyperbolis rursus duo propria privilegia: primum duas lineas est invenire, quae ex centro drawn ordinate-wise” (the usual expression for them in the early sixteenth century) simply
extra hyperbolen posito semper evadent lateribus hyperbolis propinquiores et nunquam as "ordinates".
tamen ea contingent. Manifestum est igitur et ex his quod infinitas erit alias lineas rectas 61 The fifth privilegium of the parabola (see paragraph [4] in note 58 above) is essentially the
prioribus quidem aequidistantes et in uno coniunctas puncto licitum accipere, quibus latera final conclusion of Archimedes’ On the Quadrature o f the parabola.
hyperbolis perpetuo usque in infinitum magis ac magis appropinquabunt, certo tamen spacio 82 See G. Cardano, De subtilitate libriXXI (Basel, 1554). Seepp. 418-21 for a discussion of
quodcunque voluero, semper aliquanto plus distabunt. Tertium erit quod in quacunque parte the cone and the generation of its sections, most of which is missing in the earlier edition.
circumferentiae hyperbolis duo puncta signentur, producanturque ad non tangentes binae ac The passage on the asymptotic problem occupies pp. 421-23: “ Sunt autem hyperboli tria
binae lineae mutuo aequidistantes, rectangula sub illis contenta quae ab imo veniunt atque maxime praecipua, quorum primum est quod in quaque illius parte circumferentiae duo
alio puncto invicem erunt aequalia. [4] Paraboles autem privilegia sex propria sunt. Primum, puncta sumantur, a quibus binae et binae ad non tangentes rectae lineae deducantur mutuo
ratio axis partium in ea est ut deductarum ex punctis perpendicularium ad paraboles cir­ inter se aequidistantes, rectangula contecta (/ contenta?) ab his lineis quae ab imo eveniunt,
cumferentiam duplicata. Secundum, cum fuerit ipsa perpendicularis aequalis axis parti quae atque ab his quae ab aliis punctis, invicem aequalia erunt. Secundum est quod invenire
ad verticem ab extremo eiusdem perpendicularis terminabitur, vocabitur ipsa perpendicu­ contingit duas lineas in eodem plano, quarum altera erit recta, reliqua latus hyperboles, quae
laris latus rectum paraboles, eritque haec semper talem habens proportionem ad quamcunque semper sibi invicem magis approximabuntur, et nunquam se tangent. Tertium ex secundo
perpendicularem ex axe ad circumferentiam qualis est perpendicularis ipsius ad partem axis pendet, quod erit iunctu (later eds. inventu) facile, duas lineas quae semper magis in eodem
quae ipsam perpendicularem et verticem sectionis interiacet, vocantur vero hae lineae per­ plano approximabuntur, et quanquam etiam in infinitum protraherentur, nunquam erunt
pendiculares ordinatae. Manifestum est igitur quod cuilibet parti axis paraboles ac suae proximiores mille stadiis, gratia exempli. Demonstrato enim secundo, si sumatur linea aequi-
perpendiculari semper eadem linea in continua proportione subtenditur. Tertium, quod si distans rectae ex adversa parte mille stadiis, patebit quod dictum est. Igitur demonstremus
in ea punctus praeter axem signetur, ab hoc contingens ducatur, huic vero aequidistantes secundum, quod, licet ab Apollonio demonstretur, volo tamen uti demonstratione Rabii
plurimae a circumferentia ad circumferentiam ducta ex eodem puncto contingentiae aequi- Moysis Narbonensis, exponentis dictum Rabii Moysis Aegyptii in libro cui titulus est
distans axi, omnes lineas aequidistantes contingenti ductas per aequalia secabit. Portiones Directio dubitantium, quod erat: Quaedam intelligi posse quae imaginari nequeunt: unde
quoque quomodolibet sumptae aequales habentes dimetientes etiam aequales sunt. Ipsa concludit quod intellectus ab imaginatione differat, non solum ob novitatem, sed ob facilita­
vero superficies aequalis est rectangulo ex tota basi in duas e tribus axis partes. (205) Sex­ tem et pulchritudinem. Sit igitur conus ABCD [Fig. 7.46]: nunc triangulum nullo (/ nullum)
tum, cum tres contingentes periferiam paraboles coincidunt, duae quidem extremae, media secantem intelligo. Sed per ABD intelligo connexam (/ convexam) coni superficiem, in qua
eas secat, erit proportio partium trium linearum una; scilicet partis inferioris ad superiorem, protraho AC a vertice usque ad basim. Et sit K plana superficies contangens conum in recta
et superioris alterius ad inferiorem, et mediae illarum quae ad periferiam paraboles terminan­ linea AC: quae superficies intelligatur in infinitum cum coni superficie extendi. Dico primo,
tur. [5] Omnibus autem coni sectionibus atque circulo commune est, ut etiam contrapositis, hanc superficiem planam non posse tangere coni superficiem alibi quam in linea AC: quod si
ut cum duae contingentes in unum coierint, ducta ex loco concursus recta usque ad adversam potest, tangat in G, et duo (/ duco) circulum aequidistantem per G basi BCD: cum igitur
circumferentiae partem, vel in contrapositis ad lineam rectam quae per puncta contactus (422) circulus sit in una superficie, erunt puncta contactus plani K et periferiae circuli il­
ducitur, erit totius lineae ad partem quae est extra obliquas, velut partium intra obliquas con­ lius in una recta linea, ex demonstratis in 11. Elementorum Euclidis. Quamobrem cum illa
tentarum eiusdem lineae. Cum vero trigoni cuius basis ad rectos basim hyperbolis secat latus linea iam tangat circuli periferiam in linea AC, cadet ex demonstratis ab Euclide in tertio
reliquum ducitur, medium partis exterioris axis hyperbolis centrum vocatur, exteriorque Elementorum extra circunferentiam circuli VXG, igitur non tanget illum in puncto G. Assumo
pars intra contactum versa ad quam vero versa potest, ut quadratum perpendicularis e igitur EF rectam aequidistantem AC in superficie K, et adeo propinquam rectae AC, ut
vertice coni aequidistantis axi, ad basim trianguli, ad rectangulum partium basis ipsius, superficies H, ducta ad perpendiculum super superficiem K secet conum et illius super­
vocatur recta. Igitur cum duos conos angulorum aequalium e directo verticum sibi op­ ficiem in punctis, puta 5 et G, et palam est ex dictis, partem superficiei ex H cono inclusam
positos, ut latera sint in eadem linea recta, superficies plana simul secat, conorum ambae esse hyperbolem, et lineam GS, quae est in coni superficie, esse latus hyperbolis. Constat
sectiones necessario erunt hyperboles. Has ambas Apollonius solet contrapositas vocare." I igitur iam latus hyperbolis GS esse in superficie eadem cum linea EF, scilicet in superficie H:
have made the usual changes of punctuation and consonantal "u" and vocal "v” . I have et quod istae duae lineae, cum sint in eodem plano//, nunquam se tangent: si enim se tangent,
given in parentheses two corrections from the edition of 1560 that help in furnishing better vel in linea AC, et ita AC et EF aequidistantes concurrent, quod includit contradictionem:
sense to the passage. vel extra lineam AC, et ita, cum GS semper sit in superficie coni et EF semper in superficie K,
59 This is evident from Cardano’s use of Memmo’s term defectio for an ellipse in para­ igitur ATtangeret conum extra lineam AC, cuius iam oppositum demonstravimus. Dico modo
graphs [1] and [2] in the preceding note and from his use of contrapositae for άντΜείμεναι (see quod, cum EF recta et GS latus hyperbolis sint in eadem superficie H, et protractae in in­
paragraph 15]). For these terms in Memmo’s translation, see the ed. cit., 8v, 9r (corr. ex lOr). finitum nunquam conveniunt, quod semper ut magis a vertice coni elongantur quod eo magis
348 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 349

a solution to the problem (by which he no doubt refers to Proposition II. 14 evident. Therefore, let us demonstrate the second principle. Although it is
of the Conics), he will present rather the demonstration of Rabbi Moses of demonstrated by Apollonius, still I wish to use the demonstration of Rabbi
Narbonne (/, i.e. Moise Provenzale): Moses of Narbonne, who expounds the statement of Rabbi Moses of Egypt
given in a book whose title is Guide of the Perplexed, which statement was:
There are three principles of special importance for the hyperbola, of which there are certain things which can be understood but which cannot be imag­
the first is that if two points are taken in any part of its periphery, and from each ined. Whence he concludes that the intellect differs from the imagination, not
of them a pair of lines is drawn to the asymptotes, the lines from one point being only by means of novelty but also by means of facility and beauty.
parallel to the lines from the other, the rectangles contained by these lines, Therefore, let there be a cone ABCD [see Fig. 7.46]. Now I understand no
one pair at a time, and [by similar pairs drawn] from other points, will be equal cutting triangle. But by ABD I understand the convex surface of the cone, in
to each other. The second principle is that it is possible to find two lines in the which I draw AC from the apex to the base. And let K be a plane surface that
same plane, of which lines one will be a straight line and the other a side (/) touches the cone in straight line AC, which plane I understand to be extended
[i.e., one half of the curve] of a hyperbola, which will always approach each indefinitely along with the surface of the cone. First I say that this plane sur­
other more closely but never touch each other. The third principle depends on face cannot touch the surface of the cone anywhere else but in line AC. For if it
the second, and it will be easily related to it: [it is possible to find] two lines in can [touch in another place], let it touch in G. And I draw a circle through G
the same plane which will always approach each other more closely and which, parallel to base BCD. Hence since the circle is in one surface, the points of
though extended indefinitely, would never come closer than [a set distance], contact of plane K and the circumference of the circle will be in a straight line,
for example, 1000 stades. For, with the second principle demonstrated, if a line from the things demonstrated in the eleventh [book] of the Elements of Euclid.
is taken parallel to the first line and 1000 stades beyond it, the assertion will be Accordingly, since that line already touches the circumference of the circle in
line AC, it will fall (from those things demonstrated by Euclid in the third
fiunt proximae. Et sufficiat demonstrasse de uno, utpote quod G et M sint propinquiores [book] of the Elements) outside of the circumference of circle VXG \ therefore
quam S et T: nam tunc patebit quod ubi magis procedent illae duae lineae, erunt eo proxi­ it will not touch the cone in point G. Then I assume a straight line EF parallel
miores . Capiatur igitur gratia exempli circulus P S Q , et ducatur TSR, ita quod perveniat ad op­ to AC and lying in [plane] surface K and just close enough to straight line AC
positam circunferentiae partem: et similiter ducatur MGN in superficie H, ita quod GN that the [plane] surface H drawn perpendicular to [plane] surface K cuts the
perveniat ad circunferentiam circuli VGX: et ducatur recte LT et OM in superficie H (/ K),
cone and its surface in points like 5 and G. And it is clear from what has been
quae contangent circulos QLP et XOV, quia ducuntur ex loco contactus: et quia O et M
sunt in superficie circuli OXV, nam M est terminus lineae MN, quae est in superficie circuli said that the part of the surface of the cone included in [plane] H is a hyper­
OXV, erit linea OM in superficie eiusdem circuli, et ita LT in superficie circuli PLQ. Sed bola, and that line GS, which is in the surface of the cone, is the side of the
tales superficies aequidistant, quia ambae a basi circuli: et sunt lineae OM et LT in superficie hyperbola. Hence it is already evident that the side of the hyperbola GS is in
K ambae: igitur aequidistantes. Et iam LO et TM aequidistant, sunt enim partes aequi- the same surface as line EF, evidently in [plane] surface H. And that these two
distantium, igitur LT et OM sunt aequales. Et cum contangant circulos PLQ et VOX, igitur ex lines, since they are in the same plane H, never touch each other [is clear]. For
demonstratis ab Euclide in tertio Elementorum quadratum TL est aequale ei quod fit ex TR in if they do touch each other, it will be either in line AC (and so parallels AC and
TS, et quadratum OM est aequale ei quod fit ex M N in MG, et quadratum TL est aequale EF will run together, which includes a contradiction) or it will be outside of line
quadrato OM, igitur quod fit ex TR in TS est aequale ei quod fit ex MN in MG. Igitur ex demon­ AC (and so, since line GS is always in the surface of the cone and EF is always
stratis sexto elementorum ab Euclide proportio ST ad GM est ut MN ad TR. Sed MN maior
in the [plane] surface AT, therefore K would touch the cone outside of line AC,
est TR, quia si duceretur per N , superficies aequidistans ipsum (/) caderet infra (/ extra) i?,
the opposite of which we have already demonstrated). Now I say that since
aliter occurreret K, quia diameter QP est maior XV, et superficies circulorum sunt aequi­
distantes, igitur ST maior est GM. Ducantur igitur SY et GZ ad perpendiculum super EF, straight lineFF and the side of the hyperbola GS are in the same [plane] surface
et erunt anguli S YT et GZM aequales, quia recti. Similiter anguli STY et GMZ aequales sunt, H, and [when] they are extended indefinitely, they never meet, but that al­
quia (423) ST et GM sunt aequidistantes, sunt enim ambae in superficie eadem quae est H, et ways the more they are extended from the apex of the cone the closer together
in duabus superficiebus aequidistantibus circulorum: igitur ex xxxii primi elementorum tri­ they become. And let it suffice to have demonstrated this for one case, namely
goni STY et GMZ sunt aequalium angulorum, quare per quartam sexti eiusdem proportio ST that G and M are closer together than 5 and T. For then it will be obvious that
ad GM ut 5 E ad GZ. Sed ST, ut probatum est, maior est GM, igitur SE maior GZ. Sed SE est wherever these two lines [EF and GS] proceed farther they will be closer.
minima quae possit duci ex puncto S ad lineam EF, quia ad perpendiculum eo quod omnis Hence, for example, let there be taken circle PSQ, and let TSR be drawn so
alia docta (/ ducta) ab eodem puncto ad lineam EF ex quavis parte opponitur maiori angulo that it may arrive at the opposite part of the circumference. And similarly let
quam S E, quia opponeretur recto, igitur punctus G est proximior lineae EF quam punctus S,
MGN be drawn in [plane] surface H so that GN may arrive at the circum­
quod erat demonstradum. Plerique deficiunt in hac ultima parte, admittentes paralogismum.
ference of circle VGX. And let LT and OM be drawn directly in [plane] surface
Feci igitur conum ex rapa, ut consulit Rabbi Moyses, et feci superficies K et H ex papyro, et
insciptis lineis AC, £7% SG, visae sunt non concurrentes, ut a latere vides. Sed eas nisi ea arte
K, which lines will be tangent to circles QLP and XOV [respectively], for they
inventas difficile est describere.” The author then goes on with the privilegia of the ellipse are drawn from the placets] of contact. And because O and M are in the surface
and parabola, as in the 1550 edition. I have made the usual minor changes of punctuation and of circle OXV (for M is the terminus of line MN, which is in the surface of circle
orthography described in note 58 and those that precede it. I have italicized the letters mark­ OXV), line OM will be in the surface of the same circle, and so [similarly] LT
ing points and magnitudes. will be in the surface of circle PLQ. But such surfaces are parallel because
350 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 351

each is parallel to the base circle. Further, lines OM and LT are both in [plane] construction the side of the cone and the asymptote of the hyperbola are
surface AT. Therefore they are parallel. And LO and TM are already parallel, for parallel lines in the latter plane and the hyperbola and its asymptote lie in
they are parts of parallels. Therefore LT and OM are equal. And since they the former plane. Both procedures differ from that found in Proposition 5 of
are tangent to circles PLQ and VOX, therefore (from the things demonstrated the De duabus lineis.
by Euclid in the third [book] of the Elements)TL2 —TR TS and OM2 = MN-MG, Cardano’s reconstruction of Provenzale’s proof influenced the French
and TL2 = OM2. Therefore TR-TS = MN-MG. Therefore (from the things mathematician Jacques Peletier, who gave a rather poor version of it in a
demonstrated by Euclid in the sixth [book] of the Elements) ST / GM = MN / work published in 1563.63 Peletier’s efforts were severely criticized by
TR. But MN > TR, because, if there were drawn through N a surface parallel
[to plane K], it would fall outside of R \ otherwise it would meet K because the 63 J. Peletier, Commentarii tres: I. De dimensione circuli. II. De contactu linearum; et de
diameter QP > XV and the surfaces of the circles are parallel. Therefore duabus lineis in eodem plano neque parallelis, neque concurrentibus. III. De constitutione
ST > GM. Therefore let 5 Y and GZ be drawn perpendicular to EF, and angles horoscopi (Basel, 1563), pp. 43-46:
SYT and GZM are equal because they are right angles. Similarly angles STY “ Placet huic loco subiicere quod proponit ibidem [i.e., Lib. XVI] Cardanus ex Apollonio
and GMZ are equal because ST and GM are parallel since they are both in the et Rabi Mose, de duabus Lineis in eodem Plano existentibus, quae protractae ad angulum
same [plane] surface, which is H, and in the two parallel surfaces of the circles. tendere videntur: altera ad alteram semper accedente. Et tamen magno (ut ipse Moses putat)
Therefore (from 1.32 of the Elements) triangles STY and GMZ are equiangular. miraculo, nunquam concurrunt, etiam si in infinitum protrahantur. Quod etiam obiter an­
notaverat Georgius Valla ex Ge(44)mino, libro primo suae Geometriae, cap. LIX. Et post
Wherefore, by VI.4 of the Elements, ST I GM = S Y I GZ. But, as was proved,
hunc Caelius Calcagninus ad Iacobum Zieglerum scribens, ex cuiusdam observatione,
ST > GM. Therefore S Y > GZ. But SY is the least line which can be drawn cuius nomen reticet. Veteres autem eas duas lineas primum ex Conica figuratione depre­
from point S to line EF, for any line other than the perpendicular that is drawn henderunt. Quapropter, ut earum doctrinam clariorem faciam, a Coni et Hyperboles descrip­
from S to line EF in any direction is opposite a greater angle than is 5 Y , because tione exordiar.
it [the line from S] would be opposite a right angle. Therefore point G is closer “ Conus est figura Solida descripta a Triangulo Rectangulo circa unum eorum quae rec­
to line EF than is point 5, which was to be demonstrated. Several things are tum angulum comprehendunt laterum circunducto. Ex quo apparet, in Cono esse infinita
deficient in this last part, [thereby] introducing a paralogism. Therefore, I have Trigona, illius Trianguli circunducti dupla: quorum latera in superficie Coni existunt: bases
made a cone of a turnip (rapa), as Rabbi Moses advised, and I have made sur­ sunt Diametri ipsius basis Conicae, quae ipsa Circulus est: vertex vero omnium idem cum
faces K and H of paper, and, with lines AC, EF and SG drawn, they [i.e. EF Coni vertice.
and 5G] are seen to be lines that do not meet, as you will see [in the figure] be­ “ Suscipiatur itaque unum ex iis Trigonis, quod designetur tribus notis, A , B , C [Fig. 7.46
cum legendis]: scilicet ut A sit vertex, B vero ipsius basis extremitas sinistra, et C dextra.
side [this passage]: II . But it is difficult to describe them unless they have been
Licet enim quae hic scribimus commode non possint in Planitie figurari, ea tamen annotatic
found by this art.
per elementa memoriae (ut dicunt) localis vice futura est. Ac iam intelligatur Plana Super­
ficies AD, super Cono iacens secundum longitudinem: quae quidem Superficies unica sui linea
The reader will immediately see that Cardano has not based his proof on
Conum tanget. Constat enim ipsa infinitis lineis rectis: et Conus infinitis Circulis. Linea vero
the principal demonstration given by Provenzale (which was in the tradi­ recta, sive secet Circulum, sive tangat, eum in unico puncto secat et tangit. Ea autem linea
tion of the De duabus lineis) but rather on the obscure second demonstra­ tangens Conum sit AD, eiusdem scilicet notae cum superficie Plana in qua est. Intelligatur
tion given in the appended discor so. I suspect that he made this choice be­ rursus altera superficies EF, quae ipsam AD superficiem secet ad perpendiculum in linea
cause the principal demonstration had already been clearly given by recta, quae et ipsa sit EF, parallelus ipsi AD lineae (intelligo punctum E, a parte verticis Coni:
punctum vero F, a parte basis eiusdem) simulque secet Conum utrinque. Ex qua Coni sec­
Werner, while the second proof was presented so obscurely by Provenzale
tione efficietur figura Trilatera quae Hyperbole dicitur, cuius duo latera hinc inde extant
that he thought it needed to be refashioned. In so doing this, Cardano has in Superficie Coni: basis vero ipsius secat basin BC Trigoni ABC, inequaliter quidem, sed ad
reconstructed the cone and abandoned Provenzale’s effort to treat the angulos rectos: et utrinque ad peripheriam basis Coni (qui Circulus est) educitur. Latus autem
problem simply by projecting the cone on its base. Unfortunately, Car­ Hyperboles quod extat a parte AD Superficiei dicatur GH (altero latere innominato, ut cuius
dano’s reconstruction was itself quite loose, and indeed Barozzi felt it neces­ hoc loco nullus sit usus futurus) ut punctum G sit a vertice, H vero a basi ipsius Coni. Axis
autem ipsius Hyperboles non aequidistabit lateri Trigoni opposito, quod est AB: sicut ex
sary to present a corrected version of it (see Text B below, Extract III,
superioribus apparet. Sic enim esset Parabole, eo quod axis Hyperboles maiorem facit an­
Second Demonstration, marg. pg. nos. 66-76) and to single out Cardano’s gulum cum AC latere Trigoni secto ad punctum G quam faceret axis Paraboles.
errors (ibid., Extract VI, marg. pg. nos. 180-91). At any rate, we should “ Ac iam duas lineas habemus: unam EF in superficie AD secta, et in secante EF com­
note that Cardano’s proof is similar to that of Conclusion 11 of the Speculi muniter existentem: alteram GH latus Hy(45)perboles. Suntque eae duae lineae a veteribus
almukefi compositio. Cardano’s construction of the hyperbola and its deprehensae quae, quum in uno sint Plano (utraque enim in ipsa superficie EF) et perpetuo
fiat proprior altera alteri, nunquam tamen concurrunt. Quod illi longa quidem demonstra­
asymptote is essentially the same as that found in the earlier work. For,
tione ostenderunt, eamque enarrat Cardanus plenius: ad quem remitto Lectores. Habet enim
while the earlier author generated the hyperbola by passing a plane through illa Demonstratio multa probationum capita scitu digna, e Geometria deprompta, ad con­
the cone that was parallel to the plane of an axial triangle of the cone, firmandum lineae appropinquationem: sed quae tamen sensui per se est obvia. Nam quum ipsa
Cardano constructs the hyperbola by cutting the cone with a plane that is GH, latus scilicet Hyperboles, perpetuo et gradatim ascendat, eleveturque versus Coni basin,
perpendicular to a plane that is itself tangent to a side of the cone. In Cardano’s nimirum versus punctum F: propterea quod circuli, [ex] quibus Conus ipse constat, continue
352 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 353

Barozzi (see Text B below, Extract VI, marg. pg. nos. 197—201) and except Paleotti (see Text B below, Extract II, marg. pg. no. 11). There he speaks
for Peletier’s mention of Maimonides and Cardano he scarcely belongs to of the work “ coming forth into light” under your (i.e. Camilio’s) name and
our investigation of the fortunes of the medieval traditions of the asymp­ our whole academy (i.e. the Academia Bolognese).” This date agrees well
totic problem in the sixteenth century. with the internal evidence. The latest work on the asymptotic problem
The final stage of the medieval traditions of the asymptotic problem is cited by Barozzi is that of Jacques Peletier (1563), which I have mentioned
found in Francesco Barozzi’s omnibus Admirandum illud geometricum above. Brief notice of Commandino’s translation of the Conics (which ap­
problema tredecim modis demonstratum: Quod docet duas lineas in peared in the year of 1566) is of interest because Barozzi says that before
eodem plano designare, quae nunquam invicem coincidant, etiam si in the Latin Conics “ were to be edited by Federigo Commandino,” he him­
infinitum protrahantur: et quanto longius producuntur, tanto sibiinvicem self had corrected Memmo’s translation and used the same Greek manu­
propriores evadunt (Venice, 1586). Its preface is dated 1566 and the work script as had Commandino, thus leaving the impression that they were both
was apparently written during his visit to Bologna in 1565-66. At least this working on the Conics at about the same time (see Text B below, Extract
is the apparent deduction from the preface, which is dedicated to Camillo IV, marg. pg. no. 92).64
It is not my intention to discuss the whole work of Barozzi, which is very
maiores fiunt versus id ipsum F punctum: et stet ipsa recta EF in sua superficie EF im­
long, but rather to limit myself to producing extracts that illustrate the very
mutabilis: proculdubio ipsa GH, secundum augmentum Circulorum, continue etiam proprior significant parts of it that came indirectly from the medieval traditions. It
fiet ipsi EF lineae: et tamen cum ea nunquam concurret. Intelligatur enim extendi ipsa EF is quite evident from Extract I of Text B that Barozzi had read all of the
superficies cum Cono interminate: tum linea GH aut concurret cum EF in linea AD aut extra published authors who had mentioned or treated the asymptotic problem
eandem. In ipsa AD, minime: quia oporteret ipsam quoque EF cum AD concurrere: quum from Giorgio Valla to Jacques Peletier. Before producing thirteen ways of
tamen positae sint paralleli. Neque extra AD lineam concurret: quia, quum EF sit in Super­
ficie AD, et GH in superficie Coni: oporteret ipsam AD superficiem tangere Conum in
demonstrating the existence of the curved and straight lines that have the
puncto concursus: quod fieri non potest, quum probatum sit, Superficiem AD unica sui linea asymptotic relationship, he gives a long section on definitions, postulates
AD tangere Conum. Nunquam ergo concurrent GH et EF, quod fuit demonstrandum. and axioms that I shall omit in order to concentrate on his demonstra­
“ Ea est inventio antiquorum, quae quidem acuta et Geometrica prorsus est: sed quae ex tions. In this section he describes an instrument that he had invented to
quo genere linearum sit ipsa GH, non explicet. Nonnullis enim videri possit recta, propterea draw conics and also one that was communicated to him by Jacomo Con-
quod rectissime procedere videtur in Superficie Coni. Sed lex Geometrica non permittit
esse rectam: quia duas lineas rectas non parallelos, protractas aliquando concurrere, constat
tarini, who had himself received it from Count Giulio Thiene.65 Following
ex Conversione definitionis Parallelorum. Et quidem nulla recta linea potest in Super­
ficie Coni extendi nisi quae a puncto verticis ducatur ad basin: sitque latus talis Trigoni, quale 64 P. L. Rose, “ A Venetian Patron and Mathematician of the Sixteenth Century: Francesco
designavimus ABC. Nam quum linea recta a puncto ad punctum ducitur, utrunque punctorum Barozzi (1537-1604),” Studi Veneziani, N. S., Vol. 1 (1977), pp. 119-78. On page 149 Rose
in eodem Plano esse intelligitur. Atqui duo G et H puncta, non in plana, sed in Conica sunt says that it was probably through the Accademia Bolognese that Barozzi met Commandino.
superficie: et quae ab altero ad alterum ducitur recta in Conica superficie non est, sed in Incidentally, by a slip, Rose speaks (p. 143) of MS Dublin D.2.9 (390) as containing Gerard of
corpus penetrat. At dices, duo quoque puncta A et D in Coni esse superficie simulque in Cremona’s translation of the De mensura circuli of Archimedes. It is rather the version I have
Plana. Verum id quidem est: sed ipsa AD linea, quum sit secundum ductum lateris Trianguli attributed to Plato of Tivoli (set Archimedes in the Middle Ages, Vol. 1, pp. 16-29). Earlier
Conum describentis, id habet libertatis, ut possit deserere Superficiem Conicam: et suo accounts of Barozzi’s life and works were given by B. Boncompagni, “ Intorno alia vita ed ai
ductu planam superficiem creare. Scilicet si in(46)telligatur punctum in media AD immobile, lavori di Francesco Barozzi,” Bullettino di bibliografia e di storia delle scienze matematiche
super quod ipsa circunducatur, non dubium quin creetur Circulus cuius ipsa erit Diameter, e fisiche, Vol. 17 (1884), pp. 795-848, G. Spiazzi, “ Barozzi, Francesco,” Dizionario bio-
qui quidem Circulus iacebit super Coni superficie et ipsam tanget in sola linea. At linea GH grafico degli Italiani, Vol. 6 (1964), pp. 495-99, and Μ. N. Boyer, “ Barocius, Franciscus,”
idem efficere non potest, quia non est in eadem superficie plana cum axi ipsius Coni. Quaprop­ Dictionary of Scientific Biography, Vol. 1 (1970), 468.
ter si ipsa moveri intelligatur in orbem super puncto sui medio, subsultabit circunductio, ob 65 Ed. of 1586, p. 29: “ Placet autem hic Instrumentum quoddam a nobis olim inventum
inaequalitatem interiacentis superficiei Conicae. apponere, quod Conicam superficiem, ipsosque Conos, tam Rectos, quam Scalenos; tum
“ At vero quoniam in Plano haec deliniari non possunt, attentionem requirunt. Vel Conus Rectangulos, tum Obtusangulos, tum etiam Acutangulos commode generat: Necnon tres
extruendus, et linea aliqua ductilis in ipsius superficie pro latere Hyperboles aptanda est. Sed Conicas Sectiones, Parabolem scilicet, Hyperbolem, et Ellipsim aptissime describit. Ad cuius
et ipsam Hyperboles descriptionem, et duarum linearum propinquitatem citra concursum, a Instrumenti nostri similitudinem Circinus quoque simplex duorum crurium fabricari potest
nobis modo expositam, qui facilius assequi cupiet, sumat materiam aliquam tractabilem, ut (ut Clarissimus, eruditissimusque vir Iacobus Contarenus alter aetatis nostrae Archimedes
raphanum, quem in Coni figuram designet: tum ex charta conficiat duas illas superficies EF et me nuper commonefecit huiuscemodi Circinum repertum, sibique ostensum, ac traditum
AD, quarum haec super Cono incumbet: illa vero et superficiem AD, et Conum ipsum secabit, fuisse ab Illustrissimo Comiti Iulio Tiene, viro praestantissimo, omnibus in scientiis,
lineis sectionum in charta depictis.” I omit the further analysis (pp. 46-48) as being un­ Arteque Militari egregie versato) quo etiam facillime tres iam dictae Conicae Sectiones
related to my treatment of the medieval traditions of the asymptotic problem. I have made the designantur. Cuius Circini alterum crus, quod circunvoluendum est, concavum esse debet,
usual changes of punctuation and orthography and have italicized the letters marking points habens in concavitate stylum dentatum mobilem, qui contrahendo se, ac protrahendo
and magnitudes. Below Fig. 7.461 have indicated the changes that Peletier made in Cardano’s cuiusdam denticulatae rotulae, et laminae circa eam clavo circunvolutae artificio sursum,
figure. deorsumque feratur. Nam si planum in quo Sectiones ipsae Conicae designandae sunt
354 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 355

this section o fprincipia, Barozzi gives three propositions which he has re­ ment of greater circle oq and consequently of less curvature, it is evident
worked from Werner’s Libellus (see Text B below, Extract III, marg. pg. that the versine of segment pr is less than the versine of segment oq [i.e. rq
nos. 49-56). He fashions the first of these propositions from Werner’s Ele­ < qx]." Barozzi’s corollary states: “ In the smaller segments of unequal
ment XVIII, the second from Element XVII, and the third from Element circles that have equal bases (i.e. chords), the parts of the diameters of
XIX (of which he was particularly critical; see Extract VI, marg. pg. nos. these circles that bisect both the segments and their bases are unequal,
175-79). and the part of the diameter of the smaller circle is larger and that of the
Then follows the first of thirteen methods of demonstrating the asymp­ larger circle is smaller. But in the larger segments [of these circles] the part
totic relationship (ibid., Extract III, marg. pg. nos. 57-66). It is essentially of the diameter of the larger circle is larger and that of the smaller circle is
a paraphrase of Element XX of Werner’s work, and thus is equivalent to smaller.” This is an instance of Barozzi’s overelaborate proofs. For in fact,
Proposition V of the De duabus lineis. It is evident that Barozzi had not at least in the case of the Speculi almukefi compositio, Conclusion 11, the
seen the De duabus lineis itself because he points out Werner’s lapse in not statements concerning the sines and versines of the larger and smaller
finding at the end of the proof the minimal distances between the curve and circles are patently obvious and hardly need the ponderous proof given by
the straight line and claims that nobody but Cardano had found these lines Barozzi. Particularly unnecessary are the second halves of the lemma and
(ibid., Extract VI, marg. pg. no. 178). But, as I noted in my description of its corollary that concern the larger segments of both circles.
Proposition 5 of the De duabus lineis (see Chap. 2 above), the author of Following the exposition of the second case presented by the author of
that treatise had done so, and surely Barozzi would have mentioned this the Speculi almukefi compositio (which Barozzi knew in Gogava’s edition
had he seen the medieval tract. of 1548 and which he entitled Innominatus . . . de Sectione Conica,
Barozzi’s second demonstration of the asymptotic relationship (ibid., quae Parabole dicitur) and by Fine, Barozzi then presented the first case
Extract III, marg. pg. nos. 66-76) constitutes an effort to reconstruct given in the earlier works (see Text B below, Extract III, marg. pg. nos.
Cardano’s proof, of which he (Barozzi) was critical on the grounds of its 88-90), which demonstrated the existence of a curve and straight line hav­
obscurity and its inclusion of logical errors (ibid., Extract VI, marg. pg. nos. ing the desired asymptotic relationship but not lying in the same plane. This
180-91). case presents rather lines that both lie on the surface of the cone. Barozzi’s
Barozzi’s third demonstration (ibid., Extract III, marg. pg. nos. 82-87) account needs no further elaboration. This case is followed by the one con­
comes directly from Oronce Fine’s De speculo ustorio, Proposition X, cerning two curves that have an asymptotic relationship. It was taken from
which, we recall, was itself taken from Conclusion 11 of the Speculi al- the third case given in the Speculi almukefi compositio, Conclusion 11, and
mukefi compositio. It concerns the case where the straight line and the not from Fine since that case is missing in Fine’s work. Barozzi demon­
curve are in the same plane, i.e ., the second case given by the author of the strates that the curves can either be in the same plane with a common
Speculi almukefi compositio and by Fine. Before presenting the proof of asymptote or they can be in different planes that are parallel to an axial
his third demonstration, Barozzi presents a lemma not given by the earlier triangle and have the asymptotic relationship to one side of the axial tri­
authors (ibid., Extract III, marg. pg. nos. 76-82): “ Equal straight lines in angle. Barozzi’s account of the third demonstration and its additions
unequal circles intercept unequal arcs in both the larger and smaller seg­ should be read in conjunction with his critique of Fine’s text (ibid., Ex­
ments [of these circles]; thus in the smaller segments [the equal lines] inter­ tract VI, marg. pg. nos. 192-96). I have not included the fourth, fifth and
cept a greater arc from the smaller circle and a lesser arc from the larger sixth demonstrations and the auxiliary Apollonian geometry included with
circle, while in the larger segments they intercept a greater arc from the them (ed. cit., pp. 91-156); they comprise solutions by Apollonius and
larger circle and a lesser arc from the smaller circle.” From this lemma is Pappus and so these demonstrations were not a part of the medieval tradi­
deduced a corollary that is applied to Barozzi’s third demonstration, a tions of the problem. However, I have given the beginning of the section
corollary that is merely alluded to in Conclusion 11 of the Speculi almukefi that served to lay the groundwork for Apollonius’ demonstration (Text B
compositio, where its author says that “ since pr [Fig. 4.13c] is the seg- marg. pg. nos. 91-92). It is there that Barozzi mentions the translations of
the Conics made by Memmo and Commandino and his own efforts toward
parallelum quidem Axi Instrumenti nostri vel immobili cruri Circini iam dicti positum fuerit, revising Memmo’s translation.
Parabole dubio procul describetur: Si vero planum ipsum Axi vel cruri iam dicto non paral­ After the material from Apollonius and Pappus, Barozzi gives his seventh,
lelum, sed inclinatum versus Instrumenti seu Circini summitatem sit, Hyperbole designatur. eighth, and ninth demonstrations of the existence of the curve and straight
Si autem planum Axi sive Cruri ipsi non parallelum, sed e contrario in partem oppositam,
scilicet versus inferiorem Instrumenti vel Circini partem inclinatum ponatur, Ellipsis
line with an asymptotic relationship (ibid., Extract V , marg. pg. nos. 156-
describitur. Quippe quod Instrumentum, necnon Circinum ipsum clare figurae sequentes 63). These demonstrations were drawn from Werner’s Libellus: the seventh
ostendunt [Fig. 7.46bis].” and eighth from Element XXI and the ninth from the lemma to that ele­
356 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 357

ment. Incidentally, there is a systematic error in Barozzi’s seventh demon­ Since Profiat’s discussion of the asymptotic problem appeared as a com­
stration (see Fig. 7.56) when he assumes the length BG such that BG2 ment to Chap. 1.73 of Maimonides’ Guide o f the Perplexed, Barozzi pref­
= F E E C . Comparison with Werner’s text and a careful examination of aces his treatment of the discussion with the full text of Maimonides’
Barozzi’s argument show that Barozzi meant to say everywhere that chapter and a philosophical discussion of it (ed. cit., pp. 251-59), all of
BG2 = F C E C . For it is clear that he could only arrive at the equation which I omit here as being of no concern to my mathematical investigations.
GH2 = EC2 + 2 DE -EC by using BG2 = FC EC. The eighth demonstra­ I have a few final words on the Extracts that make up Text B. As in the
tion differs only slightly from the seventh and needs no comment. Nor does case of the texts of Werner’s Libellus in Chaper 6 and of Fine’s De speculo
the ninth, which is a straightforward adaptation of Werner’s lemma. ustorio in this chapter, I have made some changes in punctuation and
Barozzi’s tenth and eleventh demonstrations (ed. cit., pp. 164-73) de­ capitalization (at least in the demonstrations) in order that the aberrant
rive from Peletier’s work and have no interest for our investigation of the punctuation of the original may not confuse the reader. I have also removed
medieval traditions. I have omitted these demonstrations in my extracts. the grave accents from the adverbs (e.g. I give “ vero” instead of “ vero” ),
After the eleventh demonstration Barozzi inserts his digressions against expanded “ e ” to “ ae” , changed consonantal “ u” to “ v ” and vocal “ v ” to
Werner, Cardano, Fine, and Peletier (see Extract VI, marg. pg. nos. 175— “u” , converted the ampersand to “ et” , and changed “ ij” to “ ii” . I have
202), which attempt to show the defects he found in their accounts as he also corrected the text’s frequent “ indirectum” to “ in directum” when “ in
reworked his demonstrations. Many of his criticisms are rather trivial, a straight line” was clearly intended by the author. As usual, I have itali­
though there is no doubt that he helped to make sense out of Cardano’s cized the letters marking points and magnitudes. Barozzi’s marginal
adaptation of Provenzale’s supplementary demonstration (and thus also to legends have been converted to footnotes to permit a reduction in the mar­
make sense of Peletier’s even more confused presentation of Cardano’s gins of the printed pages. The marginal numbers are, of course, the page
demonstration). Immediately following Barozzi’s digressions against his numbers of the edition of 1586. No English translations have been given
sources, he has presented a complete reworking of the tract of Moise Proven- since the extracts are essentially the summation and correction of the
zale, which Barozzi entitles Libelli Rabbi Moysis Narbonensis dilucidatio various proofs I have given earlier in the volume with translations.
(Text B below, Extract VII, marg. pg. nos. 209-49). Barozzi has recast This then completes my account of the medieval traditions of conic sec­
the work in a more strictly geometrical form with proper general enuncia­ tions from the first introduction of the fragments of the Arabic version of
tions and cogent proofs. He has eliminated the numerical examples which Apollonius’ Conics translated by Gerard of Cremona in the twelfth century
Provenzale had substituted for proofs. In its recast form the principal part to the summarizing efforts of Francesco Barozzi in 1565-66. It should have
of the tract emerges as quite similar to Barozzi’s first demonstration (and become evident to the reader that the medieval traditions centered pri­
thus similar to the demonstrations found in the De duabus lineis and in marily in optical investigations, even when purely mathematical con­
Werner’s Libellus). Provenzale’s principal proof is labeled by Barozzi as cerns were introduced. Such was the case of the above-noted fragments
the twelfth demonstration of the existence of lines having the asymptotic translated by Gerard of Cremona (see Chap. 1), since that translation was
relationship. Barozzi also attempts to analyze and correct the demonstra­ made to serve as the introduction to Gerard’s translation of Alhazen’s De
tion given by Provenzale in his appended discorso. In Barozzi’s form of speculis comburentibus. The same is true of the material on the hyperbola
proof it becomes the thirteenth and last of the demonstrations of the asymp­ useful for solving the neuseis that appear in the Perspectiva of Alhazen
totic problem that he thought worthy of presentation. As I have already (again see Chap. 1), which may also have been translated by Gerard of
noted, Barozzi concluded that the projection of the cone on the plane of its Cremona, for the neuseis were necessary to demonstrate optical proposi­
base does not lead to a useful analysis of the problem, and hence he returns tions. The optical context is equally evident for the material on the parab­
to the cone itself for a three-dimensional presentation of the proof. Though ola appearing in the Perspectiva of Witelo (see Chap. 3), and for the
this treatment of the second proof of Provenzale is the last acceptable similar material given by the anonymous author of the Speculi almukefi
proof, Barozzi adds one more quite hopeless demonstration given by compositio (see Chap. 4), by Jean Fusoris in his Libellus (see Chap. 5), by
Profiat Duran (and falsely attributed to Rabbi Samtou [Ibn Shem-Tob] Giovanni Fontana in his notes on the parabola which appear in the principal
by Barozzi— see above, n. 41) in his commentary to the Hebrew text of the manuscript of Gerard’s translation of the De speculis comburentibus
Guide o f the Perplexed, published with Maimonides’ text in Sabionetta, (ibid.), and by Regiomontanus in his notes to the Speculi almukefi com­
1553. The demonstration was translated by Barozzi from the Hebrew and positio (ibid.).
then commented on by him (see Text B below, Extract VIII, marg. pg. nos. The medieval concern with optics and particularly with the explanation
259-69). As Barozzi shows it is so hopelessly muddled and inept that it is and construction of a paraboloidal burning mirror was significant in stress­
difficult to know what the author intended. At any rate, I believe that its ing the importance of the focus of the parabola, which is designated in the
ultimate origin lay in Conclusion 11 of the Speculi almukefi compositio. medieval texts as the point of combustion or point of convergence of the
358 ARCHIMEDES IN THE MIDDLE AGES THE LAST STAGES 359

rays reflected from the parabola. Though not mentioned by Apollonius, sions of the De duabus lineis and the Speculi almukefi compositio. Wite­
the focus of the parabola played an important part in Alhazen’s De speculis lo’s Perspectiva was published first in 1535 (reprinted in 1551) and again in
comburentibus,in Witelo’s account of the parabola, in the Speculi almukefi 1572 (this time with Alhazen’s Perspectiva added). The Speculi almukefi
compositio, in the Libellus of Jean Fusoris and in Regiomontanus’ notes. compositio was published by Gogava in 1548 in the version which I have as­
The interest in the parabola as a reflecting curve encouraged a new two- signed in Chapters 4 and 5 above to Regiomontanus. Finally, the same
dimensional description of the parabola as a curve from which all rays work appeared once more in a paraphrased version (based on the original
parallel to an axis must be reflected to a single point (see the discussion of Speculi almukefi compositio) that was published by Oronce Fine in 1551.
Fusoris’ Libellus in Chap. 5). The distance of that point from a point given The chief lacunae in the printed works that contained the medieval treat­
as the vertex of the curve was the burning distance (equal to one fourth ment of conics were (1) the actual text of the De duabus lineis, which,
of the latus rectum). In the course of the treatment of the parabola as a though employed by Werner, was never published, (2) the Libellus of Jean
reflector the latus rectum itself was freed from the cone, when it was shown Fusoris, which was neither published, nor, so far as I can tell, used by
that the double ordinate through the focus is equal to the latus rectum. sixteenth-century authors, and (3) the notes which Regiomontanus added
Furthermore, Fusoris and Regiomontanus show an interesting tendency to the Speculi almukefi compositio and which perhaps were known to
to consider the properties of a parabola in terms of magnitudes related to Werner but never published. And finally the reader should observe that the
the parabola solely in its plane. It is of interest, however, that when the medieval influence began to wane toward the middle of the sixteenth cen­
authors of the medieval tracts discussed the generation of a parabola from tury when three versions of Apollonius’ Conics were executed (those of
a cone, it was ordinarily the right-angled right cone from which the genera­ Memmo, Maurolico, and Commandino, with the first and the third im­
tion proceeded. This was important, for example, to the account in the mediately printed) and the Apollonian text began to be widely studied with
Speculi almukefi compositio, for the author of that tract, in Conclusion 4, a consequential preponderant influence on those who treated conics.
was able to show that the latus rectum of a parabola so generated was twice
the axial segment between the vertex of the parabola and the point of inter­
section of the axis of the parabola with the axis of the cone. He thus re­
discovered the form of the parameter that had been used by Archimedes
(see Chap. 4 above).
The medieval concentration on the optical uses of conic sections is no
doubt the reason that the sundry material on conics that appeared in the
works of Archimedes and Eutocius’ commentaries thereon which were
translated by William of Moerbeke exerted no discernible influence on the
development of the mathematical treatment of conics in the Middle Ages.
Similarly the medieval optical context for conics may also explain why Wil­
liam of Moerbeke’s translations of the Conics and Eutocius’ commentary
thereon (if indeed such translations were made, as I suggested in Chap. 3
above) also had no significant influence, except on Witelo.
The major exception to the pervasive optical context of the treatment of
conics in the Middle Ages was the translation by John of Palermo of an
anonymous tract on the asymptotic relationship of a curve (hyperbola) and
a straight line (its asymptote) which was entitled De duabus lineis (see
Chap. 2 above), a work that exerted some influence on the Speculi almukefi
compositio and later on the Libellus of Johann Werner. Through the lat­
ter it influenced the extensive summary of the various solutions to the
asymptotic problem put together by Francesco Barozzi in 1565-66 and
published in 1586.
We have seen in the last two chapters how virtually all of the medieval
accounts of conic sections became available in the sixteenth century.
Werner presented in 1522— to be sure in a novel way— the basic conclu­
 FINAEI DE SPECULO USTORIO 361

A. multa de supradicta sectione parabola, quae ex recto atque rectangulo cono

Orontius Finaeus, desumitur, suo more demonstraverit, nulla tamen arte videtur edocuisse,
qualiter inflexa seu curva eiusdem sectionis parabolae linea (a qua totum
De speculo ustorio, ignem ad propositam distantiam constat pendere negotium) fuerit describenda. Ait enim propositione
generante, liber unicus. Ex quo duarum linearum 3r quadragesimaquarta praeallegati libri no/ni, lineam quam dicimus peri-
phaeriam sectionis (parabolae velim intelligas) inveniat industria operantis,
semper appropinquantium et nunquam concurrentium etc. Ex aliis porro qui de ea re tractarunt, unicum vidimus incerti nominis
colligitur demonstratio authorem, ex Arabica lingua in Latinam adeo perplexe ac involute con­
versum (quemadmodum saepius in vertendis libris peccare solent, qui
Lutetiae linguas tantummodo callere, artes vero de quibus agitur, prorsus ignorare
1551 videntur), ut vix sensum aliquem ex ipsa potuerimus elicere litera: aut
unicam conspicere figuram quae eidem literae responderet. Decreveramus
2r / Ad clarissimum simul et eruditum virum, Do. Ioannem Massonium, nihilominus ipsum librum nostrum in meliora supprimere tempora: ni
torquatum equitem, Serenissimi Angliae ac Hyberniae Regis a secretis proborum quorundam ac studiosorum virorum authoritas in publicum
sanctioribus, eiusque oratorem apud Hernicum Gallorum Regem chris- tandem prodire iussisset. Quantum autem caeteros omnes hac in parte
tianissimum: Orontii Finaei Delphinatis, in sequentem librum de ustorio vicerimus, unicuique recto ac candido lectori submittimus diiudicandum.
speculo, praefatio. Hunc porro quantuluncunque laborem nostrum, tibi generoso ac modis
Si qua geometricarum figurarum vis sive potestas, ornatissime vir, atque omnibus erudito viro, nostri nominis dudum observantissimo, dicandum
rectilineorum angulorum proprietas, ex rebus ipsis naturalibus depre­ esse censuimus: tum in primis, ut pignus aliquod nostrae erga te voluntatis
hendatur: hoc maxime in speculorum ustoriorum clarescere videtur arti­ et obedientiae posteris relinquamus: tum etiam, ut patronum habeamus
ficio. De quibus varii non infimae authoritatis ac eruditionis mathematici, in Anglia, qui lividorum calumnias reprimere dignetur et possit. Fieri enim
diversos olim conscripsere tractatus: quos omnes Vitellio, tum subtilitate, non poterit, ut quae tu ipse iuste probaveris, ab aliis studiosis non facile
tum multitudine propositionum, longe videtur superasse: ut illius testatur recipiantur. Vale, ex musaeolo nostro, Lutetiae Parisiorum mense Octobri.
Perspectiva, decem libris absoluta: in qua nullum speculi genus praeter­ M.D.L.I.
misit, eorum potissimum, quae radiis solaribus exposita, ad communem
eorundem radiorum concursum, ignem accendere possunt. Inter omnia
porro specula quae ustoria nuncupantur: ea longe fortiorem atque celeriorem
videntur efficere combustionem, quae sic excavata sunt, ut in eorum super­
3 v / Antonii Mizaldi Monsluciani in Orontianum
ficiem incidentes radii solares ad unum certum et commune punctum
refrangantur. In huiuscemodi nanque speculis, ob universalem eorundem Speculum carmen ad Lectorem
2v radiorum collectionem et / unitam proinde virtutem (quae fortior est
dispersa) ignem celeriter et intense generari, ipso sensu fit manifestum. Lector amice, vides quam prosit Orontius Orbi,
Hoc autem ei soli videtur accidere speculo, quod in formam sectionis recti Quamque iuvet doctos, voce, labore, manu.
atque rectanguli coni, quae parabola dicitur, fuerit excavatum. Ex ipsius Hic speculum fabricat, mira quod repperit olim
itaque Vitellionis postremarum propositionum libri noni demonstra­ Arte Syracusius nobilis ille senex.
tionibus, et Apollonii Pergaei conicis elementis, una cum elementis geo­ Quo Regis vires elusit: et igne triremes
metricis ipsius Euclidis, ab hinc annis duodecim, subtile recollegimus et Illinc concepto perdidit innumeras.
demonstravimus artificium, unico libro comprehensum: quo in primis ip­ Ast usum docuit tantum: dat Orontius artem
sius recti atque rectanguli coni sectio describitur parabola: dein praefatum Veram: nec solam, nam dat utrunque tibi.
ustorium construitur speculum, in formam eiusdem parabolae sectionis Arte mathematica fabricam hic demonstrat, et usum
fabricatum. Hoc enim solaribus radiis directe suppositum, ignem ad datum Urentis speculi, denique materiam.
intervallum, super inflammabili materia poterit accendere: nempe ad tan­ Res sane mira et non visa, absque ignibus ignes
tam longitudinis distantiam, quantus (!) fuerit semidiameter oblati cuiusvis Excitat: et nullo fomite cuncta cremat.
circuli, dummodo aliquantula eiusdem circuli sectio, super quopiam Ignes e speculo longe iaculatur: et unum
oblato plano describi vel facile possit. Quanquam enim praefatus Vitellio Phoebo Vulcanum sedulo conciliat.
Quaeque procul posita in flammas convertit: et igne
360
362 ARCHIMEDES IN THE MIDDLE AGES FINAEI DE SPECULO USTORIO 363

Nusquam prospecto, quod cupit, hoc abolet. circa rectum sunt angulum circunvoluto descriptus: in cuius centrum
In terram e caelo flammam transferre Prometheus praefatus coincidit axis.
Est ausus, quondam caelica regna petens. 4. Conica porro superficies est quae ab ipso latere rectangulum sub­
Hoc maiora potest insignis Orontius, ima tendente complete revoluto causatur, in supremum axis desinens apicem:
Ex terra flammas caelitus, ecce, trahit. qui vertex eiusdem coni dicitur.
Quod nemo faciet, nisi sit de gente Deorum: 5. Omnis autem linea recta quae a vertice coni in basis deducitur cir-
Talem Finaeum dicere iure potes. cunferentiam: latus sive longitudo eiusdem coni nominatur. Hoc itaque
Ingenium mirare novum, mirare laborem modo descriptus conus, rectus in primis dicitur, quoniam illius axis ad
Summum, praeclaris Gallia nota viris. rectos super basin consistit angulos: et rectangulus ideo vocitatur, quoniam
Felix hoc partu nimium es, nimiumque superba duo illius latera ex opposito constituta rectum continent angulum: quae
Nanque tuum nomen iam super astra tulit. cum sint invicem aequalia, fit ut idem conus isosceles haud dissimiliter
Finis. appelletur. Quemadmodum obiecta coni figura aede [Fig. 7.9] utcunque
demonstrat, a rectangulo et isoscele triangulo abe circa latus ab complete
revoluto descripta: Cuius vertex est punctum a , axis autem ab linea recta,
basis vero ede, cir- (/ dei.) circulus cuius centrum b, latus denique sive
4r / Orontii Finaei Delphi- longitudo eiusdem coni, ac, vel ae, linea recta.
natis, regii mathematicarum 6. Sectio autem ipsius recti atque rectanguli coni, quae parabola dicitur,
Lutetiae professoris, De Speculo ustorio et quae ad nostrum maxime videtur spectare negotium, est plana super­
ignem ad propositam distantiam ficies, inflexa quadam li/nea per coni educta superficiem, et dimetiente
generante, Liber unicus. basis ipsius coni terminata: quae isoscelis atque rectanguli trianguli plano,
quod per verticem et axem eiusdem coni transire diffinitur et sub binis
Ut ad susceptam Speculi parabolici (sic enim iure possumus appellare) lateribus et dimetiente basis continetur et conum bifariam dirimit, ad rectos
descriptionem, praeter ipsius Euclidis elementa geometrica, quae veluti consistit angulos.
certa atque nota supponimus: nonnulla ex conicis elementis Apollonii 7. Sagitta porro sive dimetiens ipsius sectionis parabolae est linea recta
Pergaei, quae potissimum ad nostrum facere videntur institutum, prius dif­ quae eorundem planorum communis est differentia, et alterum eiusdem
finire, ac exponere duximus operae pretium. Deinde ipsum Speculum para- trianguli latus secat, alteri vero fit parallela.
bolicum in primis mathematice, postea manuali construere ac polire 8. Vertex autem eiusdem sectionis parabolae est ipsius sagittae sive
docebimus artificio: unde congruam omnium speculorum tum materiam, dimetientis punctum supremum.
tum poliendi rationem, quilibet sagax et et (/ dei.) industrius artifex col­ 9. Basis vero proprie nuncupatur ipsum latus rectum sectionis sive
ligere vel facile poterit. Ab ipsius itaque recti ac rectanguli coni, eius- conicae basis dimetiens.
demque parabolae sectionis diffinitione (caeteris conorum atque sec­ 10. Et quae huic sectioni tam auctae, quam diminutae sectiones descri­
tionum differentiis, veluti parum suscepto negocio conducentibus prae­ bentur parallelae: itidem parabolae nuncupantur: quarum diminutae, hoc
termissis) nostrum feliciter auspicemur exordium. est, truncatae a basi coni rectanguli sectiones, suscepto potissimum viden­
Recti atque rectanguli coni, eiusque sectionis quae parabola dicitur, tur inseruire negocio: cuius causa infra exponetur.
11. Omnes autem lineae rectae eidem basi sectionis ac invicem paral­
Diffinitiones XII. lelae ab altera inflexae lineae parte in reliquam ad rectos cum sagitta coin-
cidentes angulos: lineae ordidinis (/ ordinis) eiusdem sagittae sive ordinatim
1. Conus rectus atque rectangulus dicitur, figura solida, sub plano cir­ extensae nuncupantur: quas omnes sagitta bifariam dividit: et unaquaeque
culari, et ea quae ab eodem circulo in punctum unum coarctatur superficie, illarum est basis illius partis sectionis parabolae quae inter eandem lineam
comprehensa: a rectangulo et isoscele triangulo (uno eorum quae circa et sagittae verticem comprehenditur.
rectum sunt angulum latere manente fixo) integre revoluto descripta, non 12. Ea denique linea ordinis quae per medium totius sagittae punctum
differens a rotunda pyramide. inter eius verticem et basin sectionis aut (si mavis) centrum basis ipsius
4v / 2. Axis igitur ipsius recti atque rectanguli coni est, ipsum trianguli latus coni transire diffinitur latus erectum eiusdem sectionis parabolae atque
fixum, circa quod idem revolvitur triangulum rectangulum. partium ipsius sectionis a qualibet ordinis linea ad sagittae verticem com­
3. Basis vero eiusdem coni est, circulus a reliquo latere eorum quae prehensarum vocitatur. Harum autem postremarum diffinitionum exem-
364 ARCHIMEDES IN THE MIDDLE AGES FINAEI DE SPECULO USTORIO 365

pia, ex subscripta coni potes elicere figura: cuius vertex a [Fig. 7.10], et Corollarium.
axis a b , illius autem basis circulus cdef, triangulum vero rectangulum per
5v axem et verticem coni ace. Sectio autem parabola dgf / sub inflexa et para- Cum igitur radii solares in concavi cuiuspiam speculi superficiem inci­
bolica linea dgf et recta dbf comprehensa: cuius vertex g, dimetiens sive dentes ad unum quoddam certum et commune punctum ex omni parte
sagitta bg, et medium illius punctum h, basis sive latus rectum ipsa dbf. refranguntur: necessum est huiuscemodi speculum, inter omnia ustoria
Ordinis porro lineae kl et mn et quaecunque his similes: quarum erectum 6v specula, ce/lerrimae atque intensissimae fore combustionis. Tale autem
latus est ipsa kl. Caetera pervia sunt. solum esse demonstrabimus, quod instar supradictae sectionis parabolae
fuerit excavatum.
Lemma sive assumptum. His in hunc modum expositis atque diffinitis: demonstrandae sunt
aliquot propositiones, ipsius sectionis parabolae discutientes accidentia, et
Quod autem sectio communisque differentia qua superficies conica et ad mathematicam intelligentiam propositi speculi, in formam eiusdem
plana superficies per axem et verticem coni deducta rectangulum ac parabolae sectionis excavandi, perutiles ad modumque necessariae.
isosceles triangulum efficiat: tum ex ipsius coni praemissa descriptione, Quarum prima est haec.
tum ex ipsa trianguli rectanguli figura a quo huiuscemodi conus describitur,
fit per sese manifestum. Sunt enim ipsius communis et triangularis PROPOSITIO I.
sectionis latera, in superficie conica, ab illius vertice in basis peri-
phaeriam deducta: et proinde aequalia adinvicem, atque rectum an­ Si in recti atque rectanguli coni superficie duo suscepta fuerint puncta
gulum comprehendentia. Basis vero eiusdem sectionis communis est ipsius quae per ipsa punta recta connectitur linea, cadit intra conum: ni producta
conicae basis dimetiens, eandem bifariam dividens. Ipsa ergo sectio com­ in directum per ipsius coni transierit verticem.
munis, cum per verticem et axem ipsius coni transire difiniatur: conum Sit rectus atque rectangulus conus abcde [Fig. 7.12], in cuius superficie
ipsum bifariam de necessitate dividit. duo signentur puncta/ , g: aio quod connexa e x /in g linea recta cadit intra
conum: ni producta in directum pervenerit ad ipsius coni verticem, ef­
POSTULATA EX PERSPECTIVA ficiens eiusdem coni latus. Demittantur enim ex a vertice, per ipsa puncta
desumpta. f , g, in basis circunferentiam, bina coni latera afc et agd: connectaturque
recta linea cd, per primum postulatum geometricum. Cum igitur basis coni
Subroganda deinde sunt communia quaedam theoremata, ab omnibus sit circulus in cuius circunferentia sunt duo puncta c, d, cadit itaque recta
Perspectivae authoribus comprobata, quae postulata nuncupabimus. 7r cd intra circulum bcde, per secundam tertii elementorum Euclidis. / Trian­
Quorum primum est huiusmodi. gulum propterea acd conum subintrat ac ipsum dividit. In triangulo porro
6r 1. Omnes radii solares in datam quamvis speculi super/ficiem incidentes acd continetur fg linea recta. Eadem itaque linea rectafg cadit intra datum
se habent veluti quaedam rectae lineae: et proinde in geometricis demon­ conum abcde.
strationibus eam vim obtinent quam lineae mathematicae servant adinvicem. Idem aliter. Aut (si velis) suscipiatur in rectafg puntum h : et ex a vertice
2. Omnes radii solares in planum coincidentes speculum faciunt angulos per h in basin cd ipsius trianguli acd recta deducatur linea ahk. Cum igitur
incidendae angulis reflexionis semper aequales. De angulis intelligo ad eam recta cd cadat intra circularem basim ipsius coni: cadet et recta linea ahk
rectam lineam relatis quae una cum ipsis radiis in eodem plano consistit. intra eundem conum, et proinde illius punctum h, et ducta consequenter
3. Omnes insuper radii solares in convexi cuiusvis aut concavi speculi per idem punctum h, linea recta fhg. Quod ostendere oportebat.
superficiem incidentes ad praefatos angulos aequales refranguntur: sed ad
eam relatos superficiem planam, vel in eadem superficie iacentem lineam Corollarium.
rectam, quae per incidentiae punctum transire diffinitur, et ipsam con­
cavam vel convexam speculi superficiem in eodem incidentiae puncto Omnes itaque lineae ordinis praefatae sectionis parabolae intra conum
solummodo tangit. Haec duo ultima postulata ex subscriptis utcunque ipsum cadunt.
clarescunt descriptionibus. In quibus radius Solis ab reflectitur in punctum
PROPOSITIO II.
c [Fig. 7.11], efficiens angulum incidentiae angulo reflectionis aequalem:
sive radius incidat in planum de, vel in convexum fg , aut in concavum Si rectangulus et erectus conus plano secetur ipsi basi parallelo, sectio
speculum hl, ab ipso plano de in eodem puncto b contanguntur: semper communis eiusdem plani et conicae superficiei erit periphaeria circuli
enim angulus abd angulo cbe causatur aequalis. cuius centrum in ipsius coni axe constituetur.
4. A quacumque autem speculi superficie radii solares sic reflectuntur Esto rectangulus et erectus conus abcd [Fig. 7.13], cuius vertex a , basis
ut in unum coincidant et refrangantur punctum: in ipso solo puncto ignem circulus bed, et illius centrum e , axis vero coni ae, planum autem secans
generari est possibile. conum ipsi basi parallelum/g/i, per quod transeat axis coni ad punctum /:
366 ARCHIMEDES IN THE MIDDLE AGES FINAEI DE SPECULO USTORIO 367

suscipianturque in conica superficie eidem plano communia puncta/, g , h . Sit rectus iterum atque rectangulus conus abed [Fig. 7.14], cuius vertex
Dico quod linea communis intersectionis eiusdem plani et conicae super­ a basis circulus bed, et ipsius circuli centrum e , axis vero coni ae, trian­
ficiei, transiens per ipsa puncta/ , g, h, est circunferentia circuli cuius cen­ gulum porro per axem bifariam dividens ipsum conum abd. Sectio demum
trum est punctum /. Non erunt enim fg , gh, et hf eiusdem sectionis por­ parabola, ad rectos cum eodem triangulo consistens angulos, cfg, cuius
tiones lineae rectae: caderent enim intra conum, per antecedentem primam latus rectum ceg, vertex punctum / , sagitta vero ef, et medium eiusdem
propositionem: et proinde non forent in ipsius coni superficie, contra 8v sagittae punctum h , erectum / denique latus eiusdem sectionis khl. Aio ip­
hypothesin. Obliquae sunt igitur eaedem fg , gh, et hf lineales intersec­ sum erectum latus khl ipsius e f sagittae fore duplum. Cum enim triangulum
tiones: et tota consequenter fgh circunvolutio itidem obliqua. Aio quod abd sectionem parabolam super sagitta e f ad rectos dirimat angulos: coin-
7v et / circularis, cuius centrum est punctum /. Deducantur enim ex a vertice cidet erectum latus khl ad rectos itidem angulos cum plano eiusdem trian­
per ipsa puncta/ , g , h in basis circunferentiam coni latera afb, a g e, et ahd: guli abd. Transeat itaque circulus quidam conum dividens per ipsum erec­
et connectantur eb, ec, et ed semidiametri, similiter //, Ig et Ih lineae tum latus khl, basi parallelus, cuius circuli dimidium sit mln, dimetiens
rectae, per primum postulatum geometricum. His ita constructis, vero recta mn. Erit itaque ipsius circuli centrum in axe ae illius vero cir­
palam est triangula a eb , alf esse invicem aequiangula: est enim / / ipsi eb ex cunferentia in superficie conica, per antecedentem secundam proposi­
hypothesi parallela, et proinde angulus alf aequalis ipsi aeb, nec non an­ tionem. His ita constructis, quoniam latus erectum khl ad rectos cum
gulus afl angulo abe interiori et opposito ad easdem partes aequalis, per sagitta e f consistit angulos: cadit igitur hl ad rectos itidem angulos cum
vigesimamnonam primi elementorum Euclidis: et angulus qui ad verticem plano ipsius trianguli abd, et proinde cum dimetiente mn. Et quoniam an­
a utrique triangulo communis est. Haud dissimiliter ostendetur, triangulum gulus qui ad punctum / rectus est, per trigesimam primam tertii elemen­
aec triangulo alg, necnon triangulum aed triangulo alh, fore itidem aequi- torum Euclidis (nempe consistens in semicirculo mln), deducta igitur ex
angulum. Aequiangulorum porro triangulorum proportionalia sunt latera angulo recto qui ad / in basin mn perpendicularis Ih est media proportionalis
quae circum aequales angulos, et similis rationis quae aequalibus angulis inter ipsius basis segmenta mh et hn, per corollarium octavae sexti eorun­
latera subtenduntur, per quartam sexti eorundem elementorum. Sicut dem elementorum. Quod igitur ex mh fit quadratum ad id quod ex hl eam
igitur ae ad eb sic al ad//: atque sicut ipsaae ad ec sic eadem al ad ipsam Ig: habet rationem quam recta mh ad rectam hn, per corollarium decimaenonae
sicut praeterea eadem ae ad ipsam ed sic praefata al ad ipsam Ih. Atqui e b , eiusdem sexti elementorum. Atqui mh ipsius hn (ut infra demonstrabitur)
ec, et ed aequales sunt adinvicem, ut pote eiusdem circuli semidiametri: est dupla. Quod igitur ex mh fit quadratum eius quod ex hl describitur est
et eadem ad aequales eandem habet rationem, per septimam quinti praedic­ duplum. Ipsa porro mh aequalis est oppositae de, per trigesimam quartam
torum elementorum. Eadem itaque al ad ipsas //, Ig, e* Ih eandem quoque primi eorundem elementorum: parallelogrammum est enim dehm quadri-
rationem obtinet. Ad quas autem magnitudines eadem magnitudo eandem laterum. Eidem rursum de aequalis est eb: utraque enim est semidiameter
habet rationem, ipsae sunt aequales, per nonam eiusdem quinti elemen- ipsius bed circuli. Binae igitur mh et eb aequales sunt adinvicem: et ab
8r torum: / Aequales igitur invicem sunt //, Ig, et Ih. Haud dissimiliter quotquot aequalibus rectis aequalia describuntur quadrata. Aequalia rursum quadrata
ex puncto / in orbitam fgh deducentur lineae rectae, tum invicem, tum ad idem quadratum eandem habent rationem, per septimam quinti praedic­
unicuique ipsarum //, Ig, Ih, ostendentur aequales. Circulus est igitur torum elementorum. Quod igitur ex eb fit quadratum duplum est eius quod
orbicularis linea fg h , per ipsius circuli diffinitionem: et illius centrum /, per 9r ex hl. Id rursum quod ex eb fit quadratum duplum est eius quod ex ef, / per
nonam tertii eorundem elementorum. Quod demonstrandum susceperamus. quadragesimam septimam primi eorundem elementorum: rectangulum est
enim atque isosceles ipsum efb triangulum, nempe simile toti abd. Quod
igitur ex eb fit quadratum ad ea quae ex e f et hl describuntur quadrata ean­
Corollarium. dem habet rationem, nempe duplam. Aequum est ergo quadratum quod ex
Quae igitur a coni vertice in praefatum planum eidem basi conicae paral­ e f ei quod fit ex hl, per nonam quinti ipsorum elementorum. Aequalia porro
lelum comprehenditur figura conus est, et toti cono similis cuius basis est quadrata sunt quae ab aequalibus rectis describuntur: aequalis est itaque
ipse circulus fg h , communis existens eiusdem plani et conicae super­ recta e f ipsi hl. Sed ipsius hl dupla est khl, et ipsius propterea e f itidem
ficiei differentia. dupla: quae enim sunt aequalia, eiusdem sunt dimidium, per septimae com­
munis sententiae conversionem. Latus igitur erectum khl duplum est sagit­
tae ef. Quod fuerat ostendendum.
PROPOSITIO III.
Lemma sive assumptum.
In recti atque rectanguli coni sectione parabola latus erectum duplum
est sagittae eiusdem sectionis inter axem coni et ipsius sectionis verticem Quod autem mh dupla sit ipsius hn in hunc modum confirmatur. Quo­
comprehensae. niam triangulum ebf triangulo fhn est aequiangulum: parallela est enim
368 ARCHIMEDES IN THE MIDDLE AGES FINAEI DE SPECULO USTORIO 369

hn ipsi eb, et angulus propterea fhn aequalis angulo feb , necnon angulus sic recta kl ad rectam fo , per vigisimamsecundam quinti elementorum. Sed
fnh angulo fbe itidem aequalis, per vigesimamnonam primi elementorum: quadrata sunt in dupla ratione laterum, ut ex ipso decimaenonae sexti
et reliquus angulus qui ad/utrique triangulo communis. Est igitur, per elementorum elicitur corollario: et proinde ipsa latera in subdupla ratione
quartam sexti eorundem elementorum, ut eb ad hn sic e f ad fh , sed e f ip­ lOv quadratorum. Re/cta igitur kl ad rectam fo duplo maiorem rationem quam
sius fh dupla est: et eb igitur ipsius hn itidem dupla. Ipsi autem de, et ad ipsam op. Tres itaque lineae rectae kl, o p ,fo sunt invicem propor­
proinde ipsi eb, aequalis praeostensa est mh: et aequalia eiusdem sunt tionales, per decimae diffinitionis quinti eorundem elementorum con­
duplicia, per sextae communis sententiae conversionem. Dupla est igitur versionem. Sicut igitur latus erectum kl ad perpendicularem op sic eadem
mh ipsius hn. Quod proxima fuerat assumptum demonstratione. perpendicularis op ad sagittae segmentumf o . Quod oportuit demonstrasse.
Idem quoque ostendere licebit, ubi eadem perpendicularis op inter latus
9v / PROPOSITIO IIII. erectum kl et basim sectionis cd fuerit data.
Si in eadem recti atque rectanguli coni sectione parabola, inter ipsius
Corollarium I.
sectionis verticem et latus erectum, a parabola in sagittam perpendicularis
quaepiam ordinetur: idem latus erectum ad ipsam perpendicularem ean­ Quadratum igitur quod ex data quavis perpendiculari describitur
dem rationem habebit quam eadem perpendicularis ad eam sagittae partem aequum est rectangulo quod sub erecto latere et comprehensam (/ compre­
quae ipsum verticem et eandem perpendicularem intercipitur. hensa) inter ipsam perpendicularem et sectionis verticem sagittae partem
Resumatur proximae et antecedentis propositionis figura, una cum ex­ continetur. Ostensum est enim ut kl ad op sic eadem op ad fo: corollarium
positis ipsius figurae partibus, cui superaddatur perpendicularis et ordinata ergo subsequitur, per decimamseptimam sexti elementorum.
linea op [Fig. 7.15]: recipio itaque demonstrandum ut latus erectum khl ad
ipsam perpendicularem op sic eadem perpendicularis ad sagittae partem Corollarium II.
of. Describatur ergo per punctum p circulus rps, ipsi basi bed parallelus,
Quacunque praeterea linea ordinis in parabola sectione designata, si per
cuius dimetiens sit ros, ipsi nhm dimetienti consequenter parallelus. Erit
illius extremitates et verticem sectionis describatur circulus (quod per
igitur os ipsi hm aequalis, per trigesimamquartam primi elementorum quintam quarti elementorum fieri potest) centrum ipsius circuli de neces­
Euclidis, atque rursum op media proportionalis inter so et or: quemad­ sitate erit in sagitta sectionis, per corollarium primae tertii elementorum:
modum et hl media itidem proportionalis inter mh et hn, per trigesimam quoniam sagitta ipsam ordinis lineam bifariam et ad rectos dirimit angulos.
primam tertii et corollarium octavae sexti eorundem elementorum. Per
corollarium insuper decimaenonae eiusdem sexti elementorum, erit Corollarium III.
lOr ut quadratum ex mh ad quadratum quod ex hl sic ipsa / mh recta ad rectam
hn: atque rursum ut quadratum quod ex so ad quadratum quod ex op sic Pars insuper dimetientis eiusdem circuli per caput sectionis et lineae
eadem recta so ad rectam o f {! or). Sunt itaque duo ordines quatuor pro­ ordinis extremitates delineati inter ipsam lineam ordinis et circunferentiam
portionalium quantitatum, et in utroque ordine primae quantitates aequales eiusdem circuli versus basim sectionis comprehensa erecto lateri sectionis
sunt adinvicem, similiter et tertiae: est igitur ut secunda quantitas ipsius erit aequalis. Nam per trigesimam primam tertii, et corollarium octavae
primi ordinis ad secundam ordinis secundi sic quarta eisdem (/ eiusdem) sexti elementorum, proposita pars dimetientis ad dimidiam lineae ordinis
primi ad quartam ipsius secundi, hoc est ut hi ad op sic hn ad or. Sicut porro partem (quae perpendicularis appellatur) eandem habet rationem quam
hn ad or sic recta h f ad rectam of, triangula enim fh n , for sunt invicem Hr ipsa dimidia pars seu perpendicularis ad reliquam partem ipsius di/metien-
aequiangula, et proinde ut nh ad hf sic ro ad ipsam of, per quartam sexti tis quae ad sectionis parabolae finitur verticem. Eandem quoque rationem
elementorum; et permutantim quoque, per sedecimam quinti eorundem praeostensum est habere latus sectionis erectum ad ipsam perpendicu­
elementorum, ut nh ad or sic hf ad ipsam of. Habes igitur ut hf ad o f sic larem, seu dimidiam lineae ordinis partem. Quae autem ad eandem, ean­
quadratum ex hl ad id quod ex op describitur. Ipsi porro hl ostensa est dem habent rationem aequalia sunt adinvicem, per nonam quinti elemen­
aequalis ef: et ab aequalibus rectis, aequalia describuntur quadrata. Est torum. Aequalis est igitur proposita pars ipsius dimetientis eidem erecto
igitur ut hf ad ipsam o f sic quadratum ex e f ad quadratum quod ex o p . Et lateri sectionis parabolae. Quemadmodum ex praefata sectione parabola
quoniam kl dupla est ipsius ef, et eadem e f ipsius fh itidem dupla: est igitur, cfg [Fig. 7.16] ad iustam rationem antecedentis coni rectanguli abed con­
per corollarium decimaenonae sexti elementorum, ut quadratum ex kl ad structa, et directe ob oculos exposita, colligere vel facile est. In qua latus
quadratum quod ex e f sic eadem recta kl ad rectani/Tz. Sicut rursum quadra­ erectum khl, et linea ordinis nop, atque descriptus circulus fn m p, circa
tum quod ex ef ad quadratum quod ex op sic ostensa est fh recta ad rectam rectilineum triangulum fn p , illiusque circuli dimetiens f m , coincidens cum
f o . Erit igitur ex aequa ratione ut quadratum ex kl ad quadratum quod ex op sagitta ef: nam pars ipsius dimetientis om ipsi kl est aequalis.
370 ARCHIMEDES IN THE MIDDLE AGES FINAEI DE SPECULO USTORIO 371

PROPOSITIO V. tionem in ipso puncto e. Si enim non tetigerit, secabit ergo sectionem:
Si recti atque rectanguli coni sectionem parabolam recta linea tetigerit, idque aut super e punctum versus a sectionis verticem, aut sub eodem
et a contactu in sagittam utrinque productam duae ceciderint lineae rectae, puncto e versus basim bdc. Secet igitur in primis (si possibile fuerit)
altera quidem perpendicularis in sagittam, altera vero tangenti perpendicu­ in puncto h: et ab ipso puncto h in sagittam ad perpendicularis deducatur
laris, et sagitta convenerit ipsi tangenti: latus sectionis erectum eandem hl, per duodecimam primi elementorum. Et quoniam ag ipsi a f data est
rationem habebit ad partem sagittae quae inter perpendiculares quam aequalis: maior erit igitur ag ipsa al; secetur itaque ipsi al aequalis am, per
sagittae pars inter perpendicularem interiorem et praedictarum linearum tertiam eiusdem primi elementorum: erit ergo Im dupla ipsius am. Patuit
llv concursum exterius comprae/hensa ad ipsius sagittae partem quae in­ autem in antecedentis quartae propositionis demonstratione rationem fa
ad ipsam al duplo maiorem esse ratione fe ad Ih. Sicut porro fe ad Ih sic
teriorem perpendicularem et verticem sectionis intercipitur.
Esto sectio parabola abc [Fig. 7.17] ad iustam rationem saepius as­ fg ad gl, per quartam sexti et sedecimam quinti elementorum. Ratio igitur
sumpti coni rectanguli delineata: cuius vertex a, sagitta vero ad, basis fa ad ipsam al duplo maior est ratione ipsius fg ad gl. Sicut rursum fa ad
ipsam al sic eadem fg , ipsius a f dupla, ad rectam Im, duplam ipsius al, per
bdc, et erectum latus efg. Tangat autem sectionem recta quaedam linea hl
quindecimam quinti elementorum. Ratio itaque fg ad Im duplo maior est
in puncto h: et ab ipso puncto h decidat perpendicularis in sagittam, quae
sit hm , ipsi vero tangenti hl perpendicularis hn. Producta demum sagitta ad ratione ipsius fg ad gl. Prima itaque fg ad tertiam Im duplo maiorem ra­
tionem habet quam eadem fg ad secundam gl; sunt igitur invicem propor­
utrasque partes, conveniat in primis cum tangente hl in ipso puncto /: cum
ipsa vero perpendiculari hn in ipso puncto n. His constructis, aio erectum tionales, per decimae diffinitionis quinti elementorum conversionem: sicut
latus efg eandem habere rationem ad partem sagittae mn quam pars Im ad quidem fg ad gl sic eadem gl ad Im: quae potestate sunt quatuor, nam gl,
ipsam ma. Cum enim angulus Ihn sit rectus, et ab ipso angulo recto qui ad consequentis primae rationis et antecedentis secundae fungitur officio.
h in basin In perpendicularis deducatur hm : erit ut Im ad mh sic eadem mh I3r / Tota igitur fg ad totam gl se habet ut gl ablata ad ablatam Im. Et reliqua
ad ipsam mn, per corollarium octavae sexti elementorum. Sed per ante­ igitur If ad reliquam mg se habebit ut tota ad totam, per decimamnonam
cedentem quartam propositionem, latus erectum efg eandem rationem quinti elementorum. Atqui prima fg maior est tertia gl: et reliqua igitur //
habet ad ipsam mh quam eadem mh ad m a. Cum autem tres lineae fuerint reliqua mg maior erit, per decimamquartam eiusdem quinti elementorum.
proportionales, quod sub extremis continetur rectangulum aequum est ei Sunt autem If et gm , per constructionem, adinvicem aequales: quae simul
quod a medio fit quadrato, per decimamseptimam sexti elementorum. impossibilia sunt. Non secat igitur eg recta sectionem parabolam inter
I2r / Utrunque igitur rectangulum, et sub Im in ipsam mn atque sub efg in ipsam datum punctum e et ipsius sectionis verticem.
ma comprehensum, eidem quadrato quod ex mh describitur est aequale: et Aio quod neque infra versus basim bdc [Fig. 7.19]. Secet enim (si pos­
proinde alterum aequale alteri. Sunt itaque duo rectangula, et conse­ sibile fuerit) in puncto n: et ducatur no super ad perpendicularis, per duo­
quenter parallelogramma invicem aequalia, et unum angulum uni angulo decimam primi elementorum: seceturque ar ipsi ao aequalis, per tertiam
eiusdem primi elementorum. Et quoniam a f ipsi ag est aequalis, erit itaque
aequalem habentia, nempe rectum recto: habent igitur quae circum aequales
gr aequalis ipsi fo: et or consequenter dupla ipsius ao. Erit rursum, ex
angulos latera reciproce proportionalia, per decimamquartam eiusdem
quartae propositionis demonstratione, ratio ipsius oa ad rectam a f duplo
sexti elementorum. Sicut igitur latus erectum efg ad rectam mn sic recta
maior ratione on adf e , et ratione consequenter ipsius og ad gf, veluti supra
Im ad partem sagittae ma. Quod oportuit demonstrasse.
deductum extitit. Ratio quoque ipsius or (quae dupla est ipsius ao) ad ip­
sam g f (quae dupla est ipsius af) eadem erit, quae ipsius oa ad ipsam af, per
PROPOSITIO VI.
quindecimam quinti elementorum: et proinde duplo maior ratione ipsius og
Si a dato puncto sectionis parabolae ipsius recti atque rectanguli coni ad ipsam gf. Prima itaque or ad tertiam g f duplo maiorem rationem habet
perpendicularis in sagittam deducatur, et producta sagitta extra verticem, quam secunda go ad eandem tertiam gf. Sunt igitur rursum invicem propor­
ei sagittae parti quae perpendicularem et verticem intercipitur aequalis' tionales, per ipsius decimae diffinitionis quinti elementorum conver-
recta forinsecus designetur: connexa ab illius extremo in datum punctum I3v sionem: sicut quidem ro ad og sic eadem og ad ipsam / gf. Tota propterea
linea recta sectionem tanget. or ad totam og se habet ut ablata og ad ablatam g f (ipsa enim og bis sumpta,
Resumatur proxima sectio parabola abc [Fig. 7.18], cuius vertex a, totius et ablatae fungitur officio), reliqua proinde gr ad reliquam or se
sagitta vero ad, et basis recta bdc. Datum autem sectionis punctum sit e, habebit ut tota or ad totam og, per ipsam decimamnonam quinti elemen­
I2v a quo decidat e f in sagittam ad perpen/dicularis: et producta sagitta ad torum. Tota porro or maior est ablata og: et reliqua proinde gr reliqua o f
partes verticis a, ipsi a f aequalis secetur ag, per tertiam primi elemen­ itidem maior. Atqui gr ipsi or aequalis praeostensa est: quae simul impos­
torum, et connectatur eg linea recta. Dico itaque rectam eg tangere sec- sibilia sunt. Non secat igitur recta eg sectionem abc inter datum punctum e
372 ARCHIMEDES IN THE MIDDLE AGES FINAEI DE SPECULO USTORIO 373

et basim bdc; patuit quod neque inter idem punctum e et eiusdem sectionis continetur rectangulum aequum est eis quae ab insecta et quolibet seg­
verticem a. Tangit itaque recta eg sectionem ipsam in eodem puncto e. mento divisae comprehenduntur rectangulis, per primam secundi elemen­
Quod expediebat ostendere. torum: et per primam sexti eorundem elementorum, sub eadem altitudine
et in basibus aequalibus consistentia rectangula parallelogramma aequaliav
Corollarium. sunt adinvicem. Quod igitur sub feg et ipsa ao continetur rectangulum
quater sub ea et ao contento rectangulo est aequale: et proinde una cum
Si recta igitur linea sectionem tetigerit parabolam, et a puncto contactus
quadrato quod ex oe aequum est ei quod ex el describitur quadrato. Atqui
in sagittam perpendicularis deducta fuerit, productaque sagitta ad partes
latus erectum feg eandem rationem habet ad perpendicularem ho quam
verticis cum tangente convenerit: erit versavice pars sagittae sectionis
ipsa perpendicularis ad sagittae partem oa, per antecedentem quartam
verticem et punctum contactus intercepta aequalis parti eiusdem sagittae
propositionem. Tres itaque lineae rectae feg , ho, et oa continue sunt pro­
quae inter ipsum verticem et eandem clauditur perpendicularem. Osten­
portionales. Quod igitur sub extremis feg et oa continetur rectangulum
sum est enim rectam ge tangere parabolam sectionem in ipso puncto e :
aequum est ei quod a media ho fit quadrato, per decimamseptimam sexti
ubi deducta perpendiculari ef pars sagittae a f posita est aequalis ipsi ag. Et
elementorum. Quae igitur ex ho et oe quadrata describuntur aequalia sunt
proinde recta ge tangente versavice sectionem parabolam in puncto e , et
quadrato quod ex el. Ipsis porro quadratis quae ex ho et oe describuntur
deducta e f perpendiculari, sagitta ipsi tangenti conveniat in punctum g:
aequum est quadratum quod ex eh, per quadragesimamseptimam primi
erit, a conversa demonstrandi ratione, pars sagittae ag aequalis ipsi af.
ipsorum elementorum: rectus est enim angulus eoh, per ipsam construc­
tionem. Aequalia porro quadrata sunt, quae ab aequalibus rectis describun­
PROPOSITIO VII.
tur: aequalis est propterea eh ipsi el, et angulus consequenter ehl angulo elh
Si a data quovis recti atque rectanguli coni parabolae sectionis puncto aequalis, per quintam eiusdem primi elementorum. Eidem rursum angulo
egrediatur linea recta sagittae parallela, altera vero cadat in punctum sagit­ elh aequalis est exterior et ad easdem partes consistens angulus nhm, per
tae medium, per quod transit latus erectum, atque in eodem puncto dato vigesimamnonam ipsius primi elementorum: parallela siquidem est hn ipsi
I4r alia quaedam linea recta sectionem tangat: angulus qui sub contin/gente dl, per constructionem. Aequalis est itaque angulus ehl eidem angulo
utrinque producta et ea quae in punctum sagittae medium ad partes verticis nhm. Quod in primis ostendendum susceperamus.
causatur aequalis est angulo qui ex linea sagittae parallela et eadem con­ Si autem angulus leh fuerit rectus, ut in sequenti figura [Fig. 7.21]: idem
tingente versus basim efficitur. rursum nihilominus concludetur. Cum enim angulus leh sit rectus ex
Esto rursum data sectio parabola abc [Fig. 7.20], cuius vertex a, sagitta hypothesi, is erit aequalis angulo / recto leg: et ipsa proinde ho perpendicu­
vero ad, et basis bdc, sitque ipsius sagittae punctum medium e, per quod laris, et coincidens propterea ipsi eg dimidio lateris erecti. Quare rursum
transeat latus erectum feg. Datum porro sectionis punctum sit h, et con­ ae ipsi al erit aequalis, per corollarium antecedentis sextae propositionis:
nexa linea recta eh, alia quaedam linea recta Ihm tangat eandem sectionem et tota proinde el dupla ipsius ea. Sed ipsius ea dupla est similiter eg, sive
in ipso puncto h , a quo decidat hn ipsi ad parallela. Aio itaque angulum ehl oh, ut ex antecedenti propositione tertia deductum extitit: et quae eiusdem
aequalem esse angulo nhm. Producatur enim sagitta versus a , similiter et duplicia sunt, aequalia sunt adinvicem, per sextam communem sententiam.
ipsa contingens Ihm, donec conveniant in punctum /. Trianguli itaque ehl Aequalis est igitur el ipsi eg: et angulus consequenter ehl angulo elh, per
angulus leh erit vel acutus, aut rectus, vel obtusus. quintam primi elementorum aequalis. Cui quidem angulo elh, aequalis est
Sit in primis acutus, ut in proxima figurae dispositione [Fig. 7.20]: et a angulus nhm, per vigesimamnonam eiusdem primi elementorum: et
puncto h in sagittam ad perpendicularis deducatur ho, per duodecimam proinde angulus ehl aequalis ipsi angulo nhm. Quod rursum fuerat
primi elementorum, quae de necessitate sectionis parabolae cadet inter ostendendum.
puncta a c ie . Cum igitur ae utcunque divisa sit in puncto o : quod igitur sub Esto demum angulus leh obtusus, ut in ea quae sequitur figurae descrip­
ea et altero segmentorum ao quater comprehenditur rectangulum, una tione [Fig. 7.22]: et a puncto h in sagittam ad perpendicularis deducatur
cum quadrato quod ex oe reliquo segmento describitur, aequum est ei quod ho, per ipsam duodecimam primi elementorum. Cadet igitur punctum o
ex ea et ao, tanquam ex una recta linea fit quadrato, per octavam secundi inter puncta d cie: eritque recta ao ipsa ae maior. Et quoniam, per corol­
elementorum: hoc est, quadrato ipsius el, nam, per corollarium ante­ larium antecedentis sextae propositionis, al ipsi ao est aequalis: maior erit
cedentis sextae propositionis, ao aequalis est ipsi al. Et quoniam, per ter­ propterea ipsa al dimidia sagitta ae. Secetur itaque eidem ae aequalis ar,
tiam propositionem antecedentem, latus erectum feg duplum est sagit- per tertiam primi elementorum. Reliqua igitur eo reliquae rl consequenter
I4v tae a d : idem ergo latus / feg quadruplum est ipsius ae. Id autem quod sub erit aequalis: et proinde tota or aequalis toti el. His praemissis, cum recta
duabus lineis rectis, quarum altera in quotcunque segmenta divisa est, oa utcumque dividatur in puncto e: quod igitur sub tota oa et altero seg-
374 ARCHIMEDES IN THE MIDDLE AGES FINAEI DE SPECULO USTORIO 375

mentorum ae quater comprehenditur rectangulum, una cum quadrato quod recta, tangatque praefatam superficiem a parabola sectione descriptam
ex reliquo segmento eo describitur, aequum est quadrato ex oa et ae, tan- I6v recta quaedam linea Im in ipso / quidem puncto g. Clarum est itaque an­
I5v quam ex una recta linea descripto, per octavam secundi elemen/torum: gulum Igf aequalem esse angulo mgh. Per datum siquidem punctum g et
igitur et aequale quadrato ipsius or, cum ae ipsi ar per constructionem verticem a sectio transit parabola, super basim bec perpendiculariter
sit aequalis: atque demum aequale quadrato ipsius el, quae eidem or erecta, et ei sectioni ex qua descripta est superficies similis et prorsus
aequalis praeostensa est. Atqui, per antecedentem tertiam propositionem, aequalis, quam quidem sectionem bifariam dividit axis ad. Et cum eidem
latus erectum feg duplum est ipsius sagittae da\ igitur et ipsius ea qua­ axi parallela sit recta gh , per constructionem, erit eadem gh in eodem plano
druplum. Comprehensum itaque subfeg et ipsaoa rectangulum aequum est cum ipsa ad: similiter et ipsas/", Per septimam undecimi elementorum, et
ei quod sub eadem oa et ae quater continetur (ut in prima huius parte, ex proinde recta Im, quae tangit superficiem, tangit similiter et eandem sec­
prima secundi atque prima sexti elementorum conclusimus), et ipsum tionem in ipso puncto g . Aequalis est itaque angulus Igf ipsi angulo mgh,
propterea sub feg et oa comprehensum rectangulum, una cum quadrato per ipsam propositionem septimam. Idem quoque subsequi necessum est
quod fit ex eo, aequatur quadrato quod fit ex el. Eidem praeterea rectan- de datis quibusvis aliis in concavum eiusdem superficiei coincidentibus
gulo quod sub feg et oa continetur aequum est quadratum quod ex ipsa oh lineis rectis.
perpendiculari describitur. Nam, per antecedentem quartam proposi­
tionem, oh est media proportionalis inter latus erectum feg et sagittae par­ Corollarium II.
tem oa. Unde, rursum per decimamseptimam sexti elementorum, compre­ In speculo itaque iuxta recti atque rectanguli coni sectionem parabolam
hensum sub extremis feg et oa rectangulum aequum est ei quod a media excavato, et Soli radianti directe exposito: omnes radii solares in con­
oh fit quadrato. Quae igitur ex eo et oh utraque fiunt quadrata aequalia sunt cavam eiusdem speculi superficiem incidentes in unum veluti commune
ei quod ex el quadrato describitur. Ipsis porro quadratis quae ex eo et oh punctum axis reflectuntur: quod tantum distat ab ipsius speculi vertice
describuntur aequum est id quod fit ex eh , per quadragesimam septimam quantum est dimidium sagittae illius sectionis parabolae, ad cuius rationem
primi elementorum: rectus est enim angulus qui sub eoh, per construc­ datum speculum constructum extitit.
tionem. Quod igitur ex ipsa el describitur quadratum aequum est ei quod Nam propter solaris corporis respectu totius globi terrestris (ne dum
fit ex eh: et proinde ipsa el recta eidem eh aequalis, et angulus conse- exigui speculi) excessivam magnitudinem, quae secundum Alphraganum
I6r quenter / ehl aequalis angulo elh, et aequalis propterea ipsi angulo nlm. est ut 166 fere ad 1, et maximam centri ipsius Solis a mundi centro distan­
Omnibus ergo modis, qui sub contingente et ea quae in punctum sagittae tiam, quam idem Alphraganus asserit continere semidiametrum ipsius globi
medium ad partes verticis causatur angulus aequalis est ei qui ex linea sagit­ terestris millies centies et septuagesies: fit ut omnes radii solares, in huiusce-
tae parallela et eadem contingente versus basim efficitur angulo. Quod modi speculum directe coincidentes, videantur paralleli. Quemadmodum
tandem oportuit demostrasse. (praeter eas quae ex ipsa distantia construi possent demonstrationes)
fidem facit umbrarum rectarum meridianarum aequalitas, quae per aequa­
Corollarium I. lium gnomonum ad notabilem distantiam sub eodem meridiano circulo con-
Si igitur a recti atque rectanguli coni sectione parabola circa sagittam I7r stitutorum causantur inter positionem: quae minime / offenderentur
integre revoluta superficies describatur, et in datum quodvis punctum aequales, si in ipso casu radiorum, iidem radii solares parallelam inter sese
concavitatis illius recta inciderit linea axi parallela, ab eoque puncto in non observarent distantiam. Se habent itaque praefati radii solares in ip­
medium sagittae punctum (per quod transit latus erectum) altera recta con­ sum coincidentes speculum veluti quaedam lineae rectae axi eiusdem
nexa fuerit: ipsae lineae rectae aequales conficient angulos cum ea linea speculi (dum Soli directe exponitur) parallelae. Sed omnes lineae rectae
recta quae praefatam superficiem a parabola sectione descriptam in eodem in concavam superficiem quae a recti atque rectanguli coni sectione para­
puncto contangit. bola describitur incidentes tales causant angulos cum singulis lineis rectis
Utpote, si ex data coni rectanguli sectione parabolaabc [Fig. 7.23], cuius in earundem incidentium punctis extremis superficiem ipsam contingenti­
vertex a, basis vero recta bc, et sagitta ad, circa eandem sagittam integre bus quales ab eisdem punctis in medium sagittae punctum connexae lineae
revoluta describatur parabola et excavata superficies abec: cuius basis sit rectae, per primum huiusce propositionis septimae corollarium. Et, per
bce circulus, et ipsius circuli centrum punctum d, diameter autem recta praemissum tertium postulatum, omnis radius solaris in huiuscemodi con­
bdc, et divisa sit ad sagitta (quae nomen axis adepta est) bifariam in puncto cavum incidens speculum angulum incidentiae angulo reflexionis facit
/ , cuius dimidium a f quartae parti lateris erecti eiusdem sectionis para­ aequalem: super plano (velim intelligas) quod ipsius speculi parabolici con­
bolae sit aequale. Incidat autem in punctum g concavitatis eiusdem super­ cavam superficiem in eodem incidentiae puncto contangit. Corollarium
ficiei parabolae recta linea gh , axi ad parallela, et connectatur g f linea itaque fit aperte manifestum. In cuius maiorem elucidationem, sequen-
376 ARCHIMEDES IN THE MIDDLE AGES FINAEI DE SPECULO USTORIO 377

sagittam praefatam longitudinem abc: eritque recta dc axis, et ce semi-

tem adiecimus figuram [Fig. 7.24], quae habet speculum parabolicum abc,
diameter basis eiusdem coni, et proinde dimidium basis eiusdem sectionis
cuius vertex b, axis vero bd, et in eodem axe quarta pars lateris erecti be,
parabolae. Si ducantur ergo per puncta b etc rectae lineae /g et hi, invicem
hoc est, dimidium sagittae sectionis parabolae, ad cuius rationem construc­
atque ipsi de parallelae, rectos cum abc utrobique causantes angulos, ac
tum est speculum. Radios denique solares inter caeteros annotatos/g, hi,
m n, in puncta concavitatisg, l , n coincidentes, ad ipsum punctum e refrac­ eidem abc utraque bg et b f secetur aequalis, et utraque ch et ci aequalis ipsi
tos. In quod quidem punctum e caeteros omnes coincidentes radios re- dc: erit fg latus erectum et hi basis eiusdem sectionis parabolae sub in­
frangi est operae pretium: ibidemque, applicata re combustibili, ignem flexa linea hfagi et eadem basi hei comprehensa.
generari. His praemissis clarum est, neque sectionem parabolam hai neque eius
partem/«g fore describendam ut per eam propositum fabricetur speculum:
Corollarium III. cum radiorum solarium reflexio futura sit in puncto b , medio quidem sagit­
tae abc, per quod transit latus erectum ft>g: haec enim sagitta ac, vel eius
Hinc rursum colligitur huiuscemodi speculum parabolicum, hoc est, dimidia pars ab, ineptae atque excessivae videretur esse magnitudinis.
iuxta recti atque rectanguli coni sectionem parabolam excavatum, inten­ Resecanda est igitur mediocris quaedam ac non incongrua particula ip­
sioris atque celerioris esse combustionis quam aliud quodvis speculum sius distantiae propositae ab, et fabricanda sectio parabola diminuta cuius
datum. Nullum etenim praeter supradictum parabolicum offenditur specu- basis sit chorda sectionis circuli ex abc descripti, veluti aim, cuius sagitta
I7v lum a cuius universa superficie radii solares in unum / commune punc­ est ak. Haec autem sagittae sive distantiae pars ak pedalis ad summum aut
tum reflectantur. Nam si aliquod tale dari posset speculum, maxime foret sesquipededalis (/) poterit esse quantitatis: etiam quantacunque fuerit
hemisphaericum concavum: sed in illo tot offenduntur reflexionum puncta oblata distantia ab. Quanto nihilominus maior extiterit ak, tanto maior
quot sunt incidentium radiorum orbiculares revolutiones, ut ex Vitellione erit chorda Im, et sectio parabola tanto maior, et proinde futurum speculum
aliisque authoribus Perspectivae facile deprehenditur. Solum itaque specu­ tanto consequenter maius. Unde solarium radiorum tanto maior multitudo
lum pro recti atque rectanguli coni sectione parabola constructum punctum in ipsum punctum b reflectetur ; ex quo subsequetur intentior atque celerior
habet commune in quod coincidentes radii solares universaliter refran- ignis generatio.
guntur. Et cum virtus unita fortior sit ipsa dissa (/ dispersa): fit ut in prae­ Fabricetur itaque (ut ad rem ipsam deveniamus) ex ligno quopiam solido,
fato speculo parabolico, et ad communem illius refractorum radiorum con­ veluti pyro vel nuce, corpus rectan/gulum sub aequidistantibus planis
cursum, ignis celerius atque intensius accendatur quam per aliud quodvis comprehensum: tantae ad minus longitudinis quanta est chorda Im, lati­
speculum datum. tudinis autem iuxta sagittam ak, et altitudinis ad ipsius ak dimidium [Fig.
7.26a]. Cuius quidem corporis longitudo, sub quatuor lineis rectis lati­
I8r / PROPOSITIO VIII. tudinis atque altitudinis lateribus parallelis et quadrilateram atque rectan-
Qua ratione sectio describatur parabola ad constructionem speculi con­ gulam comprehendentibus figuram, bifariam ex omni parte dividatur, ut ex
cavi ignem ad propositam distantiam generantis pernecessaria. obiecta corporis figura eisdem literis a , k, l, m et adiunctis η , ο , ρ insignita
Esto distantia, seu proposita longitudo, ab [Fig. 7.25]: quae in directum deprehenditur. Sumatur deinde virga quaedam lignea vel ferrea tantae ad
et continuum producatur ad partes quidem b . Et ipsi ab aequalis seceturhc: minus longitudinis, quanta est abc, in cuius extremorum altero stylus
hoc est, sumatur ipsius ab dupla, quae sit abc. Posito itaque speculi para- promineat acutus ipsi virgae orthogonus ad longitudinem ipsius altitudinis
bolici vertice in puncto a , omnes radii solares in ipsum speculum incidentes ao: in reliquo vero extremo cursorius circinus adaptetur brevissimo cus­
refrangi debent ex hypothesi et simul convenire in punctum b . Erit prop- pide una cum perstringente clavo insignitus [Fig. 7.26b]. Uti subscripta
terea data longitudo ab dimidium sagittae illius sectionis parabolae, ad cuius descriptio monstrat.
inflectionem propositum excavandum erit speculum: et tota consequenter Describatur consequenter in dato quopiam et ad libellam praeparato
ac sagitta eiusdem sectionis integra, atque semidiameter circuli qui est plano recta quaedam linea quae sit aequalis ipsi abc: et super ipsius lineae
basis recti atque rectanguli coni ex quo proposita sectio parabola (quae altero termino media linea alterutrius faciei praefati corporis rectanguli
diminuta nuncupatur) elicienda est. Fiat igitur recta de ad angulos rectos directe constituatur, sic quidem ut punctum (verbi gratia) o eidem extremi­
cum ac, et utraque da et ae eidem ac aequalis: connectanturque dc atque tati lineae supradictae adamussim respondeat. Et posito super reliquo eius­
I8v ce lineae re/ctae. Rectus erit itaque angulus d ee: uterque enim angulus dem lineae termino styli cuspide ipsius virgae seu praeparatae regulae,
dea et ace dimidius est anguli recti, velut ex quinta et trigesimasecunda atque circino cursore ad quantitatem longitudinis ipsius abc iustificato,
primi elementorum vel facile colligitur. Et proinde triangulum dee rectan- describatur in suprema praedicti corporis superficie secti olam [Fig. 7.26c]
gulum est atque isosceles: ad cuius integram revolutionem circa latus dc ei quae in praecedenti figura delineata est similis et aequalis. Opposita
rectus atque rectangulus conus describitur, cuius sectio parabola recipit in deinde ipsius corporis rectanguli facie sursum eversa, atque media prioris
378 ARCHIMEDES IN THE MIDDLE AGES
FINAEI DE SPECULO USTORIO 379
faciei lineola in directum supradictae lineae velut antea constituta, restringa-
I9v tur in/tervallum circini pro dimidia parte ipsius ak. Immoto semper (veluti Idem Aliter.
communi centro) ipsius styli cuspide: et super altera atque priori opposita Poterit et eadem sectio parabola diminuta alio (veluti sequitur) artificio
facie corporis sectio itidem circuli describatur rst, ipsalak (! lam) minor. In delineari. Supposito itaque primo huiusce propositionis diagrammate [Fig.
hunc quippe modum, ut utraque sectionis periphaeria ad eandem corporis 7.25], describatur recta quaedam linea abc [Fig. 7.29], cuius pars ab sit
partem inclinetur: et altera earum tangat latus faciei in qua describitur in aequalis duplo ipsius assumptae distantiae, hoc est, sagittae ipsius descrip­
ipso puncto a , reliqua vero ad mediam oppositae faciei partem solummodo tae a principio sectionis parabolae. Et a puncto b ipsi abc ad angulos rectos
perviat. Quemadmodum subscriptae videntur indicare formulae. excitetur bd, per undecimam primi elementorum, quae sit aequalis dimidiae
Connexis postmodum lineis rectis Ir et mt, abscindantur omnia quam chordae Ikm, hoc est, ipsi Ik sive km eiusdem antecedentis primi diagram­
rectissime fieri poterit extra praefatas circulorum sectiones et superficiem matis. Describatur postmodum semicirculus adc, cuius centrum in hunc
Imrt comprehensa: relinquetur enim portio quaedam truncata recti atque modum promptissime reperietur. Connexa ad linea recta, describatur an­
rectanguli coni, cuius basis est circulus a praefata linea recta abc descrip­ gulus ade aequalis angulo bad, per vigesimamtertiam primi elementorum:
tus. Huius autem truncatae portionis sive corporis figuram hic habes ob ubi enim de recta diviserit rectam ab (ut in ipso puncto e) illic erit centrum
oculos expositam, iuxtapraeassumptarum linearum rationem quam melius praefati semicirculi. Quibus absolutis, sectio bc in quotcunque partes in­
fieri potuit in plano repraesentatam. Quod si demum partes extra planam vicem aequales dividenda est: sit igitur eadem bc in quatuor partes (exem­
superficiem quae transit per puncta l, s, m, k, versus r et t prominentes, pli gratia) distributa. Describantur consequenter singuli semicirculi
quam apte fieri poterit resecentur prodibit tandem proposita sectio para- 2ir quorum dimetientes inter punctum a et singula / divisionum puncta ipsius
20r bola di/minuta sub inflexa linea Ism et recta basi Ikm comprehensa, cuius bc comprehendantur: notenturque singulae praedictorum semicirculorum
vertex erit punctum s, sagitta vero connexa linea rectask, ut haec ostendit intersectiones in ipsa perpendiculari bd contingentes, sub punctis quidem
figura, pendenter cum prius assumptis delineata [Fig. 7.27]. f , g, h: ut in figura. Exponatur rursum altera linea recta, saepius expressae
Exponatur enim ob oculos quadrilaterum rectangulum aosk [Fig. 7.28], sagittae a k primae descriptionis aequalis: quae sit Im succedentis descrip­
quod videlicet sumptum a principio corpus Imnp bifariam dividebat, cuius tionis [Fig. 7.30]. Haec postmodum recta Im in tot partes invicem aequales
unum latus est ak\ et connexa linea recta as per datum punctum 's, ipsi ao dividatur in quot divisa est ipsa bc. Et per singula divisionum puncta (ex­
parallela ducatur su , per trigesimam primam primi elementorum. Parallelo- cepto / extremorum altero) singulae ducantur lineae rectae invicem paral­
grammum erit igitur aosu quadrilaterum atque rectangulum. Et quoniam, lelae et ad rectos angulos cum eadem Im coincidentes. Descripto conse­
per trigesimamquartam primi elementorum, omnis parallelogrammi latera quenter circa Im circulo, ab ipsa parallela quae per m punctum educta est
quae ex opposito et anguli aequalia sunt adinvicem: aequum est propterea geminae secentur rectae ipsi bd aequales: et a sequenti parallela, duae
latus au ipsi os, atque ao ipsi su aequale. Sed os dimidium est ipsius ak, similiter rectae aequales ipsi bf: et a succedenti parallela, totidem aequales
similiter , per ipsam constructionem; utraque igitur au etsu et ipsa con­ ipsi bg. Et deinceps in hunc modum pro data parallelarum atque sectionum
sequenter uk eiusdem ak est dimidium. Tres propterea au, uk, su aequales multitudine ipsius bd. Tandem a sectione circuli quae per punctum / de­
sunt adinvicem, et uterque angulus qui circa u verticem rectus; basis igitur scribitur in singula praedictarum linearum extremalia puncta inflexa linea
as basi sk est aequalis, per quartam primi elementorum, et qui ad easdem praefatae sectionis parabolae describatur: ut in ipsa continetur figura.
bases consistunt anguli aequales adinvicem, et proinde quilibet eorum recti
dimidius, et qui sub ask continetur angulus rectus. Completo itaque trian­
Corollarium.
gulo axy, cuius utrunque latus ay etxy duplatae distantiae abc ipsius ante­
cedentis primae figurae sit aequale, et diviso ax latere bifariam in puncto z : In quanto plures igitur partes ipsa bc recta fuerit distributa: tanto prae­
si connectatur yz, ea erit perpendicularis super ax, per octavam proposi- cisior, hoc est, minus peccans, erit eadem inflexa linea ipsius sectionis
20v tionem et decimam diffinitionem primi elemen/torum, et ipsi propterea sk parabolae.
parallela, per vigesimamoctavam eiusdem primi. Ex supradictis ergo fit
manifestum rectam yz fore sagittam parabolae sectionis illius recti atque PROPOSITIO IX.
rectanguli coni qui a triangulo rectangulo axy circa latus xy complete
revoluto describitur et cuius basis est circulus ex ay delineatus. Hinc, per Ut speculum ipsum, iuxta prius descriptam sectionem parabolam ex­
praemissam parabolae sectionis diffinitionem, sk est sagitta diminutae sec­ cavatum fabricetur, ac poliatur, tandem ostendere.
tionis parabolae cuius basis est praefata chorda Ikm. Quod fecisse ac 2iv / Fabricetur igitur ex puro et electo calibe instrumentum quoddam
demonstrasse oportuit. moderate crassum et veluti scalprum in acutiem definens: quae quidem
acuties instar praefatae sectionis parabolae ad unguen sit efformata, atque
380 ARCHIMEDES IN THE MIDDLE AGES FINAEI DE SPECULO USTORIO 381

ita indurata ut vulgatum calibem seu ferrum depuratum facile discindat atque rectangula, tantae ad minus longitudinis quanta est basis sive latus
atque radat [Fig. 7.31]. Huius autem instrumenti hanc accipe formulam. 22v rectum praeparatae sectionis parabolae: latitudinis autem paulo ma/ioris
Postmodum, ex ipso vulgari calibe, seu ferro depurato, lamina quaedam quam sit illius sagitta: et crassitudinis ad summum digitalis: veluti obiecta
incurvata fabricetur, ad inflexam lineam eiusdem parabolae sectionis figura abcd utcunque demonstrat [Fig. 7.32]. In qua quidem tabella deline­
propemodum excavata, digitalis propemodum crassitudinis. Cuius quidem etur ac tandem excavetur sectio parabola iuxta antecedentis octavae
laminae superficies concava, ad iustam inflexae lineae parabolae ipsius propositionis traditionem praefigurata: cuius inflexa linea ad unguem ex­
praeparati et indurati instrumenti, per tornatilem et artificiosam illius cir- pressa sit aed. Praeparetur consequenter ex ligno congruo, aliave tractabili
cunductionem, radendo figuretur: ac demum subtiliter et optime poliatur, materia, corpus quoddam solidum, veluti aedf: cuius basis sit circularis,
quemadodum infra declarabitur. Habebis enim optatum speculum quod et ipsius circuli diameter aequalis lateri recto praefatae sectionis para­
solaribus radiis expositum ignem ad propositam distantiam super inflam- bolae: inflexa vero superficies inflexae lineae eiusdem porabolae (/), hoc
mabili materia (ut ex praeostensis fit manifestum) generabit. est, ipsi aed excavatae tabellae abcd, quaquaversum sine aliquo discrimine
Conditiones porro boni et electi calibis ad praefati instrumenti sive scal­ congruat.
pri parabolici constructionem necessarii sunt huiusmodi: lenitas videlicet Tandem mediante huiuscemodi parabolico corpore fiat typus ipsius
exterioris superficiei absque scissuris, frangendi facilitas, et partium speculi, ex sabulo, vel arsilla, instar Campanarum: et fundatur speculum ex
contingens in fractura splendor. Facilitas etenim fractionis ipsius calibis subscripta materia metallica, cuius superficies concava inflexam sive
duritiem arguere videtur: lenitas autem superficiei exterioris et claritas convexam superficiem eiusdem praeparati corporis parabolici ex omni
partium in fracturis debitam earundem partium continuationem atque parte contangat. Erit enim hoc modo ad rationem parabolae sectionis
mundiciam eiusdem calibis aperte manifestant. Induratio autem ipsius excavata.
calibis, quae caeteris potissimum in hoc videtur praestare negotio, est Recipe igitur boni aeris et bene purgati lib. I, stanni glacialis lib. semis,
haec. Exprimatur succus raphani, et cum eodem succo permisceatur aqua marcassitae albae lA lib., salis petrae lA lib. Deinde funde haec omnia
de lumbricis terrae contusis et expressis per pannum lineum: sic quidem ut simul. Quibus fusis, superpone laminam lardi, et move diu; cum autem
22r utriusque et / succi et aquae partes sint aequales. Et intra hanc mixturam spumaverit, proiice spumam. Et proiice hanc materiam intra paratum
praefatum instrumentum ex depurato calibe fabricatum, candens vel ig­ typum sive (ut vocant) modulum speculi. Quo infrigidato, extrahatur, et
nitum bis, terve, aut pluries extinguatur: fiet enim adeo solidum et durum, figatur illius convexum super excavatum asserem, aut alio quovis modo. Et
ut ferrum commune praeciososve lapides incidat non minus facile quam cum pumice rudi et aqua communi fricetur ipsius speculi concava et para-
plumbum vel stannum. bolica superficies, quatenus ablata fuerit illius asperitas, et unita videatur.
Reliquum est de politura ipsius speculi nonnulla subiungere. Huic itaque Postea fricetur cum lapide sulphuris. Sumatur consequenter tripolitum,
rei commodissimus est lapis emerillus appellatus, colorem habens fer­ et oleum olivarum, spuma stanni, creta crocea sive massicotus lapis: et
reum, veluti magnes. Melior tamen esse videtur cui color inest citrinus et 23r fricetur rur/sum cum corio eadem interior speculi superficies. Tandem
suboscurus, silicibus in aquis claris inventis haud dissimilis. Is itaque lapis sumatur tartarum rubeum, fuligo, et cinis salicis, et cum illis extrema fiat
intra mortarium aeneum pulverisandus est, dein per cetasseum aut lineum politura: hoc enim modo, paratum erit praefatum speculum parabolicum.
pannum cribrandus et exprimendus. Et huiuscemodi pulvis aqua com­
miscendus, et commixtura ponenda super plumbum, et cum ipso plumbo Appendix I.
ita madefacto poliendum speculum. Sed in primis, cum grossiori utcunque
pulvere eiusdem lapidis emerilli poliatur: deinde cum subtiliori. Est et Adde quod si ex praeassumpto corpore parabolico (sic enim non inepte
alius lapis emerillus pochea nuncupatus, quo vulgares utuntur artifices, vocari potest) libera pars circa illius verticem auferatur, dein reliquae parti
potissimum aurifabri, ad idem utilis si tritus fuerit super lapidem. Item orbiculari typus de more paretur, et fundatur demum, atque poliatur in­
genus aliud pocheae, quod vulgo color nuncupatur, ad poliendum etiam terior huiuscemodi orbis superficies: fiet speculum annulare seu orbicu­
valet, cum ligno mundo ab omni sorde, aut cum lamina ex plumbo et stanno lare, ad truncatam superficiem parabolam (ut haec figura [Fig. 7.33] reprae­
conflata. Poterit et idem speculum eo modo poliri quo poliuntur gladii et sentat) efformatum, quod simili modo, sed non adeo vivaciter, ignem ad
enses ab illorum artificibus. propositam distantiam (si radiis obiiciatur solaribus) accendet.

Alia eiusdem speculi compositio. Appendix II.

Iuvat demum aliam eiusdem speculi materiam, fabricam, atque polituram Ex hac itaque metallica et fusili materia, et haud dissimili poliendi ra­
ostendere: ad aliorum quoque speculorum constructionem indifferenter tione, fabricari poterunt data quaevis alia specula, tam plana, quam gib­
adcommodam. Fiat igitur ex ligno quopiam solido tabella quadrangularis bosa, et excavata. De his ergo satis. Speculi parabolici finis.
 FINAEI DE SPECULO USTORIO 383
382 ARCHIMEDES IN THE MIDDLE AGES
latus pl est maior latere rn], atque [/7] eidem [kn] parallela; plus ergo distat
PROPOSITIO X. punctum i ab ipso puncto / quam punctum k ab ipso puncto n : et proinde
ex supradictis corollaria. lineafg propinquior est ipsi aci in puncto k quam in ipso puncto io (//), quod
Dato recto atque rectangulo cono, duas colligere lineas quae quanto erat ostendendum. Haud dissimiliter descripto sub kno alio quopiam cir­
longius producentur tanto propiores evadent, nunquam tamen (etiam si in culo, eidem kno parallelo, ostendimus eundem circulum secare lineam
infinitum producantur) convenient adinvicem. fg in puncto propiore ipsi lineae ad quam sit punctum k: et sic in infinitum.
23v / Dum supradictas sectionis parabolae eius coni qui rectus atque rectan- Quanto magis igitur ad et fg lineae in continuum producentur ad partes
gulus dicitur construeremus demonstrationes: succurrit nobis imaginatio quidem d et g, tanto propiores evadent: et nihilominus eas tandem con­
quaedam non praetermittenda, a non nullis olim tentata, de duabus vide­ venire est impossibile, utpote, quae in planis consistunt invicem parallelis,
licet lineis tam in eodem plano quam in diversis planis constitutis quae ex ipsa constructione, et semper tantum ad minus invicem distabunt,
quanto longius producentur tanto propiores fient adinvicem, nunquam quanta est linea recta utrique praedictarum superficierum perpendicularis.
tamen convenient, etiam si in infinitum producantur. Utraque igitur propositionis pars verissima relinquitur.
Esto igitur datus rectus atque rectangulus conus abc [Fig. 7.34], cuius Idem consequenter ostendetur ubi datae lineae sub eodem plano fuerint
vertex a , basis vero bdce circulus. Hunc itaque conum bifariam dividat constitutae. Intelligatur enim plana quaedam superficies asxd [Fig. 7.35]
triangulum rectangulum atque isosceles ade per ipsius coni verticem et super rectam ad constituta et cum ipso triangulo ade orthogonaliter erecta,
axem eductum: cuius latera sint ad et ae, basis vero recta de. Sit rursum in quam concurrat praeassumpta superficiesfgh ad partes lineaefg directe
alia quaedam superficies plana conum ipsum inaequaliter dividens, sub coextensa: sitque earundem superficierum communis et orthogonalis inter­
inflexa linea gfh et recta gh comprehensa et ipsi ade triangulo parallela: sectio recta sx. Aio lineas fg et sx, sub eodem plano constitutas, quanto
cuius vertex seu punctum ipsi a vertici propinquius sit / . Dico quod, si magis in directum producentur ad partes quidem g et x, tanto fieri pro­
lineae ad et fg , sub diversis planis in primis et invicem parallelis constitu­ piores: sed nusquam invicem convenire posse, etiam si in infinitum pro­
tae, quanto magis in continuum una cum ipso cono abc producentur, tanto ducantur. Per data enim puncta / et n ipsius ad in rectam sx rectae ducantur
propinquiores offendentur: et nihilominus easdem lineas invicem con­ lineae It et aim, ipsis ip et kr parallelae; et connectantur it et ku lineae rectae.
venire est impossibile. Suscipiantur enim in ipsa linea inflexafg duo puncta Parallelogramma erunt igitur ipsa//?/t et krnu quadrilatera atque rectangula,
i, k, per quae duo transeant circuli ipsi basi bdce atque invicem paralleli, 25r per ipsam / planorum atque linearum constructionem. Omnis autem
quorum circunferentiae sint ilm et kno. Et comprehensis inter lineas ac (/ parallelogrammi latera, et anguli quae ex opposito, aequalia sunt adin­
ad) et fg eorundem circulorum arcubus il et kn; aequales eisdem sint Im et vicem, per trigesimam quartam primi elementorum. Aequalis est igitur it
no, una cum subtendentibus chordis im atque lo (/ ko), quae de necessitate ipsi pl, et ku ipsi rn. Atqui/?/ eadem rn maior praeostensa est: et it igitur ipsa
bifariam et ad rectos dividentur angulos a plana superficie praefati trianguli ku maior est. Propior est itaque linea fg ipsi rectae sx in puncto k quam in
ade.in punctis quidem/? etr, quarum sagittae in ipso plano constitutae sint puncto /. Haud aliter, si describatur sub kno alius quispiam circulus eidem
pl et At i . His constructis, aio fg lineam propinquiorem esse eidem ac (! kno parallelus, concludetur rursum eadem lineafg sub illius sectionem cum
ad) in puncto k quam in ipso puncto i. Connectantur enim il et kn lineae eodem circulo propior esse eidem sx quam sub puncto/:: et sic in infinitum.
rectae. Et quoniam superficies trianguli rectanguli ade transit per utriusque Quanto magis igitur praefatae lineae fg et sx ad easdem partes g et x una
circuli centrum et dividit arcus ilm et kno bifariam: dividit igitur et ipsas cum ipso cono producentur, tanto propiores erunt adinvicem. [Et quod
24r chor/das im et ko itidem bifariam in ipsis quidem punctis p et r. Dupla nunquam convenient etiam si in infinitum producantur patet.] Sola enim
est igitur im ipsius ip, et ko ipsius kr. Sed ip ipsi kr est aequalis: parallela linea ad ipsius plani asxd conum ipsum abc tangit, et productum conum
siquidem est/g/i superficies ipsi a d e, per constructionem. Aequalis est tanget in omnibus suis punctis; linea porro fg , ab eodem cono nusquam
igitur et im ipsi ko, per sextam communem sententiam geometricorum ele­ 25v dimovebitur. Recta / itaque linea sx nunquam tanget eundem conum abc in
mentorum. Circulus porro ilm minor est circulo kno: nempe vicinior aliquo sui puncto: neque igitur lineae fg . Et proinde ipsas fg et sx lineas
vertici a ipsius coni a b c. Minor est itaque arcus kno ipso arcu ilm : aequales datas, et in eodem plano constitutas, invicem convenire est impossibile.
enim chordae inaequalium circulorum inaequales secant arcus, minorem Quod tandem invenisse ac demonstrasse oportuit.
quidem a maiori et maiorem a minori, quoniam plus incurvatur minor quam FINIS.
ipse maior circulus. Et sagitta consequenter pl maior est sagitta rn : ut ex Virescit vulnere virtus.
ipsa seorsum ad dextram obiecta licet intueri figura. Triangula itaque ipl et
24v krn / habent duo latera//? (U l)etpl duobus lateribus kr {! kn) et rn inaequalia
et aequos nihilominus angulos comprehendentia, nempe rectos qui ad p et
r. Basis igitur il basi kn maior est [quoniam latus ip aequatur lateri kr, et
 EXTRACTS FROM BAROZZTS ADMIRANDUM PROBLEMA 385

B. Peletarius in secundo trium suorum Commentariorum De Dimensione

Circuli, De Contactu linearum, et De Constitutione Horoscopi.
E x tr a c ts fr o m F . B a r o z z i’s
Admirandum illud geometricum
[Extract II: Preface, pp. 5-11]
problema 5 / FRANCISCI BAROCII
Tredecim modis AD CAMILLUM PALAEOTUM
VIRUM CLARISSIMUM
Demonstratum Praefatio.
Quod docet duas lineas in eodem plano Nonnulla1 in Geometria Problemata, et Theoremata sunt (Camille vir
designare, quae nunquam invicem coin- praestantissime, ac eruditisime) quae cum admiratione digna sint, hoc sibi
nomen vendicarunt, ut a Geometris admiranda vocarentur: non profecto
cidant, etiam si in infinitum protrahantur: quia Geometris admiranda videantur (qui enim rerum causas sciunt, ef­
et quanto longius producuntur, tanto sibi- fectus admirari non possunt) sed quoniam vulgo geometricas eorum
causas ignoranti absurda, incredibiliaque apparent. Tale2quidem est illud
invicem proprior es evadunt. Problema: Super una parte lateris trianguli duas rectas lineas introrsum
V e n e t iis constituere duobus reliquis eiusdem trianguli lateribus maiores, et minorem
angulum quam ea latera continentes. Quomodo enim admirabile non est,
A p u d G r a tio s u m P e r c h a c in u m , s u m p tib u s si rectae quidem lineae super toto latere introrsum constitutae, externis
I o . B a p tis ta e F a n tin i P a ta v in i. minores, et maiorem angulum comprehendentes ab Euclide demonstratae
sunt: quae vero super parte ipsius lateris constituuntur, eisdem externis
M D L X X X V I. maiores, et minorem angulum comprehendentes sint? Huiuscemodi3 vero
illud etiam est Problema, quod ait: Triangulum quadrilaterum reperire
6 habens angulum externum tribus internis aequalem, tres au/tem internos
[Extract I: pp. 3-4]
duobus rectis minores. Nonne hoc etiam admiratu dignum erit, cum
3 / Autores de lineis nunquam coincidentibus, et semper sibi magis ap­ definiat Euclides triangulum esse trilateram figuram, demonstretque
propinquantibus in eodem plano in infinitum protractis. omnis trianguli angulum externum duobus internis ex opposito iacentibus
QUI REI MENTIONEM tantum fecerunt: PROCLUS in suis Com­ esse aequalem: necnon tres internos duobus rectis aequales esse? Ex
mentariis in primum librum Elementorum Euclidis multis in locis. GEMINUS admirandis4 etiam est Pythagoricum illud Theorema: Tria sola Multian-
in libro sexto suarum Geometricarum Enarrationum. GEORGIUS Valla gula totum, qui circa punctum unum est locum replere possunt, nempe
Placentinus in libro primo suae Geometriae Cap. LIX. RABBI MOYSES Triangulum aequilaterum, Quadratum, et Sexangulum aequilaterum
Aegyptius in primo libro Cap. 73 sui Operis inscripti Director dubitantium. simul, et aequiangulum. Si enim Triangulum, Quadratum, et Sexangulum
COELIUS Calcagninus in quadam sua Epistola. locum ipsum replent; cur Quinquangulum etiam, et Octangulum, caeteraque
QUI REM IMPERFECTE demonstarunt: APOLLONIUS Pergaeus in Multangula eum replere non possunt? Ex admirabilium5 numero illud
prima, et quartadecima Propositione secundi libri Conicorum. PAPPUS quoque Theorema est: Figurarum planarum rectilinearum quaedam habent
4 in suis Scholiis in quintum librum Conicorum Apollonii. / EUTOCIUS ambitus quidem, sive circuitus aequales, areas vero inaequales: et e con­
Ascalonita in suis Commentariis in secundum librum Archimedis de verso, areas quidem aequales, ambitus vero inaequales, et quae quidem
Sphaera, et Cylindro: et in librum secundum Conicorum Apollonii. IN­ minores habent ambitus, quandoque aequales quoad aream, quandoque
NOMINATUS Antiquus Autor in fine libelli de Sectione Conica, quae maiores sunt iis, quae maiores ambitus habent: quae vero quo ad aream
Parabole dicitur. ORONTIUS Finaeus in suo libello de Speculo Ustorio.
IOANNES Vernerus in vicesima Propositione suorum Elementorum 1 mg.: Admiranda in Geometria Problemata, et Theoremata quae sint, et cur admiranda
Conicorum. RABBI MOYSES Narbonensis in Opusculo, quod de hac re vocentur.
2 mg. : Admirandum Problema.
composuit. RABBI SAMTOU in expositione Cap. 73 libri primi operis
3 mg. : Aliud admirandum Problema.
inscripti Director dubitantium Rabbi Moysis Aegyptii. HIERONYMUS 4 mg.: Admirabile Pythagoricum Theorema.
Cardanus Mediolanensis in libro sextodecimo de Subtilitate. JACOBUS 5 mg.: Aliud admirabile Theorema.

384
386 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZFS AD M IRAN D U M PROBLEMA 387

minores sunt, si maioribus quoad aream comparentur, aliquando aequali, Philosophi, de quibus alibi sermo nobis etiam erit, ubi9 hoc Rabbi Moysis
aliquando maiori ambitu fruuntur. Talia sunt ea Problemata, et Theore­ dictum eiusque verba plenius exposituri sumus. Verum10 enim vero quum
mata Geometrica, quae admirabilia vocantur. Ex omnibus6 autem ad­ celeberrimam, ac mirabilissimam iam dictam propositionem in medium
mirandis in Geometria propositionibus una est caeteras admiratione, mihi attulisses, eiusque demonstrationem a me petiisses: cupiens ego
stuporeque superans, quippe quae demonstrat duas in eodem plano posse desiderio tuo pro viribus satisfacere, quoddam onus non leve suscepi, a
describi lineas, quae nunquam adinvicem coincidant, etiam si in infinitum quo tandem Dei Opt. Max. ope expeditum me video. Opus enim de hac
protrahantur: et quanto longius producuntur, tanto sibiinvicem propiores re integrum composui, in quo tredecim modis praefatam propositionem
7 evadant. Unde7 Rabbi Moyses Ae/gyptius primo lib. cap. 73 sui divini problematice demonstravi, et quicquid ab omnibus tum antiquis, tum re-
operis inscripti Director dubitantium, volens ostendere quod imaginatio centioribus, quos vidi, Autoribus de hac re dicta fuere, in unum collegi:
non sit mentis operatio, sed a mente differat, hac usus est ratione. Quoniam et ea quidem, quae ab eis vel imperfecte, vel male demonstrata fuerant,
scilicet quaedam mente percipiuntur, quae imaginatio non capit, quam 9 ad perfectionem, et exquisitam demonstrationem redegi: ea vero, quae / ab
utique rationem ex hoc confirmat, quod iam dicta omnium admira­ eis falso dicta sunt, rationibus confutavi, atque destruxi, ut rei veritas
bilissima propositio mente quidem percipitur, sub imaginationem vero ab omni contradictione, controversiaque immunis redderetur. Talem au­
non cadit. Quod sane Rabbi Moysis dictum si ita intelligatur ut a multis tem in hoc opere ordinem servavi. Primum quidem more Mathematicorum
exponitur, falsum nimirum esse videtur, ideoque a nonnullis tanquam principia quaedam posui, atque declaravi, quae una cum Euclidis Ele­
falsum refellitur. Nil enim a mente percipitur, quod etiam ab imagina­ mentis confirmant quidquid in toto opere a me dictum est. Post principia11
tione non capiatur, quamvis diversis modis, nempe a mente quidem in- vero tres propositiones demonstravi, quae tanquam totius operis Elementa
telligenter, ab imaginatione vero imaginanter. Mens nanque cuncta sim­ sunt. Deinde undecim diversas institutae propositionis demonstrationes
pliciter, et indivisibiliter rationibus intelligit, imaginatio autem composite, posui: quarum etiam Elementa, et Sumptiones ante eas semper demon­
et partim divisibiliter, partim indivisibiliter formis in phantasia impressis strationibus confirmavi; Corollariaque necessaria ex eis excerpsi. Post
res sibi subiectas imaginatur, atque cognoscit. Cum enim phantasia inter undecim autem demonstrationes, errores insigniores, et imperfectiones
mentem et sensum media sit, ut docent Aristotelici, ac Platonici: ima­ Autorum de hac re tractantium declaravi: falsasque eorum opiniones
ginatio etiam, quae circa Phantasiam versatur, inter mentem sensumque redargui. Postremo denique libellum Rabbi Moysis Narbonensis de hac
media erit. Cum autem nil sit in mente, quod prius non fuerit in sensu, re compositum dilucidavi, in quo dilucidando reliquas etiam duas eiusdem
ut Aristoteles docet: necessario quicquid mens percipit, imaginatio etiam propositionis demonstrationes illustravi, atque perfeci: dictum quoque
capit. Non datur si quidem ab extremo ad extremum, nisi per medium Rabbi Moysis Aegyptii diligenter exposui, ac demum12 Divino auxilio
transitus. Aliter8 igitur dictum Rabbi Moysis intelligendum esse existimo. finem operi dedi. Cuncta vero haec a me quamvis non exiguo labore,
Quod scilicet quaedam mente quidem percipiantur, quae imaginatio non libenter tamen peracta sunt cum ut integre tibi satisfacerem, tum ut studio­
capiat. Hoc est quibusdam rebus imaginatio non assentit, donec discur­ sis omnibus maxime prodessem. Omnes enim qui diligenter huic nostrae
rens cogitatio causas earum inveniat, mensque eas tanquam evidentes lucubrationi operam navarint non solum admirandam illam proposi­
8 percipiat. / Quod enim duae lineae in eodem plano designatae, si in tionem perfecte, diversisque modis demonstrare scient: verum etiam ita
infinitum producantur, nunquam invicem coincidant, semperque magis, in rebus Conicis exercebuntur, atque instruentur, ut omnes libros de Co­
ac magis sibiinvicem appropinquent; sensus primum videt in rebus ma­ nicis scriptos, praesertimque Apollonii Elementa perfecte intelligere po-
terialibus, et sensi[bi]libus et imperfectis: deinde imaginatio perfectiori 10 terunt: nec non a multis errolribus, in quibus Autores elapsi sunt, sese
quodam modo in Phantasia id imaginatur, in rebus imaginabilibus, et a abstinebunt. Quam13utilis autem doctrina de Conicis sit, multi gravissimi
materia sensibili separatis, et perfectioribus: postea vero cogitatio dis­ viri attestantur, quippe qui in ea tradenda maxime insudarunt: ut Conon,
currens reperit huiusce rei causas, quibus rem ita se habere demonstrat: Apollonius, Serenus, Archimedes, et alii. Ex Conicorum enim doctrina
Postremo demum Mens ipsa rem iam demonstratam veluti veram percipit, multa humano usui emolumenta proveniunt. Diversa nanque Specula tum
tuncque imaginatio, et sensus menti consentientes conquiescunt. Ecce Conica, tum etiam Columnaria ea Perspectivae scientiae pars, quae Specu­
igitur quod mentis operatio ab imaginationis operatione differt, ut pulcre laria dicitur, construere docet, quae porro mirabiles effectus nobis sup-
concludit ipse Rabbi, si recte verba eius intelligantur. Sunt autem aliae
quoque mentis ab imaginatione pulcherrimae discrepantiae, ut docent 9 mg. : Vide hoc in fine operis.
10 mg. : Propositum, et Subiectum operis.
6 mg.: Propositio admirabilissima omnium Geometricarum Propositionum. 11 mg. : Ordo.
7 mg.: Rabbi Moysis Aegyptii dictum. 12 mg. : Finis.
8 mg. : Quomodo dictum Rabbi Moysis intelligendum sit. 13 mg. : Utilitas.
388 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZI’S AD M IRAN D U M PROBLEMA 389

peditant. Fieri autem non potest ut dicta Specula recte construantur ab Quum5 enim ratio rectae lineae D ad latus quadrati B sit sicut ratio
eo, qui Conicorum Elementorum ignarus existit. Nam Speculum illud om­ lateris quadrati C ad rectam E, igitur per primam partem vicesimae se­
nium Speculorum alioqui utilissimum, quod per reflexionem radiorum cundae propositionis sexti lib. Element. Eucl. eadem etiam erit ratio quad­
Solis magna etiam, et durissima corpora comburit, quo Archimedes rati lineae D ad quadratum B quae est quadrati C ad quadratum lineae
quoque in Syracusis naves comburebat, nonne ex Conica illa Sectione E. At quadratum ipsius D per Constructionem est aequale Rectangulo
fit, quae Parabole appellatur? Praeterea centra gravitatum inveniri non A. Ergo per primam partem septimae propositionis quinti libri Element.
possunt sine Conicarum Sectionum adminiculo, ut patet ex libris Archi­ Eucl. quam habet rationem quadratum ipsius D ad quadratum B eandem
medis de aeque ponderantibus, seu centris gravium planorum, et ex libro habebit Rectangulum A ad idem quadratum B. Quare per undecimam
Trepl οχονμένων, quem nonnulli inscribunt de insidentibus aquae, alii vero, propositionem eiusdem quinti parallelogrammum rectangulum A eandem
de iis quae vehuntur in aqua. At centri gravium cognitio, nonne admodum habet rationem ad quadratum B quam habet quadratum C ad quadratum
necessaria est ad multas Machinas cum in bello, tum in pace utilissimas ipsius E rectae lineae. Datis6 igitur parallelogrammo rectangulo A et duo­
extruendas? Rursus Perspectivae Scientiae, quae adeo utilis est, omnia bus quadratis B, et C, repertum est tertium quadratum ipsius E, ad quod
fundamenta ex Conicis dependent, quandoquidem omnis visio per Conum eam habet rationem alterum datorum quadratorum, nempe C, quam habet
fit. Quinetiam multas alias utilitates praebet Conicorum doctrina Astro­ parallelogrammum rectangulum A ad reliquum B quadratum, quod erat
logiae, Mechanicae, et Architecturae, quas in praesentia, ne taedio te faciundum (/ faciendum).
afficiam, silentio pertranseo. Utilissima itaque Conicorum doctrina est,
ll ad quam / opus hoc nostrum breviter instituit, dum propositionem praecipue / Propositio Secunda, Theorema primum.
nobis institutam variis modis demonstrat. Accipe14 igitur Camille nobilis­
sime, atque doctissime hunc nostrum laborem, qui sub tui nominis, toti- Si1 trianguli per Axem Coni alterum latus versus Coni Verticem in
usque Academiae nostrae foelicissimis auspiciis in lucem prodiens, non directum producatur, et ab eius extremitate extra Conum existente ad
parva cum autoritate in manibus hominum versabitur. Bononiae Kalendis conicae Basis Dimetientem recta linea ducatur secans reliquum trianguli
Ianuarii Anno Salutis M. D. LXVI. latus, atque in eadem recta linea quotcunque signa intra Conum suscipi­
antur, ab eisque rectae lineae plano ipsius per coni Axem trianguli ad
rectos angulos erigantur conicae occurrentes superficiei: erit ratio quadrati
[Extract III: pp. 49-91] uniuscuiusque ipsarum ad rectos angulos erectarum ad rectangulum con­
49 / ELEMENTA tentum a tota sua conterminali versus Coni Verticem extensa et parte ipsius
Propositio prima, Problema primum. conterminalis intra Conum existente sicut ratio quadrati cuiuslibet aliarum
Dato1 Parallelogrammo Rectangulo et duobus quadratis: invenire ter­ ad rectos angulos erectarum ad Rectangulum a tota similiter sua conter­
tium quadratum ad quod eam habeat Rationem alterum datorum quad­ minali et eius parte intra Conum existente comprehensum.
ratorum quam habet datum Rectangulum ad reliquum ipsorum quad­ Sit2 ABC triangulum per Axem Coni, et conicae Basis dimetiens BC
ratorum. [Fig. 7.48]. Et ipsius trianguli latus AB in partem A quantumlibet pro­
Sit2 datum quidem parallelogrammum rectangulum A , data vero duo ducatur per secundam pet. primi libri element. Euc. usque adD signum, a
quadrata B et C, volo invenire tertium quadratum ad quod eam habeat quo ad quodvis signum E in conicae BC basis dimetiente sumptum per
rationem alterum quadratorum B et C, verbi gratia ipsum C, quam habet primam petitionem eiusdem recta DE ducatur linea secans necessario AC
rectangulum A ad reliquum quadratum B [Fig. 7.47]. Reperiatur3 iiaque reliquum eiusdem trianguli latus in signo/7, per 32 prop, et 9 com. sent, et 5
per ultimam propositionem secundi libri Elementorum Euclidis latus po­ pet. primi libri Elementorum Eucl. et per 7 com. sent, huius. Si recta
tens aream A, quod sit recta linea D; et per duodecimam propositionem scilicet DC linea ducta intelligatur, atque in EF recta linea quotlibet utcun­
libri sexti eorundem Elementorum inveniatur recta linea E, ad quam latus que assumantur signa ut G, H, a quibus plano trianguli ABC ad rectos an­
50 quadrati C habeat eandem rationem quam habet recta D ad latus qua/drati gulos rectae lineae per 12 prop. lib. xi. Elementorum Euclidis erigantur oc­
B. Dico4 quadratum ipsius E illud esse quod quaeritur. currentes conicae superficiei in /, K signis. Dico3 quod ratio quadrati rectae

14 mg.: Dicatio. 5 mg.\ Demonstratio.

6 mg. : Conclusio.
1 mg. : Propositio.
2 mg. : Expositio. 1 mg.: Propositio.
:i mg. \ Constructio. 2 mg. : Expositio.
4 mg.\ Determinatio. 3 mg.: Determinatio.
390 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZTS AD M IRANDUM PROBLEMA 391

52 lineae GI ad re/ctangulum a DG, GF comprehensum est sicut ratio quad­ quartae propositionis quinti lib. Elementorum Euc. ratio quadrati lineae
rati lineae HK ad rectangulum quodam (/ quod) DH, HF continetur. Intel- GI ad Rectangulum a DG, GF comprehensum est sicut ratio quadrati
ligantur4 itaque duo plana conicae Basi parallela secantia Conum rectamque lineae HK ad Rectangulum a DH, HF contentum. Quod est Propositum.
DE lineam in signis G ,H , quorum utique planorum, et plani trianguli ABC Si8 igitur trianguli per Axem Coni alterum latus versus Coni Verticem in
communes sectiones erunt per 3 prop, xi lib. Elem. Euc. rectae lineae quae directum producatur, et reliqua, ut in Propositione. Quod demonstrasse
sint LGM et NHO. Communes5 autem eorundem planorum et superficiei oportuit.
conicae sectiones erunt per 4 Petitionem huius circunferentiae circulorum
LIM et NKO, quorum Dimetientes sunt ipsae LGM et NHO, ipsae vero Corollarium.
GI et H K rectae lineae per 2 prop. lib. xi. Elem. Eu. in eisdem cum lineis Hinc sit perspicuum quod recta linea HK maior est quam GI.
LGM et NHO sunt Planis. Ipsae demum LM et NO rectae lineae ipsi BC Nam per Constructionem et 29 prop, primi et 4 prop, sexti et 9 Com. Sen.
parallelae sunt per 16 prop, eiusdem lib. xi; unde etiam inter se parallelae 54 primi lib. Elem. Eucl. et 9 Com. Sent, huius HN ma/ior est quam GL, et
sunt per 30 propositionem primi libri eorundem. His6 ita constructis, si HO maior quam GM, ergo per 3 Com. Sent, huius Rectangulum contentum
rectae lineae LI, IM, et N K , KO ductae intelligantur, quoniam anguli ab NH , HO maius est Rectangulo ab LG, GM contento. At quadratum
quidemL/M et NKO per 31 prop, libri 3 Elem. Euc. recti sunt, rectae7vero quidem ipsius HK aequale ostensum est Rectangulo ab NH, HO compre­
lineae GI et HK per Constructionem et per 3 definitionem lib. 11 eorundem henso, quadratum vero ipsius GI aequale itidem Rectangulo ab LG, GM
ad rectos sunt angulos ipsis LM et NO rectis lineis: erunt per Corollarium contento. Igitur per primam, et secundam partem 7 propositionis lib. quinti
octavae propositionis libri sexti Elem. Eu. ipsae GI et HK mediae propor­ Elem. Eucl. et 9 Com. Sent, huius bis sumptam quadratum ipsius HK
tionales inter LG, GM, et NH, HO rectas lineas. Quare per primam partem maius est quadrato ipsius GI. Quare per 2 Com. Sent, huius recta linea
53 17 propositionis eiusdem sexti Elem. erit Rectangulum ab LG, GM / com­ HK maior est quam GI, et hoc est quod a Corollario proponitur. Eodem
prehensum aequale quadrato ipsius GI: et Rectangulum ab NH , HO con­ autem modo in aliis etiam omnibus huiuscemodi rectis lineis liquebit, quod
tentum aequale quadrato ipsius HK. Quoniam autem per prop. 23 eiusdem semper Basi conicae propinquiores ab eadem Basi remotioribus maiores sunt.
sexti aequiangula parallelogramma eam habent inter se rationem quae ex
laterum suorum rationibus componitur, omnia vero parallelogramma
Propositio tertia, Theorema secundum.
rectangula per 4 petitionem primi lib. Elem. Euc. aequiangula etiam sunt:
igitur ratio rectanguli aDG , GF contenti ad rectangulum ab LG, GM con­ Si1duo parallelogramma rectangula duobus quadratis ita adiungantur; ut
tentum componitur ex duabus rationibus, quarum una est lateris DG ad unum quidem cum ipsis commune latus habeant, duo vero duobus in direc­
latus GL, altera ipsius FG ad GM. Similiter ratio Rectanguli a DH, HF tum iacentia, atque duo haec aggregata invicem aequalia fuerint, si unum
comprehensi ad Rectangulum ab NH, HO comprehensum componitur ex unius rectanguli latus ex in directum iacentibus uni ex eisdem alterius
ratione lateris DH ad latus HN et ratione ipsius FH ad HO. At per Con­ rectanguli lateribus aequale fuerit: quadrata illa aequalia invicem erunt. Si
structionem et 2 partem prop. 29 primi et quartam propositionem sexti libri autem dicta latera inaequalia fuerint: quadratum cuius lateri maius rectan­
Elem. Euc. ratio ipsius DG ad GL eadem est quae ipsius DH ad H N, et guli latus in directum iacet reliquo quadrato minus erit.
ratio ipsius FG ad GM eadem quae ipsius FH ad H O . Igitur ratio composita Sint2 duo Parallelogramma rectangula ABCD et EFGH [Fig. 7.49], quae
ex rationibus laterum DG ad GL et FG ad GM, et ratio composita ex ra­ adiungantur quadratis BCIK et FGLM ita ut recta quidem linea BC sit com­
tionibus ipsorum DH ad HN et FH ad HO eaedem sunt per primam com. mune latus rectanguli ABCD et quadrati BCIK; et similiter ipsaFG sit com­
sent, huius. Qua propter per undecimam propositionem quinti libri Ele­ mune latus rectanguli EFGH et quadrati FGLM: duo vero AB et DC latera
mentorum Euclidis quater et secundam partem 7 propositionis eiusdem bis in directum iaceant ipsis BK et C/ lateribus, et pari modo duo EF et HG
sumptas ratio Rectanguli a DG, GF contenti ad Rectangulum ab LG, GM 55 ipsis FM et GL. Atque duo haec aggregata, nem/pe rectangula AKID et
comprehensum, seu ad ipsi aequale quadratum lineae GI, est sicut ratio EMLH invicem aequalia sint. Dico3 quod si AB recta linea rectae EF
Rectanguli a DH, HF comprehensi ad Rectangulum ab NH , HO con­ aequalis est, quadratum etiam BCIK aequale est quadrato FGLM. Si vero
tentum, seu ad ipsi aequale quadratum lineae HK. Ergo per Corollarium
8 mg. : Conclusio.
4 mg. : Constructio.
5 mg.: Hoc obscure elicitur ex Vemero. 1 mg.: Propositio.
6 mg.: Demonstratio. 2 mg. : Expositio.
7 mg.: Hoc etiam obscure ex Verneri verbis elicitur. 3 mg.: Determinatio.
392 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZFS AD M IRAN DUM PROBLEMA 393

AB maior fuerit quam EE, quadratum4BCIK quadrato FGLM minus erit. 57 / PROBLEMATIS PRAECIPUI DEMONSTRATIO PRIMA
Sint4 primum AB et EF aequales [Fig. 7.49a]. Si itaque BCIK quadratum Duas1in eodem plano designare lineas alteram rectam, et alteram curvam,
FGLM quadrato aequale non fuerit, aut minus ipso aut maius esse necesse quae nunquam adinvicem coincidant, etiam si in infinitum protrahantur: et
est. Quod si minus: ergo et BK latus ipso FM latere, e t/X ipso FM, minus quanto longius producuntur, tanto sibiinvicem propiores evadant.
erit per secundam Com. Sent, huius. Quare per 4 Com. Sent, primi lib. Sit2 Conus ABCD [Fig. 7.50], cuius Vertex A, Basis BDC circulus, tri­
Elemen. Eucl. tota A K minor erit quam tota EM, est autem et IK minor angulum per Coni Axem ABC, cuius latus AB producatur per secundam
quam LM, igitur per 3 Com. Sent, huius Rectangulum AKID Rectangulo petitionem primi lib. Elemen. Eucl. in partem A quoadlibet, usque in sig­
EMLH minus est, quod est suppositioni contrarium.5 Si vero quadratum num E, et in trianguli ABC BasiBC accipiatur quodcunque signum F, a quo
BCIK quadrato FGLM maius esse dicatur, eisdem rationibus concludetur ad signum E per primam pet. eiusdem primi recta ducatur linea FE secans
AKID Rectangulum ipso EMLH Rectangulo maius esse, quod etiam sup­ necessario ratione superius dicta latus AC eiusdem trianguli in signo G,
positioni oppugnat. Non6 est igitur quadratum BCIK quadrato FGLM 58 quae per 2 prop. lib. 11 Element. Eucl. erit in plano trianguli / ABC, et a
minus; neque maius, ergo ipsi aequale, quae est prima propositionis pars. signo G per 12 prop, eiusdem 11 lib. erigatur recta GH ad angulos rectos
Sint7 modo AB et EF inaequales, AB scilicet maior quam EF, ut in plano trianguli ABC, quae per 3 definitionem eiusdem 11 ad rectos angulos
secunda figura [Fig. 7.49b]. Si igitur quadratum BCIK quadrato FGLM est rectae lineae EGF, et per 2 prop, eiusdem erit in uno plano cum recta
minus non fuerit, aut aequale ipsi aut ipso maius erit.8 Quod si aequale sit linea EGF, quod porro planum secat Conum praeter Verticem si usque ad
56 latus BK lateri FM, et latus IK lateri LM / aequalia erunt per iam dictam Coni Basim protractum esse intelligatur, et erectum est ad planum trianguli
secundam Com. Sent., ergo per eandem etiam 4 Com. Sent, tota AK maior ABC per 4 definitionem eiusdem 11, et communis quidem sectio horum
erit quam tota EM. Cum autem IK sit aequalis ipsi LM, quae per 4 defini­ duorum planorum erit recta linea EF per constructionem et 3 prop, eiusdem
tionem sexti lib. Elem. Euc. sunt altitudines parallelogrammorum AKID libri 11: communis vero sectio plani EGH et conicae superficiei curva
et EMLH, erit per primam prop, eiusdem sexti ratio parallelogrammi quaedam est linea Hyperbolica seu latus Hyperboles per 5 pet. et 21 et 24
AKID ad parallelogrammum EMLH sicut ratio basis A K ad basim EM. definit, huius, et sit GID, quam utique tangit recta GH in signo G per con­
Atqui basis AK maior ostensa est basi EM. Igitur per 9 Com. Sent, huius et structionem et per secundam partem 28 propositionis primi lib. Elem. Eucl.
parallelogrammum AKID parallelogrammo EMLH maius erit, quod est et per primam partem 32 propositionis primi libri Conicorum Apollonii.
contra suppositionem. Si demum BCIK quadratum maius FGLM quadrato Subinde suscipiatur in ipsa GID curva linea quodcunque signum E , et ab
quis esse dixerit, per eandem secundam Com. Sent. BK erit maior quam ipso per 11 propositionem lib. 11 Elem. Eucl. ducatur in trianguli ABC
FM , et IK maior quam LM. Unde per quintam Com. Sent, huius tota AK 59 planum perpendicularis recta linea KL, quae per / 38 prop, eiusdem lib. 11
maior erit quam tota EM. Ergo per 3 Com. Sent, huius Rectangulum cadit inEE communem sectionem duorum iam dictorum planorum, ipsique
AKID Rectangulo EMLH maius erit, quod iterum suppositioni adversatur. EF ad rectos angulos est per tertiam definitionem eiusdem 11. Posthaec
Existente9 igitur AB linea maiori quam EF, neque maius est quadratum linea recta EG per 10 prop, primi lib. Elem. Euc. secetur in duas partes
BCIK quadrato FGLM, neque ipsi aequale, sed minus. Atqui ostensum est aequales in signo M, et per praedemonstratum problema fiat GH recta
etiam quod AB et EF lineis invicem aequalibus existentibus BCIK quadra­ linea tantae longitudinis ut ad eius quadratum eam habeat rationem quad­
tum de necessitate FGLM quadrato aequale est. Patet ergo utraque Theo­ ratum rectae lineae GM quam habet rectangulum contentum ab E L, LG
rematis huius pars. Si10 itaque duo parallelogramma rectangula duobus rectis lineis ad quadratum ipsius LK. Quo demum facto, ducatur MH recta
quadratis ita adiungantur, et reliqua ut supra. Quod oportebat demonstare. linea per primam pet. eiusdem primi, quae erit in eodem plano EGH in quo
Verum demonstratis iam tribus sequentium omnium demonstrationum est etiam GID curva linea per 2 prop. lib. xi. Elem. Euc. His hoc modo
Elementis, age modo ad institutam nobis Problematicam admirandamque constructis dico3 quod si duae lineae, nempe recta MH et curva GID in
Propositionem accedamus.
eodem EGH plano existentes, in infinitum protrahantur (intelligendo
scilicet planum EGH ex parte GH et Conum ex parte BCD basis in infinitum
4 mg.: Demonstratio primae partis. produci) nunquam adinvicem coincident: et quanto longius producuntur,
° m8·'· Duos casus habet haec prima pars, quos vide infra in digressione contra Vernerum. tanto sibi invicem propiores evadent. Coincidant4 autem, si id fieri potest,
e mg.: Conclusio primae partis.
7 mg.: Demonstratio secundae partis. 1 mg.: Propositio.
8 mg. : Tres habet casus secunda haec pars, quos vide infra in digressione contra Vernerum. 2 mg. : Constructio.
9 mg.: Conclusio Secundae partis. 3 mg.: Determinatio.
10 mg.: Conclusio totius. 4 mg.: Demonstratio primae partis.
394 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZFS AD M IRAN DUM PROBLEMA 395

in aliquo signo, verbi gratia in signo/, a quo per xi. prop. lib. xi. Elem. Euc. cidat in signo O cum recta linea MH in longum producta [Fig. 7.51]. Neces­
in trianguli ABC planum perpendicularis ducatur IN recta linea, quae per sario siquidem coincident per quintam pet. primi lib. eorundem, quoniam
38 prop, et 3 definitionem eiusdem cadit perpendiculariter in EGF com­ 62 angulus quidem MLK ex constructione rectus /' est, angulus vero LMH per
munem duorum planorum sectionem, et per 6 prop, eiusdem xi parallela 32 prop, eiusdem primi minor est recto. Deinde in ipsa GID curva linea
est ipsis GH et KL rectis lineis. Quare per primum Theorema superius infra signium K suscipiatur quoddam signum P, a quo super EGF rectam
demonstratum ratio rectanguli ab EN, NG rectis lineis comprehensi ad lineam per 12 prop, primi lib. Elem. Euc. perpendicularis PQ ducatur, quae
quadratum rectae IN est sicut ratio rectanguli ab EL, LG contenti ad in partes P producta occurrat ratione iam dicta in puncto R ipsi MH pro­
quadratum rectae lineae LK. Verum quae est ratio rectanguli ab EL, LG ductae. Postea vero quoniam per constructionem et 32 prop, primi lib.
comprehensi ad quadratum ipsius LK eadem per constructionem posita est Elem. Eucl. GH et KO et PR rectae lineae ad rectam MR perpendiculares
ratio quadrati rectae MG ad quadratum rectae GH. Ergo per xi. prop, libri non sunt, a punctis G ,K ,P ipsius inflexae lineae ad rectam lineam MR per
quinti Elem. Euc. eandem habet rationem rectangulum ab EN, NG con­ 12 prop, eiusdem primi ducantur perpendiculares G S,K T, PV rectae lineae,
tentum ad ipsius NI quadratum quam habet quadratum quod fit a linea MG quae quidem per 19 prop, eiusdem erunt minimae distantiae quibus puncta
ad quadratum quod a recta GH describitur. At ipsius MG ad ipsius GH G ,K ,P distant a recta linea MR. Aio7 itaque GS distantiam esse maiorem
quadratum eandem habet rationem quam habet etiam quadratum ipsius 63 KT di/stantia, et similiter ipsam KT ipsa PV\ nec non si infinitae eiusmodi
MN ad quadratum ipsius NI (in triangulo enim iuxta suppositionem recti- distantiae ducantur, minores continue fieri eas quae Basi Coni proximiores
lineo MNI recta linea NI rectae GH parallela posita est. Unde per 2 partem sunt, versus quam fit duarum non coincidentium linearum productio.
29 propositionis primi Elem. Euc. duo triangula MGH et MNI aequiangula Volentibus igitur nobis ostendere lineam GS linea KT, et ipsam KT ipsa PV,
sunt, et ideo per 4 prop, sexti eorundem habent latera proportionalia, maiorem esse, prius ostendendum est ipsam GH ipsaFO, et ipsam KO ipsa
60 videlicet MG ad GH sicut MN ad NI. Quamobrem per primam / partem 22 PR, esse maiorem. Quod8 itaque GH maior fit quam KO patet si per 31
propositionis eiusdem sexti quadrata etiam harum quatuor linearum sunt prop, primi elem. ducatur per signum K parallela ipsi MR secans lineam
proportionalia). Igitur per xi. proposi. libri quinti Elem. Eucl. rectangulum GH in puncto X (secabit enim eam per 5 pet. eiusdem primi si recta linea
ab EN, NG rectis lineis contentum ad quadratum lineae NI eandem habet GA ducta esse intelligatur, cum angulus quidem LGX rectus sit, angulus
rationem quam habet quadratum lineae MN ad eiusdem NI lineae quadra­ vero LKX aequalis per 2 partem 29 propositionis primi lib. Elem. Euc.
tum. Ergo per primam partem 9 propositionis eiusdem quinti rectangulum ipsi H OK, et ideo minor recto) nam per 34 proposi. eiusdem H X aequalis
ab E N , NG comprehensum aequale est quadrato rectae lineae M N . Verum- est ipsi KO , unde tota GH eadem KO maior est per 9 com. sent, primi Ele­
tamen cum in constructione recta EG in duas partes aequales in signo M mento. Euc. et 7 com. sent, huius. Quod vero KO maior sit quam PR, sic
divisa sit, eique in rectum adiiciatur recta GN, procul dubio per sextam liquebit. Ratio9 rectanguli ab E L ,L G contenti ad quadratum ipsius KL est
prop, secundi lib. Elem. Euc., rectangulum ab EN, NG comprehensum sicut ratio quadrati ipsius GM ad quadratum ipsius GH per construc­
superabitur a quadrato lineae MN, quadrato ipsius MG rectae lineae. Sed tionem, atque propterea sicut etiam ratio quadrati ipsius LM ad quadratum
hoc rectangulum eidem quadrato aequale etiam iam ostensum fuit, quod ipsius LO per propositiones vicesimamnonam primi et 4 et 22 sexti et xi
est maxime absurdum. Nam fieri non potest ut eaedem quantitates invicem quinti lib. Elem. Euc. Erit igitur per 19 prop, eiusdem quinti ratio quadrati
aequales simul et inaequales sint. Hoc equidem inconveniens sequutum est ipsius GM (quod per prop. 6. lib. 2. eorundem est excessus quo quadratum
quoniam suppositum fuit lineam rectam MH et inflexam GID in eodem ipsius LM superat rectangulum ab EL, LG contentum) ad ipsius OK quadra­
plano protractas coincidisse adinvicem in ipso I signo. Similiter autem tum, et duplum eius quod a KO, KL continetur (quod quidem totum est dif­
61 / idem sequetur incommodum si etiam in quocunque alio signo ipsae duae ferentia qua ipsius KL quadratum ab ipsius LO quadrato exceditur per 4
lineae adinvicem coincidere ponantur. In nullo igitur signo coincident, prop, eiusdem secundi) sicut ratio quadrati lineae ML ad quadratum ipsius
etiam si in infinitum protractae fuerint. Patet5 itaque prima quaesiti nostri LO, hoc est quadrati lineae GM ad quadratum lineae GH , eaedem enim
pars, descriptae nanque sunt in eodem EGH plano duae lineae recta MH et istae duae rationes sunt, ut ostensum est. Igitur per 2 partem prop. 9. lib. v.
inflexa GID nunquam adinvicem coincidentes quantumcunque protrahantur. Elemen. Euc. quadratum ipsius GH aequale est quadrato ipsius KO , et
Praeterea demonstrandum est quod quanto longius producuntur, tanto duplo eius quod ab OK, KL continetur, si quidem ad utrunque eorum
sibi invicem propiores fiant. Producatur6 itaque per secundam petitionem quadratum ipsius GM eandem rationem habet. Similiter quoque demon­
primi lib. Elem. Euc. recta linea LK in continuum et directum donec coin- strabitur quod quadratum ipsius GH aequale est quadrato ipsius PR et

7 mg. : Determinatio 2ai partis.

5 mg.: Conclusio primae partis. 8 mg.: Demonstratio 2a<' partis.
6 mg.: Secundae partis constructio. 9 mg.: In hoc Vernerus obscurus est.
396 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZTS AD M IRAN D U M PROBLEMA 397

duplo eius quod ab RP, PQ comprehenditur (nam rectangulum abEQ, QG Corollarium.

contentum ad quadratum PQ eandem habet rationem quam rectangulum ab Ex demonstratione secundae partis huius Problematis emergit nobis
EL, LG comprehensum ad quadratum KL per primum Theorema ante 66 Corollarium quod quotiescunque Hyberbolle cum recta linea ipsi non
demonstratum. Unde per constr. et per xi prop. lib. v. Elem. Euc. rectan- coincidente in eodem plano descriptae fuerint, quadratum ipsius KO, et
64 gulum ab EQ, QG ad quadratum / lineae PQ se habet ut quadratum ipsius duplum eius quod ab OK, KL continetur aequalia sunt quadrato ipsius PR
MG ad ipsius GH quadratum, et reliqua ut superius). Ergo per 1 com­ et doplo (! duplo) rectanguli ab RP, PQ comprehensi, idemque verum est
munem sent, primi eorundem Elem. quadratum ipsius KO una cum duplo in omnibus lineis ordinate ductis intra Hyperbolem, atque ad non coin­
rectanguli ab OK, KL contenti aequale est quadrato ipsius RP et duplo cidentem usque rectam lineam productis.
eius rectanguli quod ab RP, PQ comprehenditur. Si itaque duo rectangula
a PR, PQ comprehensa ita sibiinvicem in directum coniungantur ut per Secunda Eiusdem Problematis Demonstratio.
primam propositionem secundi libri Elem. Eucl. unum ex ipsis confectum
rectangulum aequale sit duplo rectanguli a PQ, PR contenti, necnon Sit1 Conus ABCD [Fig. 7.52], cuius Vertex quidem A, Basis vero BCD
quadratum lineae PR ipsi totali rectangulo sic adiungatur ut unum cum ipso circulus, et a puncto C per medium semicircunferentiamBCD dividente ad
commune latus habeat: idemque similiter de duobus rectangulis ab LK, Verticem A ducatur recta linea CA, quae tota erit in conica superficie per
KO contentis, et quadrato lineae KO fiat: quoniam per Corollarium primam petitionem huius. Et ab eodem puncto C per 12 prop. 11 lib. Elemen.
65 / primi praeostensi Theorematis PQ maior est quam K L , erit per secundam Eucl. recta erigatur linea CE ad rectos angulos plano trianguli per axem
partem secundi ante demonstrati Theorematis quadratum lineae KO maius habentis latus AC, quae quidem CE linea per 3 definit, eiusdem 11 erit ad
quadrato lineae PR. Per 2 igitur Com. Sent, huius recta linea KO maior est rectos angulos in puncto C tum dimetienti circuli BCD, tum ipsi AC rectae
quam ipsa PR. Verum10 enim vero quandoquidem iam demonstratum est lineae. Quare per Corollarium 16 prop. lib. tertii Elem. Eucl. eadem AC
rectam lineam GH recta KO, et ipsam KO ipsa PR, maiorem esse: reliquum recta linea erit tangens circulum BCD in puncto C. Duae igitur lineae CA et
est ut itidem rectam lineam GS recta KT, et ipsam KT recta PV, maiorem CE in eodem sunt plano per 2 prop. lib. 11 eorundem, quod quidem planum
ostendamus. Quoniam11 igitur anguli GSH, KTO, PVR per construc­ necessario tangit conicam superficiem a Vertice usque ad Basim recta linea
tionem recti sunt, ideoque per 4 Pet. primi lib. Elemen. Eucl. invicem AC. Ducatur itaque per punctum A recta linea AF parallela et aequalis
aequales; similiter autem anguli GHS, KOT, PRV per constr. et 2 par­ rectae CE per 31 et 3 prop, primi lib. Elemen. Eucl. Et ducatur per primam
tem 29 prop, primi eorundem Elemen. inter se aequales sunt: ergo per Pet. eiusdem EF, quae etiam per 33 prop, eiusdem primi rectae lineae
32 prop, et per 3 Com. Sent, eiusdem primi triangula GHS, KOT, PRV AC parallela et aequalis erit. Factum est ergo ACEF parallelogrammum
aequiangula sunt. Quare per 4 prop. 6. lib. eorundem quemadmodum se rectangulum planum tangens superficiem conicam in linea AC. Quod autem
habet GH ad K O , et KO ad P R , ita etiam GS ad KT, et KT ad PV. Sed GH hoc planum hincinde productum nullibi nisi in linea AC conicam super­
maior est quam KO , et KO quam PR, ut ostensum est. Igitur etiam GS ficiem tangere possit, facile convincitur. Si2enim tangeret ipsam in quodam
quam KT, et KT quam PV, per 9 Com. Sent, huius maior erit. Hae autem alio puncto, ut in C, intelligatur planum parallelum basi secans Conum per
sunt minimae distantiae quibus signa G, K , P in curva linea existentia signum G, cuius plani et conicae superficiei communis sectio erit circulus
distant a recta MR linea. Ergo signum P propius est rectae lineae MR quam per 4 pet. huius, quippequi circulus sit//G /secans lineam AC in signo H. Et
signum K , et signum K quam signum G. Et quoniam idem de quocunque 67 quoniam planum circuli HGI secat / planum ACEF in signis G, H ipsius con­
alio puncto in eadem obliqua linea suscepto eodem modo usque in infinitum tactus, erit per 3 prop. lib. 11 Elemen. Eucl. communis eorum sectio recta
demonstrari potest, perspicuum12 est quod quanto amplius recta linea linea GH, quae cum duo puncta connectat in circuli circunferentia existen­
MHR et inflexa seu Hyperbolica GID in eodem plano EGH producuntur, tia, seu in conica superficie praeter Coni Fastigium iacentia, cadat intra
eo magis sibiinvicem appropinquant. Atque hoc erat secundum Quaesiti 68 circulum ipsum atque intra Conum per 2 prop. 3 lib. Elem. Eucl. / seu per 2
membrum. Quapropter utraque propositi Problematis pars manifesta Pet. huius. Cum autem ipsa linea GH sit in plano ACEF, iuxta quam seca­
claraque habetur. Duas13 itaque in eodem plano designavimus lineas al­ tur a circuli plano, necessario planum etiam ACEF caderet intra Conum
teram rectam, et alteram inflexam, quae nunquam adinvicem coincidunt secaretque ipsum in linea AC per 15 Definitionem huius, quod est contra
etiam si in infinitum protrahantur: et quanto longius producuntur, tanto Constructionem, quandoquidem planum ACEF positum fuit tangens
sibiinvicem proximiores evadunt, quod faciundum (/) erat. 69 superficiem Coni in linea AC. / Non potest enim idem planum in eadem
recta linea eandem conicam superficiem et tangere simul et secare. Secans
10 mg.: Hoc obscure ex Verneri verbis haberi potest.
11 mg.: In hac parte deficit Vernerus.
12 mg.: Conclusio secundae partis. 1 mg. : Constructio.
13 mg.: Conclusio un[i]versalis. 2 mg.: Indirecta ostensio. In hoc probando deficit Cardanus, et Paralogismum committit.
398 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZLS AD M IRAN D U M PROBLEMA 399

siquidem intra Conum, tangens vero extra cadit, ut ex 15 et 16 Defini­ duae lineae non coincidentes quo magis a culmine Coni elongatur eo magis
tionibus huius patet. Praeterea planum ACEF non posse tangere conicam invicem proximae fiunt. Et satis sit in duobus tantum lineae inflexae punc­
superficiem alibi quam in AC linea, sic etiam directa demonstratione tis, utpote G et M hoc demonstrare, quod scilicet in puncto G proximior
habetur. In3 conica nanque superficie infiniti circuli excogitari possunt sit inflexa GM linea rectae EF quam in puncto M , quandoquidem eodem
rectam lineam AC secantes, a quibus sectionibus per 31 prop, primi lib. modo in omnibus etiam aliis ipsius curvae lineae punctis idem ostendetur.
Elemen. Eucl. in plano ACEF rectae lineae duci possunt parallelae ipsi Capiatur8 itaque circulus NMO parallelus basi BCD et priori circulo HGI.
CE, atque idcirco per 8 propo. 11 lib. Elem. Eucl. rectae erunt ad planum Et ducantur per primam Pet. primi lib. Elem. Eucl. in planis circulorum
dimetientium eorum circulorum, et per tertiam Definitionem eiusdem ad NMO et HGI iuxta communes eorum et plani EFKL sectiones rectae
rectos angulos ipsis dimetientibus, et per 16 prop. 3 lib. eorundem et eius lineae GP et M Q, et producantur in partes G et M per 2 pet. eiusdem quous­
corollarium tangent circulos ipsos in illis tantum punctis sectionum totae que secent lineam EF, ipsa quidem PG in signo R , ipsa vero QM in signo S .
extra circulos, extraque Conum cadentes. Unde manifestum est quod Et erunt per 16 prop. lib. 11 Elem. Eucl. ipsae PR et QS invicem parallelae.
planum ACEF, in quo sunt iam dictae lineae, nullibi nisi in linea AC coni­ Similiter per eandem primam Pet. ducantur in plano ACEF rectae lineae
cam superficiem tanget. Sit igitur ACEF planum tangens in AC tantum HR et N S , quae quidem erunt etiam in planis eorundem circulorum per 2
recta linea Coni superficiem, cuius plani latitudo A F , seu CE, tanta sit, ut si prop, eiusdem 11 Elementorum bis sumptam; cum lineis enim PR et QS
aliquod planum cuius communis sectio cum plano ACEF sit recta EF seinvicem secant.9 Igitur per 16 propositionem eiusdem 11 sibi invicem et
erigatur super planum ACEF, producaturque ex parte Coni, necessario ipsi CE parallelae erunt. Necnon circulos HGI et NMO tangent rationibus
secet Conum praeter Verticem. Sit itaque huiuscemodi planum productum superius dictis. At etiam H N et RS ex Constructione parallelae sunt. Ergo
EFKL secans conicam superficiem in signis G et M. Et quoniam per Con­ per 34 prop, primi lib. eorundem element. HR et NS invicem sunt aequales.
structionem et 4 Definitionem et 14 prop. 11 Elem. planum EFKL paral­ Et10quoniam contingunt circulos HGI, NM O: erit per 36 prop. lib. 3 eorun­
lelum est plano trianguli per axem Coni illius, scilicet cuius unum latus est dem quadratum ipsius NS aequale parallelogrammo rectangulo quod fit
AC , communes etiam duorum Ltorum planorum et plani trianguli ABD per ex multiplicatione ipsius QS in SM\ et quadratum ipsius HR aequale ei
axem Coni sectiones parallelae invicem erunt per 16 proposi. lib. 11 Elemen. quod fit exPR in RG. Quadrata autem ipsarum///? et TVS aequalia sunt per
Eucl. Quare cum altera istarum parallelarum sectionum per 10 et 18 Defini­ Constructionem et per 2 Com. Sent, huius. Rectangulum igitur quod fit ex
tionem huius sit axis Coni, reliqua nimirum sectio per 29 et 5 et 32 prop, PR in RG aequale est rectangulo quod iit ex QS in SM per primam Com.
primi lib. Elem. Eucl. exibit a minoribus duobus rectis angulis cum latere Senten. primi lib. Elemen. Eucl. bis sumptam. Quare per secundam partem
AB, ideoque per 5 Pet. eiusdem primi occurret ipsi AB lateri in partem A 16 propositionis lib. 6 eorundem Elementorum ratio rectae lineae PR ad
producto. Quamobrem per 5 Pet. et 21 Definitionem huius linea GM in rectam QS est sicut ratio lineae MS ad GR. Sed PR maior est quam QS.
superficie conica iacens inflexa, mista, et Hyperbolica linea est, seu latus Ergo MS etiam maior est quam GR per 9 Com. Sent, huius. Quod autem
Hyperboles per 24 Definitionem huius. Constatque ex Constructione 72 PR maior sit quam QS sic probetur. Sin in Plano EFKL per / signum P
ipsam GM curvam lineam in eodem esse plano EFKL cum recta EF. ducatur per 31 prop, primi lib. Elemen. Eucl. parallela ipsi EF, dubio
Dico4 itaque hasce duas lineas, inflexam scilicet MG et rectam EF, in procul11 extra signum Q cadet, quod sic deducatur. Recta linea AF per
70 eodem EFKL plano continue productas (quod planum / intelligatur in in­ Construet, et per 2 prop. lib. 11 eorundem Elemento, est tum in plano
finitum cum tota Coni superficie Basim versus extendi) nunquam sibi ACEF, tum in Plano trianguli ABD per Axem Coni. Si igitur Axis Hiper-
invicem occurrere. Si5 enim sibi occurrant, aut hoc fiet in linea AC, et ita 73 boles iam de/scriptae in eiusdem trianguli plano per 21 et 22 Definitionem
AC et EF parallelae coincident, quod est inconveniens: aut praeter AC huius existens producatur in directum versus Coni verticem, necesse est
lineam, et ita cum MG quidem inflexa linea semper sit in superficie Coni, ut rectam AF secet in aliquo puncto. Cum autem idem axis sit etiam per
71 recta vero EF / maneat semper in plano ACEF, necessario planum ACEF constructionem in plano EFKL, in nullo alio signo quam in ipso F rectam
tangeret Conum alibi quam in linea AC, quod fieri non posse iam demon­ AF lineam secabit. Producatur igitur axis ipse, qui per 22 Definitionem
stravimus. Nunquam ergo GM et EF lineae coincident etiam si in infinitum huius secet PG ordinate ductam per medium, et ad rectos angulos in signo
protrahantur. Et haec6 est prima Problematis pars. Dico7 modo quod istae T. Quoniam igitur trianguli FTR angulus qui ad T rectus est, erunt duo
reliqui eius anguli per 32 prop. lib. primi Elemen. Eucl. acuti. Angulus
3 mg.: Directa ostensio. In hoc probando deficit Peletarius, et quoddam falsum dicit.
4mg.: Determinatio primae partis. 8 mg.: Constructio secundae partis.
5mg.: Demonstratio primae partis. 9 mg. : In hoc deficit Cardanus, et falsum dicit.
6 mg.: Conclusio primae partis. '° mg.: Demonstratio secundae partis.
7 mg.: Determinatio secundae partis. 11 mg.: Hoc male probat Cardanus.
400 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZTS AD M IRAN DUM PROBLEMA 401

igitur FRG acutus est. Eadem ratione etiam angulus FSM acutus erit. Si erat secundo demonstrandum. Descriptae13sunt igitur in eodem plano duae
itaque protracta recta RP in partem P quantumlibet eius extremitas con- lineae altera recta, et altera inflexa, et reliqua ut in propositione. Quod
iungatur per rectam lineam cum signo F, ostendetur similiter angulum qui fecisse oportuit.
ex his duabus rectis lineis extra Conum fiet acutum esse. Quamobrem si in
plano EFKL per punctum P per 31 prop. lib. primi eorundem Elemen. Corollarium.
ducatur recta linea parallela ipsi RS, necesse est ipsam cadere extra Coni Hinc manifestum est quod si planum EFKL ex parte lineae EF pro­
superficiem, quia ipsi a puncto F extra Conum ultimo per imaginationem ducatur, in ipsoque per 31 propositionem primi lib. Elemen. Eucl. una
ductae lineae occurrere debet per 5 pet. primi lib. Elem. Eucl. quando­ recta linea parallela ipsi EF ducatur, quae duae parallelae rectae lineae
quidem in signo P cum parte protracta ipsius RP acutum facit angulum per distent ab invicem quodam determinato spatio, gratia exempli, mille
secundam partem 29 prop, eiusdem primi lib. Elem. Cum itaque recta linea Stadiis: erunt designatae in eodem plano duae lineae altera recta, ultimo
quae per signum P ipsi SR parallela ducitur extra Coni superficiem cadat, scilicet ducta, et altera curva, nempe Hyperbolica ipsa, quae cum eodem
manifestum est quod si SQ recta linea extra Conum protrahatur quousque plano et superficie Coni in infinitum productae semper sibi invicem magis
occurrat ipsi parallelae ductae per punctum P , in infinitumque productae proximabunt; nunquam tamen mille Stadiis sibi proximiores erunt. Alioqui
(coincident enim necessario, quod si quis neget, per ultimam Definitionem linea GM inflexa rectae EF oc/curret, quod fieri non posse iam demonstra­
et 30 prop, primi lib. eorundem Elemen. facilie probari potest) evadet tum fuit. Hoc autem Corollarium maxime admirandum est.
aequalis ipsi PR per 34 prop, eiusdem primi lib. Elemen. Maior igitur est
PR quam QS per 9 Com. Sent, eiusdem. Verum quoniam in longum pro­ Ante quam ad tertium instituti Problematis demonstrationem accedamus
vecti sumus ut hoc probaremus iam a digressione revertendo ad institutum quoddam Theorema nobis praedemonstrandum est, in quo tota vis illius
dicimus quod cum PR sit maior quam Q S, eam autem (ut ostensum fuit) ra­ demonstrationis consistere videtur. Quidam enim tanquam manifestum
tionem habet PR ad QS quam habet MS ad G R . Et MS igitur quam GR hoc supponentes demonstrare se credidere, cum tamen nugentur in Geo­
maior est per 9 Com. Sent, huius. At GR et MS ad rectos angulos ipsi EF metricis namque demonstrationibus nil tanquam manifestum assumendum
minime sunt; cum anguli qui ad R et S acuti iam ostensi sint atque prop- est, quin ipsum vel ab aliis satissuperque demonstratum, vel ab omnibus
terea ipsae MS, GR non sunt brevissima intervalla quibus G et M signa in­ tanquam principium nulla demonstratione indigens concessum, recep­
flexae lineae a recta EF distare possint: eo quod ab eisdem punctis ad ip­ tumque sit. Theorema igitur, quod praemittimus, sit huiusmodi.
sam EF rectam lineam perpendiculares duci possunt, quae per 19 prop.
74 primi lib. Eleme. Eucl. ipsis MS, GR / breviores erunt, imo brevissimae Lemma seu Assumptum Sequentis tertiae
omnium quae ab eisdem signis ad rectam EF duci possint. Omnes enim Demonstrationis.
aliae ab eisdem punctis ex quavis parte ductae perpendicularibus ipsis Aequales1rectae lineae in circulis inaequalibus inaequales cum maiorum,
maiores sunt per eandem 19 prop, eum maiorem angulum, scilicet rectum, tum minorum Segmentorum auferunt circunferentias, in minoribus nempe
subtendant. Nam una tantum perpendicularis ab eodem puncto ad eandem Segmentis maiorem quidem a minori, minorem vero a maiori: in maioribus
75 re/ctam lineam in eodem plano duci potest, quod patet ex 17 prop, primi autem Segmentis maiorem quidem a maiori, minorem vero a minori circulo
lib. Elemen. Eucl. Quae cum ita se habeant, ducantur per 12 prop, primi lib. circunferentiam.
eorundem Elemen. a signis G, M ad lineam rectam EF in plano EFKL Sint2 duo inaequales circuli, ABC quidem minor, DEF vero maior
perpendiculares GV et M X: et erunt anguli GVR et MXS per 4 pet. eiusdem [Fig. 7.53], in quibus duae rectae lineae AC et DF invicem aequales una
primi lib. Elem. Eucl. aequales, quia recti per 10 Definitionem eiusdem cum circulorum circunferentiis minora quidem Segmenta circulorum
sunt. Quoniam autem GR et MS parallelae ex Constructione sunt: anguli contineant ABC et DEF: maiora vero AGC ,D H F . Dico3 quod in minoribus
etiam GRV et MSX per secundam partem 29 prop, primi lib. eorundem quidem Segmentis circunferentia ABC circuli minoris est maior quam cir-
Elem. sunt aequales. Ergo per 32 prop, et per 3 Com. Senten. eiusdem cunferentia DEF circuli maioris: in maioribus autem Segmentis aio circun­
triangula GRV etM SX aequiangula sunt. Atque idcirco per4prop. sexti lib. ferentiam DHF maioris circuli circunferentia AGC minoris circuli maiorem
eorundem Elemen. ratio ipsius MX ad GV est sicut ratio ipsius MS ad GR. esse. Dividatur4 itaque recta linea DF per decimam / prop, primi lib. Elem.
Sed MS maior est quam GR (ut probatum fuit); ergo et MX quam GV per 9
Com. Sent, huius maior est. Sunt autem MX et GV minimae distantiae 13 mg.: Conclusio universalis.
quibus signa G, M Hyperbolicae lineae a recta linea EF distare possint.
1 mg.: Propositio.
Igitur12 signum G est proximius rectae EF quam signum M, quod quidem 2 mg.: Expositio.
3 mg.: Determinatio.
mg.: Conclusio secundae partis. 4 mg.: Constructio primae partis.
402 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZTS AD M IRAN D U M PROBLEMA 403

Eucl. in duas partes aequales in signo /, a quo erigatur ipsi DF ad angulos necesse est si super recta linea DF in partes P simile et aequale segmentum
rectos per 11 propo. eiusdem recta linea quae utrinque per secundam peti­ ipsi ORP constituatur, ut eius circunferentia cadat extra DEF circun­
tionem eiusdem producta secabit circunferentiam DEF, verbi gratia, in ferentiam segmenti circuli maioris: alioquin QR aequalis esset ipsi IE vel
signo E, et transibit per centrum circuli DEF per Corollarium primae minor quam ipsa, cum tamen maior esse ostensa iam sit. Verum si circun­
propositionis tertii libri Elem. Eucl., et secabit ex altera parte eiusdem ferentia ipsi ORP vel ABC aequalis cadit extra circunferentiam DEF,
circuli circunferentiam, utpote in sign o//. Deinde quoniam semidimetiens perspicuum est ipsam ORP, seu ABC, ipsa DEF esse maiorem per defini­
circuli DEF est maior semidimetiente circuli ABC per suppositionem et per tionem lineae rectae. Quandoquidem recta linea definitur minima omnium
30 Definitionem huius: resecetur per 3 prop, primi lib. Elem. Eucl. a semi­ eosdem cum ipsa terminos ha/bentium linearum: vel a puncto in punctum
dimetiente maiore semidimetienti minori aequalis recta linea EK: et centro brevissima extensio: vel quae ex aequo inter signa sua sita est, ex his enim
K , intervallo autem KE, describatur per 3 pet. primi lib. Elem. Eucl. cir­ rectae lineae definitionibus clarum est quod omnes curvae lineae eosdem
culus ELMN, qui necessario tanget intrinsecus circulum DEF, et nullibi cum recta terminos possidentes, et ad easdem partes constitutae, inter se sunt
78 nisi in signo E per 11 et 13 prop, tertii lib. Elem. Eucl. Quamobrem seca/bit inaequales: et remotiores quidem a recta proximioribus semper maiores,
eiusdem ELMN circuli circunferentiam linea recta DF in duobus signis, ut ut sensui conspicuum, ab omnibusque concessum est. Patet6 igitur prima
puta L, N . Igitur recta linea LN est minor per 9 Com. Sent, primi lib. huiusce Theorematis pars. Secunda vero sic constabit. Sint7 duo circuli
eorundem Elemen. quamDF, hoc est quam AC. Atque idcirco remotior est inaequales, minor quidem ABC [Fig. 7.54], maior vero DEF: etin ipsis duae
a centro circuli ELMN quam recta linea ipsi AC aequalis per conversam rectae lineae AC et DF invicem aequales continentes cum circunferentiis
quintaedecimae prop. lib. tertii Elem. Eucl. minorque est circunferentia segmenta minora quidem ABC, DEF: maiora vero AGC ,D H F . Dico8 quod
79 LEN quam ABC per ultimam propositionem lib. sexti Elem. Eucl., quia / si DHF circunferentia maior est quam AGC. Dividatur9 igitur ut superius
a centris circulorum ABCG, ELMN ad AC, LN signa rectae lineae ductae rectaDF in duas partes aequales in signo/, et erigatur IE ad angulos rectos,
intelligantur, erunt anguli ad centra circulorum aequalium constituti, et / producatur utrinque ut transeat per centrum et secet circunferentiam
quorum ille quidem qui ABC circunferentiae insistit maior erit eo qui LEN circuli in signis E ,H . Deinde circa centrum K circulus aequalis ipsi ABC eo
circunferentiae insisteret per 25 prop, primi lib. Elemen. Eucl., latera enim modo quo superius describatur, tangens circulum maiorem in signo E, et
eorum essent aequalia alterum alteri et basis AC base LN maior. Quare secans rectam quidem DF in signis L, N , dimetientem vero EH in signo M.
cum circunferentia LEN minor sit quam ABC, etiam AGC circunferentia Et quoniam his ita iacentibus superius ostensum est circunferentiam ABC
minor erit quam LA/N per 4 com. senten. huius. Unde per eandem ultimam esse maiorem circunferentia LE N , inaequales igitur angulos capiunt per ul­
sexti segmenta ABC et LEN inaequales capiunt angulos. Ergo per 31 defini­ timam propositionem sexti lib. Elem. Eucl. Segmenta AGC et LM N , et
tionem huius dissimilia sunt. A dato igitur circulo ELMN abscindatur per ideo dissimilia sunt per 31 Definitionem huius. Abscindatur itaque a circulo
34 propositionem 3 lib. eorundem Elementorum segmentum OMP capiens ELMN per 34 prop, tertii, lib. Elem. Eucl. Segmentum ONEP suscipiens
angulum aequalem cuilibet angulo rectilineo in segmento ABC existenti, angulum aequalem cuivis angulo in Segmento AGC existenti, quod utique
quod quidem OMP segmentum erit simile segmento ABC per 10 defini­ Segmentum erit simile et aequale AGC Segmento rationibus superius
tionem eiusdem tertii atque circunferentia ABC per 26 prop, tertii et tertiam dictis. Potest autem aliter etiam abscindi Segmento AGC simile et aequale
Com. Sent, primi lib. Elem. Eucl. est aequalis circunferentiae OMP, et per Segmentum ONEP, scilicet accommodando per primam propositionem
29 propo. tertii libri eorundem OP recta linea aequalis est rectae AC, et quarti libri Elemen. Eucl. in circulo ELMN rectam PO aequalem rectae
totum segmentum ABC toti segmento OMP per 24 propositionem eiusdem. AC, erit enim per 28 prop, tertii lib. Elem. Eucl. circunferentia ABC
Cum autem OP centro K propinquior sit (ut iam dictum est) quam LN , aequalis circunferentiae OP. Quare per 26 prop, eiusdem tertii lib. Seg­
minor est distantia ipsius OP a centro K (quae sit KQ producta quousque menta AGC et ONP suscipiunt angulos aequales. Unde per 10 Definitionem
secet circunferentiam OMP in signo R) quam distantia KI. Quoniam vero eiusdem similia sunt. Cum autem sint super aequalibus rectis lineis consti­
omnes eiusdem circuli semidimetientes aequales sunt, QR maior sit quam tuta, proculdubio per 24 prop, eiusdem aequalia quoque erunt. Hoc itaque
IE per 4 Com. Sent, huius. Cum5igitur QR maior sit quamIE, et OP aequalis facto recta linea OP, basis nempe Segmenti OEP, secetur per medium in
ipsi AC, hoc est ipsi DF, est autem IE quidem ad angulos rectos ipsi DF, signo Q , a quo erigatur ipsi OP ad rectos angulos recta linea, quae protracta
eamque per medium dispescens per constructionem, QR vero similiter ad utrinque transibit per centrum K per Corollarium primae propositionis
rectos angulos ipsi OP, et per medium ipsam secans per 4 definitionem et
secundam partem tertiae propositionis tertii libri Elementorum Eucl. 6 mg.: Conclusio primae partis.
7m g Expositio secundae partis.
8 mg.: Determinatio secundae partis.
5 mg.: Demonstratio primae partis. 9 mg.: Constructio.
404 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZLS AD M IRAN D U M PROBLEMA 405

tertii lib. Elem. Eucl., secabitque circunferentiam circuli minoris in signis sectio plani trianguli ABC et plani GFH, coincidet per quintam petitionem
R>S. His ita constructis,10 quoniam OP maior est quam LN (ut superius primi libri Elementorum Eucl. cum latere AC ipsius trianguli extra Coni
fuit ostensum), ergo centro K est propinquior per conversam 15 proposi­ Verticem producto. Quoniam per Constructionem et 16 prop. 11 lib.
tionis tertii lib. eorundem Elementorum. Igitur QK distantia minor est Elemen. Eucl. ipsa communis sectio axi Coni parallela est; et ideo si
quam distantia IK. Quare QS maior est quam IE per 4 Com. Sent, huius: protracta intelligatur rectaBC, faciet per 12 Definitionem huius et per 29, et
nec non QR minor est quamIM per eandem; unde multo minor quamIH. Si 32 prop, primi lib. Elem. Eucl. cum iam dicta exeunte et AC rectis lineis
igitur super recta linea DF in partem H constitutum fuerit Segmentum intra Conum duos angulos duobus rectis minores. Huius itaque Hyper-
ONP, necesse est eius circunferentiam cadere intra circunferentiam DHF: 84 bo/lis planum intelligatur directe et interminate protractum ad partes FG.
alioquin/D esset aequalis ipsi QR aut minor quam ipsa, quod est contra ea Subinde quoddam aliud planum superficiei Coni sic applicetur ut ipsam
quae ostensa sunt. Quapropter ex Definitionibus rectae lineae superius tangat in tota recta AD linea, planoque trianguli ADE rectum sit: hoc autem
82 dictis patet circunferentiam DHF esse maiorem circunfe/rentia ONP, seu fiat quemadmodum in superiori Constructione. Coextenso2igitur hoc plano
quavis alia quae sit ipsi AGC aequalis. Atque11 hoc erat secundo demon­ versus Hyperbolem, secabit eius planum. Cum enim secet planum trianguli
strandum. Patet autem haec secunda pars etiam ex sexta Communi Senten­ ADE parallelum plano GFH, necessario et ipsum GFH secabit, aliter neque
tia huius. Aequales12 igitur rectae lineae in circulis inaequalibus, et reliqua etiam ipsum ADE secaret. Quoniam plana quae eidem plano sunt parallela
ut in propositione. Quod erat demonstrandum, atque praeassumendum. et inter se parallela sunt (ut recte demonstrant Campanus in 16 prop. lib. 11
85 Elemen. Eucl. et Vitellio / in 14 prop, primi libri suae Perspectivae). Paral­
Corollarium. lela autem plana sunt quae invicem non coincidunt per 8 Definit, eiusdem
undecimi lib. Elem. Eucl. Secent igitur sese duo haec plana, et sit com­
Ex demonstratione huius Theorematis constat quod in minoribus in­ munis eorum sectio per 3 prop, eiusdem undecimi lib. recta linea IK , quae
aequalium circulorum Segmentis aequales bases habentibus, partes erit parallela rectae AD per 16 prop, eiusdem. Atque idcirco omnes rectae
dimetientium ipsorum circulorum, quae tum Segmenta ipsa, tum eorum lineae quae in plano ADKI ducentur ad angulos rectos ipsi AD erunt etiam
bases per medium dividunt, inaequales sunt: maior quidem minoris di­ ad rectos angulos ipsi IK communi sectioni per 29 prop, primi lib. Elem.
metientis, minor vero maioris. In maioribus autem inaequalium circu­ Eucl. Quamobrem recta AI ducta ad angulos rectos ipsi ADE plano ad
lorum Segmentis praefatae dimetientium partes e contrario sunt in­ rectos angulos est ipsi IK. Est autem etiam per 3 Definitionem et 16 prop. 11
aequales: maioris quidem dimetientis maior, minoris vero minor. et 29 prop, primi lib. Elementorum Eucl. ad angulos rectos ipsi IF axi
Patuit enim in prima configuratione [Fig. 7.53] ipsam QR esse maiorem Hyperbolis, si ducta intelligatur. Ergo per quartam prop, eiusdem 11 lib.
ipsa/E , pariterque in secunda descriptione [Fig. 7.54] ipsam QR ipsa IH AI recta linea plano Hyperboles ad rectos erit angulos. Et propterea per 4
minorem esse. Quod sane Corollarium et pulcrum est, et sequenti Demon­ Definitionem vel per 18 prop, eiusdem planum ADKI ad planum ipsius
strationi maxime opitulaturum. Hisce autem praemissis modo ad institu­ Hyperboles rectum erit. His ita expositis atque constructis, dico3 quod
tum revertamur. recta linea IK et inflexa GF in eodem plano existentes nunquam sibi oc­
current, etiam si in infinitum una cum ipso Cono protrahantur: et quanto
EIUSDEM PRAECIPUI PROBLEMATIS magis producuntur, tanto propiores erunt adinvicem. Quod4 igitur sibi nun­
DEMONSTRATIO TERTIA. quam occurrant ex eo manifestum est, quod recta quidem IK in plano
Sit1 igitur Conus ABCD [Fig. 7.55], cuius vertex A, basis vero BDCE ADKI iacet, quod (ut in superioribus ostensum est) nunquam tanget
circulus, et triangulum per axem ADE, cuius basis recta linea DE, et aliud Conum alibi quam in recta AD linea, inflexa vero GF linea nusquam a Coni
per axem triangulum ABC. Sit rursum aliud quoddam planum GFH plano superficie dimovebitur. Quare nunquam recta IK tanget Coni superficiem:
trianguli ADE parallelum, conum ipsum sub inflexa GFH linea inaequaliter neque5 igitur lineam FG, ipsi conicae superficiei inhaerentem, quod erat
secans, quae quidem inflexa linea per 21 Definitionem huius una cum recta primo demonstrandum.
83 GH continet / Hyperbolem sectionem conicam. Quandoquidem si per sig­ Quod autem istae duae lineae quanto magis producuntur, tanto propiores
num E (quod erit vertex ipsius Hyperbolis) exire intelligatur communis adinvicem sint dilucide ostendetur, paucis prius constructis. Suscipiantur6

10 Demonstratio secundae partis. 2 mg. : Hoc non probat Orontius.

11 mg.: Conclusio secundae partis. 3 mg.: Determinatio.
12 mg.\ Conclusio universalis. 4 mg.: Demonstratio primae partis.
5 mg.: Conclusio primae partis.
mg.: Expositio, et Constructio. 6 mg.: Constructio secundae partis.
406 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZTS AD M IRAN DUM PROBLEMA 407

itaque in ipsa inflexa linea duo quaelibet signa L, M, per quae transeant una cum ipso Cono producentur, tanto sibi invicem proximiores evadent.
duo circuli sibi invicem et Basi Coni paralleli, quorum circunferentiae in Quod secundo demonstrandum erat. Duas10 igitur in eodem plano lineas,
superficie conica ab eorundem planis Conum secantibus designatae sint et reliqua ut superius. Quod facere oportebat.
LNO quidem minoris culminique propinquioris: MPQ vero maioris basique
proximioris. Et comprehensis inter lineas AD et FG eorundem circulorum / DE DUABUS LINEIS, RECTA ET CURVA,
circunferentiis LN et MP fiant eis aequales NO et PQ circunferentiae; quod NON COINCIDENTIBUS,
fiet per primam petitionem bis sumptam et 23 propositionem primi libri ET MAGIS SEMPER INVICEM APPROPINQUAN­
Elementorum Euclidis, seu per eandem primam petitionem et primam propo- TIBUS IN DIVERSIS PLANIS.
86 sitionem quarti eorundem. Deinde per eandem primam petitionem / ducantur Priusquam1 reliquas instituti Problematis demonstrationes persequar
rectae lineae LO, MQ, quae per 3 Com. Sent, primi et 27 et 29 proposi­ commodum mihi videtur hoc in loco demonstrare hanc eandem affec­
tionem tertii et 4 prop, et decimam Definitionem primi libri Element. Eucl. tionem veram esse de duabus etiam lineis, altera similiter Hyperbolica et
in duas aequales partes et ad rectos dividentur angulos a communibus altera recta, ambabus in Coni superficie, sed non in eodem plano, tacenti­
sectionibus planorum utriusque circuli et plani trianguli ADE. Sit igitur bus. Quod etiam ab Orontio in suo libello de speculo ustorio, et ab antiquo
secta ipsa LO in puncto R, ipsa vero MQ in puncto S. Et ducantur com­ innominato Autore in fine libelli de sectione Parabolae, quamvis satis ob­
munes sectiones NR et PS, quae erunt parallelae per 16 prop. 11 lib. Ele- scure, imperfecteque demonstratum tamen fuit.
men. Eucl. Ducantur praeterea per 31 prop, primi lib. eorundem Elem. per
Maneant2 igitur cuncta sic disposita ut in superiori proxima Construc­
puncta N ,P ipsis RL, SM parallelae rectae lineae N T ,P V , quae productae
tione, et per primam petitionem primi libri Element. Euclid, ducantur LN
87 in plano ADKI secabunt IK rectam de necessita/te per 30 propositionem
et MP dimetientes parallelogrammorumLRNT et MSPV [Fig. 7.55]. Dico3
et ultimam Definitionem primi libri eorundem Element. Secent ipsum in
duas lineas FG inflexam et AD rectam in superficie conica non in eodem
punctis T, V, et ducantur LT et MV rectae lineae, quae per 16 prop, libri
plano iacentes in infinitum cum ipso Cono protractas semper magis atque
11 eorundem Element, erunt parallelae ipsis RN et SP. His ita constructis,7 magis sibi invicem proximari, nunquam tamen sibi occurrere. Quum4 enim
quoniam per Constructionem et 29 et 34 prop, primi lib. eorundem Elemen.
rectae lineae LR, MS in eisdem parallelis sint planis, erunt parallelae et
parallelogramma rectangula sunt ipsa LRNT et MSPV, et latera ex op­
aequales per 16 prop. 11, et 34 prop, primi lib. Elemen. Eucl., intelligendo
posito habent aequalia, igitur LT ipsi RN , et MV ipsi SP, aequales sunt.
scilicet rectas LM, RS lineas esse ductas. Atqui R N maior est quam SP
Sed RN maior est quam SP per Corollarium praeoste[n]si Lemmatis. Et
per Corollarium praeassumpti Lemmatis. Triangula igitur LRN, MSP per 4
LT igitur ipsa MV maior est per 14 propositionem quinti libri Element.
Com. Sent, primi lib. Element. Eucl. habent duo latera LR, R N simul
Eucl. Si itaque LT et MV ad rectos essent angulos ipsi IK , haberemus
sumpta duobus lateribus MS, SP simul sumptis maiora, et rectos nihilo­
intentum. Quoniam autem non sunt, angulis ITL et IVM acutis existentibus
minus angulos comprehendentia. Quadratum ergo ipsius LN quadrato
(ut patet per Constructionem, et octavam propositionem, et tertiam Defini­ ipsius MP per 47 prop, eiusdem primi lib. maius est. Unde per 2 Com.
tionem 11, et 29 et 32 prop, primi libri Elem. Eucl. si IF axis Hyperbolis, et
Sentent. huius LN recta linea maior itidem est quam ipsa MP. Si itaque
ipsae TL, VM protractae intelligantur) ducantur per 12 prop, eiusdem primi
LN et MP rectae lineae ad ipsam AD rectam lineam perpendi/culares
lib. Element, a punctis L, M ad rectam IK perpendiculares LX et MY
sunt, habemus intentum. Sin minus, ducantur a punctis L , M ad rectam AD
rectae lineae. Cum igitur LT et MV (ut iam ostensum est) parallelae sint,
perpendiculares, et ostendetur superioribus rationibus inflexam FG
proculdubio triangula LTX et MVY aequiangula sunt per 29 prop, et 4 pet.
lineam rectae AD propinquiorem esse in puncto M quam in puncto L: et sic
et 32 prop, et 3 Com. Senten. primi lib. eorundem Element. Quare per
in infinitum, et nihilosecius nunquam coincident, cum in duobus planis
quartam prop, sexti lib. eorundem Elemen. ratio ipsius LT ad MV est sicut
parallelis sint, sed aliquod semper inter eas erit interstitium maius quam
ratio ipsius LX ad MY. Atqui LT maior quam MV fuit ostensa; ergo per recta linea utrique plano perpendicularis. Si enim aliquando coinciderent,
9Com. Sent, huius et LA ipsa ML maiorest. Propior8est itaque lineaFG in­
dubioprocul plana quoque ipsa parallela tunc sibi occurrerent, quod per 8
flexa rectae IK in puncto M quam in puncto L. Haud dissimiliter autem si
Definitionem 11 lib. Elementorum Euclid, nequaquam fieri potest. Duae5
describatur sub MPQ circulo alius quispiam circulus ipsi MPQ parallelus,
concludetur iterum eadem FG linea in eius cum eodem circulo sectione pro­ 1(1 mg.: Conclusio universalis.
pinquior esse eidem IK rectae lineae quam in signo M\ idemque in infinitum
ostendi potest. Quanto9 magis igitur praefatae lineae ad inferiores Coni partes 1 mg.: Prop.
2 mg.: Expositio, et Constructio.
7 mg.: Demonstratio secundae partis. Λmg.: Determinatio.
Kmg.\ In hoc deficit Orontius. 4 mg.: Demonstratio.
9 mg.: Conclusio secundae partis. 5 mg.: Conclusio.
408 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZTS AD M IRAN D U M PROBLEMA 409

90 itaque lineae in una superficie conica sed / in diversis planis describi pos­ enim intermedio lateri trianguli per axem semper magis magisque ap­
sunt quae quanto magis in continuum producentur, tanto sibi propin­ propinquent, cum ipsoque nunquam coeant (ut supra patuit) quod etiam
quiores evadent, nunquam tamen invicem coincident, etiam si in infinitum sibi continue in infinitum propiores fiant, nunquam tamen invicem con­
protractae fuerint. Quod erat demonstrandum. veniant, luce iam clarius relinquitur.

DE DUABUS LINEIS CURVIS [Extract IV: pp. 91-92]

NON COINCIDENTIBUS, QUAEDAM ELEMENTA CONICA
ET MAGIS SEMPER SIBI INVICEM PROXIMANTIBUS QUARTAE DEMONSTRATIONI
TUM IN EODEM, TUM IN DIVERSIS DESERVIENTIA
PLANIS.
H ic1amplius praetereundum non est, quod quemadmodum duas lineas, Declaratis iam atque restauratis tribus praecedentibus Demonstra­
inflexam scilicet et rectam, cum in eodem, tum in diversis planis, ac­ tionibus, in praesentia consequens est propositum nobis Problema iuxta
cidentia supradicta patientes iam descripsimus: ita duas etiam curvas doctrinam Apollonii Pergaei de Conicis tractantium principis demonstrare.
lineas, tum eodem, tum diversis in planis describere possumus, quip- Ipse enim, licet imperfecte, exactius tamen caeteris omnibus Autoribus,
pequae iisdem passionibus afficiantur. duas iam dictas lineas in eodem plano in infinitum productas nunquam
Iacentibus2 itaque cunctis quemadmodum in secunda, et tertia propositi invicem coincidere et continue sibi propinquiores fieri theorematice
Problematis demonstratione, si construatur alius Conus iam extructo demonstravit in prima et quartadecima propositionibus secundi libri
aequalis et similis; priorique posterior ita incumbat ut recta linea AD sit suorum Conicorum Elementorum, in quarta vero eiusdem secundi libri
communis utrique, planum vero Hyperbolis FG triangulari ADE plano demonstratione (quae Apollonii non est, sed potius Eutocii, ut ibi recte
parallelum continue directeque extendatur donec secat posteriorem, Federicus Commandinus adnotavit) iam dictae non coincidentes lineae
incumbentemve Conum eodem modo quo secet anteriorem succumbentem: problematice secundum Apollonii doctrinam describuntur. Volentibus
statim liquebit propositum. Communis3 enim sectio iam dicti Hyperboles igitur nobis de mente Apollonii nostrum admirandum Problema demon­
FG plani et conicae superficiei novissimi Coni erit linea curva similis in­ strare, necessarium erit tres commemoratas propositiones, primam scilicet
curvationis cum inflexa FG linea, quae duae curvae lineae non desinent et quartam et quartamdecimam secundi libri Conicorum Elementorum
continue propiores invicem fieri, quoniam utraque ipsarum per ea, quae Apollonii / in unam problematicam propositionem reducere. Quoniam
hucusque demonstrata sunt magis magisque appropinquant rectae lineae vero tres iam dictae propositiones in unum reduci perfecteque intelligi non
IK eodem in plano eas interiacenti: et nihilosecius nunquam sese tangent, possunt ni prius quaedam Elementa conica ab Apollonio in primo libro
etiam si in infinitum una cum duobus Conis producantur, quandoquidem suorum Elementorum conicorum demonstrata recte percipiantur: idcirco
neque etiam cum recta IK inter ipsas media concurrent. Haec autem necessarium esse existimo ante quam ad propositi Problematis demon­
imaginatione potius quam ostensione indigent, quippe cum ab iis qui strationem accedamus ea primum hic declarare. Et praesertim quoniam
91 superio/res Conos extructos prae oculis habent perpulere quidem excogi­ Apollonius in Elementis suis propter nimiam brevitatem et multorum Lem­
tari possunt. Atque idcirco nullam huiusce rei configurationem subiicere matum suppositionem obscurissimus est, ego autem (ut hoc opus nostrum
libuit. omnino clarum sit) aliquantulum prolixiori, sed perfectiori quoad fieri
Age4 modo ostendamus quomodo duae curvae lineae in una superficie poterit sermone ea declarabo, verbis ipsius Apollonii me nequaquam
conica, diversis tamen in planis affectionibus supra dictis succumbere of­ obligans: sed potius mentem ipsius amplioribus lucidioribusque verbis
fendantur. Si igitur in quolibet Cono duo plana sibi invicem et plano trian­ explicans. Nam Graeca quidem Apollonii Elementa adhuc edita non sunt.
guli per axem parallela superficiem conicam intersecent: fient porro in ipsa Antequam autem a Federico Commandino Latina ederentur, ego ea
conica superficie duae Hyperbolicae lineae, quae quanto amplius una cum maximo cum sudore in ipsa perversa ac sordida Ioannis Baptistae Memi
Cono producuntur, eo magis ac magis sibi proximant, nunquam tamen tralatione ex ingenio correxi atque dilucidavi: postea vero cum quodam
coincidentes adinvicem, quamvis etiam in infinitum extendantur. Cum5 etiam exemplari Graeco manuscripto (quod apud ipsum Commandinum
erat) ea contuli. Erant enim propter tralatoris illius Graecarum literarum
et Mathematicarum scientiarum ignorationem adeo imperfecta ut nil
1 mg.: Propositio. faedius unquam legi posset: quandoquidem maximo cum labore quid sibi
2 mg.: Expositio, et Constructio in eodem plano.
3 mg.: Demonstratio.
velet Apollonius vix coniicere quispiam poterat. Quapropter ex iisdem a
4 mg.: Expositio, et Constructio in diversis planis. nobis iam correctis Elementis conicis ea quae in praesentia proposito
5 mg.: Demonstratio. nostro deserviunt, qualia tunc pro viribus instauravimus, talia nunc in
410 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZFS AD M IRAN DUM PROBLEMA 411

medium sumus allaturi. Primum igitur Elementum Conicum a nobis de­ exactius propositum habebitur. Earundem itaque sectarum partium prima
clarandum, ac illustrandum duodecima primi libri propositio apud Apol­ sit GK, et per signum K ipsi DEC rectae lineae parallela ducatur per tri­
lonium est. Quoniam autem in eius expositione quoddam ab Apollonio cesimamprimam propositionem primi lib. Elementorum Euclid. K L . Atque
Lemma supponitur, quod in fine undecimae propositionis eiusdem primi ex ipsa KL dematur KM per / tertiam propositionem eiusdem aequalis
libri breviter et particulatim Eutocius demonstrat: propterea priusquam rectae lineae, cuius quadratum per ultimam propositionem secundi libri
dictam duodecimam propositionem declaremus, illud Lemma exquisitiori eorundem Elementorum factum sit aequale parallelogrammo rectangulo ab
et universaliori demonstratione confirmabimus, quippequae diversa sit ab IK, KG comprehenso. Quod etiam commode fieri potest, quoniam quadra­
illa Eutocii. Erit enim Problema non inutile huiusmodi. tum lineae XL maius quidem est rectangulo abIK ,K G contento, rationibus
ante dictis. His itaque sic constructis, dico3 quod si iam dictae parallelae
[Extract V, pp. 156-63] crebriores quoad fieri poterit peragantur, atque in ipsis similia signa,
156 /. . . DEMONSTRATIO qualia suntL, M pari Constructione capiantur, eaque rectis connectantur
SEPTIMA. lineis: inflexa quaedam creabitur linea Hyperboles lateri haud absimilis,
cui AB, BC rectae lineae continue propiores fient; nunquam tamen oc­
Sit1 quodcunque planum propositum; volo super ipso duas describere current, etiam si in infinitum protractae fuerint. Quum4enim per Construc­
lineas, alteram rectam, alteram inflexam, quae duas iam saepe dictas af­ tionem angulusBDC rectus sit, et GH parallela ipsi CD: erit per vicesimam­
fectiones subeant. Suscipiatur2 in proposito plano rectus angulus ABC, nonam propositionem primi libri Elementorum Euclidis angulus BGH
[Fig. 7.56], qui dividatur per nonam propositionem primi libri Elemen­ rectus. Sed angulus GBH itidem per Constructionem est recti dimidium:
torum Euclidis in duas partes aequales producta recta linea BD, et a signo ergo per tricesimamsecundam propositionem eiusdem angulus etiam BHG
D per undecimam propositionem eiusdem erigatur ipsi BD ad rectos an­ recti dimidium existit. Unde per sextam propositionem eiusdem GH
gulos recta linea quae utrinque producta secabit per Constructionem et aequalis est ipsi BG, cuius quadratum aequale per Constructionem est
quintam petitionem eiusdem rectas AB,BC in signis quae sint A , C. Deinde rectangulo ab FE (! FC), EC contento. Igitur per secundam Com. Sent,
inter signa D ,C , seu D , A , quodlibet accipiatur in ipsa ADC recta linea sig­ huius et primam Com. Sent, primi lib. Elem. Eucl. quadratum etiam ip­
num E, sitque illud in praesentia susceptum inter signa D, C, et ab ipsa sius GH eidem rectangulo ab FE (! FC), EC comprehenso aequale est.
AD ipsi DE per tertiam propositionem primi libri Elementorum Euclid, Quare per 2 Com. Sent. pri. Ii. eorundem Ele. rectangulum ab FE (/ FC),
abscindatur aequalis DF, nanque DA et DC et DB inter se aequales sunt; EC contentum una cum quadrato ipsius DE est aequale quadrato ipsius GH
157 necnon AB ipsi BC per Constructio/nem, et 32 et sextam prop, eiusdem simul cum eodem ipsius DE / quadrato. At rectangulum ab FE (/ FC), EC
primi lib. Elementorum. Subinde ex BD auferatur per eandem tertiam contentum cum quadrato lineae DE per sextam propositionem secundi libri
primi pars BG aequalis uni rectae lineae cuius quadratum sit per ultimam eorundem Elementorum aequale est quadrato lineae DC. Ergo per primam
propositionem secundi libri Elementorum Eucl. factum aequale paral- Com. Sent, eiusdem primi quadratum etiam ipsius GH cum quadrato ipsius
lelogrammo rectangulo ab FE (/ FC), EC contento. Hoc enim commode DE aequale existit eidem quadrato lineae DC. Quadratum autem lineae
fieri potest, quandoquidem quadratum lineae DC maius quidem est iam DC per quartam propositionem secundi libri eorundem Elem. aequale est
dicto rectangulo per sextam propositionem secundi et nonam Comm. Sent, quadratis linearum DE, EC et duplo eius quod a DE, EC continetur rectan­
primi lib. Elementorum Euclidis; aequale vero quadrato lineae BD per gulo. Igitur per primum Com. Sent, eiusdem primi Elementorum et qua­
Constructionem et secundam Comm. Sent, huius. Demum per G signum dratum lineae GH cum quadrato ipsius DE eisdem duobus linearum DE,
ducatur per tricesimamprimam propositionem primi lib. eorundem Ele­ EC quadratis et duplo rectanguli a DE, EC comprehensi aequalia sunt.
ment. GH parallela ipsi DC secans necessario per vicesimamnonam Quamobrem per tertiam Comm. Sent, eiusdem primi Element., communi
propositionem et quintam petitionem eiusdem primi BC lineam in sign o//. ablato quadrato lineae DE, quadratum ipsius GH aequale est quadrato
Et producatur per secundam petitionem eiusdem primi in alteram partem ipsius EC simulque duplo rectanguli a DE, EC contenti. Praeterea quoniam
quousque per easdem secet etiam lineam AB. Postea vero producatur DB per Constructionem IB aequalis est ipsi BG, et GK in rectum additur: erit
in partem B interminate, et fiat per tertiam propositionem eiusdem primi per sextam propositionem eiusdem secundi Elementorum quadratum ip­
BI aequalis ipsi BG. Ipsa denique GD in aliquot utcunque secetur partes, sius BK aequale rectangulo ab IK, KG contento et quadrato ipsius BG. Et
atque per sectionum signa ipsi AC parallelae ducantur secantes AB, BC quia per Constructionem et 29 et 32 et sextam propositionem primi libri
rectas lineas; quanto autem crebriores ipsius GD sectiones fient, tanto
1 mg.: Expositio.
3 mg.: Determinatio.
2 mg. : Constructio.
4 mg.: Demonstratio.
412 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZFS AD M IRAN D U M PROBLEMA 413

eorundem Elementorum linea KL aequalis est lineae B K : erit per secundam omnia signa ipsis Μ, E, similia rectae continuentur lineolae. Atque2 ar­
Com. Sent, huius et primam Com. Sen. eiusdem primi libri quadratum ip­ guatur ut in praecedenti demonstratione (scilicet ibi [Praeterea quoniam
sius KL aequale rectangulo ab IK, KG contento et quadrato ipsius BG. per Constructionem IB, etc.] et fiat bis illa argumentatio) et propositum
Verum per Constructionem rectangulum ab IK , KG comprehensum concludetur.
quadrato ipsius KM est aequale. Ergo per secundam et primam Com. Sent,
primi lib. Elemen. Euclidis quadratum ipsius KL aequale est quadratis ip­ DEMONSTRATIO
sarum BG et KM. Sed per quartam propositionem secundi libri eorund. NONA
Element, quadratum KL est aequale quadratis linearum KM, ML et ei quod
Sit1 rursus quemadmodum superius angulus ABC rectus divisus per
bis a KM , ML continetur rectangulo. Ergo per eandem primam Com. Sent,
lineam rectam BD per medium, et ipsa AC ducta, et DB interminate pro­
et duo ipsarum BG, KM quadrata eisdem duobus ipsarum KM, ML quadratis
ducta, et in ipsa BD acceptum quodlibet signum G, factaque B l aequalis
et duplo rectanguli a KM , ML contenti aequalia sunt. Quare, ablato com­
ipsiRG [Fig. 7.56]. Sumatur deinde recta linea NO [Fig. 7.57] aequalis ipsis
muni quadrato ipsius KM , erit per tertiam Comm. Senten. eiusdem primi
ID, DG in directum coniunctis, fiatque per tertiam propositionem primi
libri Elemen. quadratum ipsius BG, seu ipsius GH (aequalia enim sunt per
libri Elementorum Euclidis OP aequalis ipsi D G , et erit PN quoque ipsi
secundam Com. Sent, huius), aequale quadrato ipsius ML et duplo eius
ID aequalis per tertiam Communem Sententiam eiusdem. Postea vero a
quod a KM , ML comprehenditur. Atqui paulo ante ostensum est idem
signo P erigatur ad angulos rectos ipsi NO per undecimam propositionem
quadratum lineae GH esse aequale quadrato ipsius EC et duplo rectanguli
eiusdem primi Elementorum recta PQ interminata ex parte Q, et secetur per
160 a DE, EC contenti: / igitur per primam Com. Sent. pri. lib. Elem. Eucl.
decimam propositionem eiusdem recta NO per medium in signo R. Et
quadratum ipsius ML et duplum rectanguli a KM, ML comprehensi aequalia
centro R, spatio autem RN , per tertiam pet. eiusdem describatur semi­
sunt quadrato EC lineae et duplo eius rectanguli quod a DE,EC rectis lineis
circulus NQO secans rectam PQ in signo Q. / Rursus in quot partes ipsa
continetur. Idem autem eodem modo potest ostendi in omnibus etiam aliis
GD secta fuit in totidem aequalesque ipsis per 10 prop, sexti lib. eorundem
parallelis per sectiones ipsius DG, tam in partem C, quam in partem A
Elem. ipsa quoque PO secetur, quarum OS aequalis sit ipsi DK. Et similiter
ductis. Quapropter5per conversum Corollarii primae superiorum praecipui
centro R et intervallo RS designetur semicirculus STV secans ipsam
Problematis demonstrationum inflexa EGF Hyperboles linea tali Construc­
quidem PQ rectam lineam in signo T, ipsam vero NR in signo V. His ita
tionis artificio in infinitum producta, semper ipsis magis, magisque ap­
constructis, dico2 quod recta linea PQ potest parallelogrammum rectan­
propinquabit, nec tamen cum ipsis unquam coincidet. In proposito itaque
gulum ab ID, DG contentum, et recta TP potest rectangulum quod ab IK,
plano duas in utraque parte descripsimus lineas, alteram inflexam et al­
KG comprehenditur, quod sic demonstrabitur. Quoniam3per 31 prop, tertii
teram rectam, inflexas quidem GE, GF, rectas vero BC, BA, quae iam
lib. Elemen. Eucl. angulus NQO (si NQ, QO rectae lineae ductae intel-
dictas duas subeunt affectiones. Quod faciendum erat.
ligantur) rectus est, et PQ per Constructionem perpendicularis ipsi NO:
erit per Corollarium octavae prop, sexti lib. eorundem Elem. ipsaP(? inter
DEMONSTRATIO
ipsas NP, PO media proportionalis. Quare per primam partem 17 prop,
OCTAVA
eiusdem quadratum ipsius PQ est aequale rectangulo quod ab Ν Ρ ,Ρ Ο con­
Sit1 ut pr[i]us angulus ABC [Fig. 7.56] rectus divisus per lineam BD in tinetur. Pari ratione quadratum ipsius PT aequale est rectangulo ab VP,PS
duas partes aequales, et ipsa AC ducta, et ipsaDR producta interminate, et contento. At rectangulum ab NP, PO contentum rectangulo ab ID, DG
in ipsaDR accipiatur quodlibet signum G, et fiat B l aequalis RG; deinde per contento et rectangulum ab VP, PS comprehensum rectangulo ab IK, KG
ultimam propositionem secundi lib. Elementorum Euclidis fiat quadratum comprehenso aequalia sunt per tertiam Com. Sent, huius. Nam NP qui-
aequale rectangulo ab ID, DG contento, et a linea DC per tertiam proposi­ /dem ipsi ID, et PO ipsi GD, per Constructionem aequales positae sunt. VP
tionem primi libri eorundem Elementorum auferatur DE aequalis lateri autem ipsi IK , et PS ipsi K G , sunt etiam aequales. Cum enim PO ipsi D G , et
161 iam / dicti quadrati; subinde similiter in linea GD suscipiantur crebriora SO ipsi D K , per Constructionem aequales sint: ergo F i”ipsiXG per tertiam
quoad fieri potest signa, per quae ducantur utrinque parallelae ipsi AC, Com. Sent. pri. lib. Ele. Eucl. aequalis est. Cum autem NR ipsiRG, et VR
quemadmodum ipsa K L , et per easdem ultimam secundi et tertiam primi ipsi R S , per 15 defin. eiusdem primi aequales sint: igitur, ablatis VR, RS
abscindatur ab ipsa KL pars KM potens parallelogrammum rectangulum
ab IK, KG contentum, idemque in caeteris parallelis fiat. Et a signo G per 2 mg.: Conclusio. Note that in this sentence the square brackets are Barozzi's, not mine.

5 mg.: Conclusio. 1 mg. : Constructio.

2 mg. : Determinatio proposito deserviens.
1 mg. \ Constructio. 3 mg.: Demonstratio eiusdem.
414 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S FR O M B A R O Z Z T S AD M IRAN D U M PROBLEMA 415

aequalibus, erit per eandem tertiam Com. Sent. NV aequalis ipsi SO : quare defecit, quoniam Theorema illud particulatim proposuit atque demon­
et ipsi KD per primam Com. Sent, eiusdem. Atque idcirco, ablatis NV et stravit, cum tamen universe verum sit, atque in proposito nostro in univer­
KD ab ipsis ID et NP aequalibus, remanent per eandem tertiam Com. Sent. sum tum proponi, tum demonstrari necessario debeat, alioquin proposito
VP et IK aequales. Cum4itaque haec ita sese habeant, manifestum est quod problemati quibusdam in Casibus deservire non poterit. Sic enim illud
PQ et PT rectae lineae possunt rectangula abID ,D G et ab IK, KG compre­ Vernerus proposuit. Si duo data rectangula inaequalium longitudinum
hensa. Quamobrem si ex DC abscindantur per tertiam prop, primi lib. quadratis suarum latitudinum iungantur, fuerintque haec duo aggregata
Elem. Eucl. DE aequalis ipsi PQ, et ex KL similiter KM aequalis ipsi PT invicem aequalia: erit quadratum aggregati maioris longitudinis minus
(quod rationibus superius dictis factu commodum est), ipsae etiam DE et quadrato aggregati brevioris longitudinis. Si igitur in proposito nostro (ut
KM eadem iam dicta rectangula poterunt. Similiter5 autem si centro R et in primae nostrae demonstrationis secunda figura [Fig. 7.51]) ipsaXO et PR
intervallis reliquis ipsius PO sectionibus semicirculi rectam PQ secantes latera quadratorum quibus parallelogramma rectangula adiungi debent,
describantur; caeterae quoque parallelarum per signa ipsius DG ductarum haud latitudines ipsorum rectangulorum, sed longitudines ambo; vel al­
partes talia rectangula potentes in ipsa PQ reperientur. Unde si ab ipsis terum quidem longitudo, alterum vero latitudo fuerint: quid nam dicendum
parallelis iam dictae etiam reliquae partes per tertiam prop. pri. lib. eorun­ erit? utrum in his etiam Casibus iam dictum Theorema nobis deservit?
dem Elem. ressecentur, et a signo G per omnia earum puncta ipsis E, M nonne inutile prorsus erit, cum de illis tantum rectangulis loquatur, quae,
punctis simila rectae lineolae continuentur: dubio procul superioribus cum inaequales habeant longitudines, quadratis suarum latitudi/num
argumentationibus propositum nobis quaesitum factum esse demon­ iunguntur? Quid enim siquis dubitet, ubi rectangulorum latitudines in­
strabitur. Institutum6 itaque nobis Problema in quocunque prostrato plano aequales supponuntur, ipsaque rectangula quadratis suarum longitudinum
sine ullo Conicorum corporum auxilio tripliciter hucusque iuxta tres iunguntur: vel quando latitudo unius longitudini alterius inaequalis sup­
diversas Constructiones demonstravimus. ponitur, ipsorumque rectangulorum alterum quidem quadrato suae longi­
tudinis, alterum vero quadrato suae latitudinis iungitur, sunt autem duo
[Extract VI, pp. 175-202] aggregata aequalia, an in his etiam duobus Casibus Theorema verum sit,
AUTORUM DE HAC RE TRACTANTIUM nec ne? quod scilicet quadratum aggregati maioris latitudinis minus sit
ERRORES quadrato aggregati minoris latitudinis, vel quod quadratum aggregati
175 / DIGRESSIO CONTRA VERNERUM. maioris longitudinis minus sit quadrato aggregati brevioris latitudinis.
Quod itaque in his duobis Casibus Theorema illud ita ut a Vernero pro­
Verum enimvero ut iam rem ipsam aggrediar Ioannes Vernerus Nurem- ponitur demonstraturque nullum nobis auxilium afferat, perspicuum est.
bergensis Mathematicus Clarissimus in libello suo de vigintiduobus Ele­ Quod vero propositum problema iuxta primum nostrum demonstrandi
mentis Conicis prope finem Problema de quo nunc agimus duobus modis modum duos etiam quos diximus Casus suscipere possit, quisque cog­
demonstravit; quorum alter quidem est ille quem nos in prima nostra noscere poterit si modo longe a summitate, modo prope summitatem
demonstratione instauravimus: alter vero quem in septima et octava et Hyperbolis parallelas ipsas duxerit. Tres3 enim omnino Casus inveniet,
nona nostris demonstrationibus illustravimus atque ampliavimus. Ante­ quorum unus est, quando4 ambae parallelae intra Hyperbolem ab uno eius
quam autem ad ipsius Problematis demonstrationem accederet, eas tres latere ad alterum ductae ambabus parallelis inter Hyperbolem et non coin-
ipse propositiones demonstravit quas nos etiam ante primam propositi cidentem extra Conum sibi in directum iacentibus maiores sunt, ut Vernerus
Problematis demonstrationem praedemonstravimus. Sunt1 autem apud accepisse videtur, quem Casum tanquam commodiorem nos etiam sus­
ipsum decimum septimum, decimum octavum, et decimum nonum Ele­ cepimus; et in hoc Casu rectangula inaequalium longitudinum quadratis
menta conica. Quanvis perperam ipse has tres propositiones ordinaverit, suarum latitudinum iunguntur. Secundus5 vero Casus est, quando ambae
quoniam secundam loco primae, et primam loco secundae posuit: cum iam dictae internae parallelae ambabus eisdem externis in directum sibi
tamen in ostendendo proposito prius secunda quam prima abutatur. Nos iacentibus minores sunt, verbi gratia, si in ipsa iam dicta nostra figura
vero eo ordine ipsas disposuimus quo ipsis utimur. In2 tertia itaque harum duplae ipsarum K L ,P Q rectarum linearum rectis K O ,PR minores essent;
trium propositionum, quae apud nos etiam tertia est, maxime Vernerus atque in hoc Casu rectangula inaequalium latitudinum quadratis suarum
4 mg.: Conclusio eiusdem. longitudinum adiunguntur. Tertius6autem Casus est, quando altera quidem
5 mg.: Applicatio ad propositum.
e mg.: Conclusio universalis. :i mg.: Tres casus secundae partis tertii nostri Elementi in principio positi.
4 mg. : Primus casus.
1 mg.: Ioannis Verneri prava ordo. 5 mg. : Secundus casus.
2 mg.: Ioannis Verneri defectus primus. 6 mg. : Tertius casus.
416 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S F R O M B A R O Z Z T S AD M IRAN DUM PROBLEMA 417

dictarum internarum parallelarum externa sibi in directum iacente parallela longitudinis propterea minus / est, quod rectangulorum longitudines
minor est, altera vero earundem internarum maior quam externa ei in quidem inaequales sint, aggregata vero aequalia (nam si latitudines etiam,
directum iacens; ac demum in hoc casu rectangula latitudinem longitudini vel latitudo, et longitudo inaequales supponantur, aggregata autem aequa­
inaequalem cum habeant, alterum quidem eorum quadrato suae longi­ lia; eadem affectio sequitur, ut iam diximus); sed quia rectangula quadratis
tudinis, alterum vero quadrato suae latitudinis iungitur. Quod igitur pro­ adiuncta unum quidem commune cum ipsis latus habent, duo vero in
positum Problema iuxta primum nostrum demonstrandi modum duos etiam directum iacentia, et in uno rectangulo maiora quam in altero, aggregata
hosce ultimos Casus suscipere possit credo nemini dubium esse. Potest autem aequalia sunt, ut secunda Theorematis pars proposuit. At si tum
autem hoc etiam sic confirmari. Pappus Alexandrinus ostendit (ut superius latera in directum iacentia unius lateribus in directum iacentibus alterius,
177 vidimus) / quod circa non coincidentes rectas lineas duae Hyperbolae tum aggregata aequalia fuerint: quadrata etiam quibus rectangula eo modo
describi possunt quae etiam inter sese non coincidentes sunt, et semper sibi adiunguntur aequalia sunt, ut in prima Theorematis nostri parte proposuimus.
magis appropinquant in infinitum productae. Iuxta hanc doctrinam igitur Haec igitur sunt subiecta quibus primo et per se et quatenus talia duae
circa non coincidentes rectas lineas huiuscemodi Hyperboles infinitas dictae aequalitatis et inaequalitatis affectiones insunt; non secus ac Trian­
unam intra aliam describere possumus. Unde manifestum est quod iam guli tres angulos duobus rectis aequales habere. Si enim duo haec subiecta
dictae internae parallelae continue minores, externae vero maiores fient. auferantur, hae quoque duae affectiones primo auferuntur: et si haec
Quod7 vero iam dictum Theorema in omnibus hisce Casibus universe ponantur, hae quoque primo ponuntur, cum aliis prius non insint. Sicuti
verum sit, facile ostendetur si demonstratio illa qua nos secundam eius etiam Triangulo ablato, affectio haec habere tres angulos aequales duobus
partem demonstravimus cunctis Casibus coaptabitur. Nos enim univer­ rectis primo aufertur: positoque primo ponitur, quoniam huic primo inest.
saliori quodam modo Theorema illud proposuimus atque demonstravimus. Ablatis autem rectangulorum longitudinis inaequalitate et aggregatorum
Cuius secunda quidem pars tribus iam dictis opitulatur Casibus, prima aequalitate, affectio inaequalitatis quadratorum non aufertur, quoniam
vero quamvis proposito nostro nullum afferat iuvamentum (quoniam inest etiam quadratis quibus rectangula inaequalium latitudinum; sive
nunquam parallelae intra Hyperbolem ab uno eius latere ad alterum ductae longitudinis, et latitudinis, adiuncta aggregata aequalia faciunt. Positis
aequales invicem sunt, sed basi Coni propinquiores remotioribus semper rursus illis, affectio primo non ponitur, cum aliis etiam (ut ostendimus)
maiores, ut ex decima Com. Sent, huius, vel ex Corollario primi prae- subiectis prius inesse possit. Verumtamen quoniam haec iam conspicua
demonstrati Theorematis patet), nihilominus Theorematis universalem sunt, ad alium Verneri defectum ostendendum accedamus. In vicesimo10
doctrinam nobis ostendit. Habet8 autem et illa prima pars duos Casus: aut itaque suo Elemento Conico, ubi Problema de quo sermonem habemus
enim latera illa quae ibi supponuntur aequalia longitudines rectangulorum demonstrabit, in demonstranda secunda eius parte, paralogismum quen-
sunt ambo, aut ambo latitudines. Nam alterum quidem longitudo, alterum dam commisit, quem etiam omnes alii, quos vidi huiusce rei Autores
vero latitudo esse non possunt, cum aggregata aequalia est debeant, ut admiserunt, praeter Hieronymum Cardanum, qui recte quo ad hoc con­
consideranti liquet. Cum itaque iam dictum Theorema universale sit, et cludit. Paralogismus autem talis est. Cum Vernerus (ut in primae nostrae
prima quidem eius pars duos suscipiat Casus, secunda vero tres, qui porro demonstrationis secunda figura) parallelam KO parallela PR maiorem esse
tres Casus eius, de quo sermonem habemus Problematis Constructioni ostendisset, statim concludens subiunxit: Ergo signum P propius est
accidere possunt: necessarium mihi visum fuit universe illud proponere rectae lineae MHproductae quam signum K (quamvis corrupte ibi legatur,
atque demonstrare, ut quod etiam prae manibus habemus Problema uni­ quam signum O ). Horum autem utrunque signorum K, P (et si ibi etiam
verse construi, demonstrarique posset. Maximum etenim in scientiis mendose legatur, O, R) existit in Hyperbolica sectione GID. Et quoniam
vitium est (ut docet Aristoteles)9 ea quae universe demonstrari possunt idem de omni alio / puncto quod in eadem obliqua linea Hyperbolicae
particulatim ostendere. Qui namque omne Aequilaterum, aut Aequicrure, sectionis GID extiterit eodem modo demonstrari poterit usque in infinitum:
aut Scalenum ostendit tres habere angulos duobus rectis aequales, non igitur quanto amplius recta linea MH et inflexa linea Hyperbolicae Sec­
demonstrat universe, etiam si in unaquaque specie hoc demonstraverit, tionis GID producantur, eo amplius appropinquant, quod Secundo demon­
sed qui omne Triangulum quatenus Triangulum est. Hunc igitur errorem strare oportuit. Haec sunt eius verba, in quibus concludit lineae Hyper­
Vernerus mihi perpessus esse videtur, cum Theorema illud particulatim bolicae signum P propius esse rectae MH lineae quam signum K , eo quod
proponat atque demonstret, credens tamen se universe demonstrare. Non KO recta linea maior est quam PR, quae quidem conclusio esset optima si
enim quadratum aggregati maioris longitudinis quadrato aggregati minoris KO et PR perpendiculares essent ipsi MR rectae lineae, tunc enim ipsae
essent minimae distantiae quibus signa K ,P a recta linea MR distare pos-
7 mg.: Quomodo Theorema illud nostrum in tribus casibus demonstretur.
* mg.: Duo primae partis casus qui sunt.
9 mg.: Lib. Posteriorum. 10 mg.: Secundus Verneri defectus.
418 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S F R O M B A R O Z Z F S AD M IRAN D U M PROBLEMA 419

sint. Quoniam autem perpendiculares non sunt (ut ibi ostendimus) atque vexam (quanvis male ibi legatur connexam) Coni superficiem, in qua pro­
propterea neque minimae distantiae, idcirco paralogismus in conclusione traho AC a vertice usque ad has in. Et sit K plana superficies contangens
committitur, quandoquidem a causa remota et extrinseca propositum con­ Conum in recta linea AC: quae superficies intelligatur in infinitum cum
cluditur. Nam causa maxime propinqua et immediata maioris appropin­ Coni Superficie extendi. Dico primo hanc superficiem planam non posse
quationis signi P ad rectam lineam MR quam signi K ad eandem est mini­ tangere Coni Superficiem alibi quam in linea AC: Quod si potest, tangat in
mae signi K distantiae maior longitudo minimae signi P distantiae longi­ G, et duco (licet ibi depravate legatur duo) Circulum aequidistantem per
tudine: non autem cuiusvis distantiae signi K a recta MR maior longitudo G basi BCD: (vel legatur melius, Circulum per G aequidistantem basi
cuiusvis distantiae signi P longitudine. Quando enim rem aliquam alicui BCD:) cum igitur Circulus sit in una superficie, erunt puncta contactus
propinquiorem alia quadam ostendere volumus, non dicimus illam minus plani K et peripheriae Circuli illius in una recta linea, ex demonstratis in
quam haec distare iuxta quaslibet earum distantias, sed iuxta minimas undecimo Elementorum Euclidis. Quamobrem cum illa linea iam tangat
quibus ambae ab ipsa tertia distare possint. Quamvis itaque parallelarum Circuli peripheriam in linea AC, cadet ex demonstratis ab Euclide in tertio
K O ,P R inaequalitas perpendicularium KT, PV inaequalitatis causa sit (ut Ele/mentorum extra circunferentiam Circuli VXG, igitur non tanget illum
in superioribus patuit), non ob id tamen hic manendum est, ex hacque in puncto G. Haec sunt verba Cardani, quae quantum obscura sint, et non
remota causa propositum concludendum: verum ulterius progrediendum Geometrice dicta, versatis in Geometria iudicandum relinquo. Hoc autem
quousque immediata reperiatur causa ex qua propositum recte concludi in primis animadvertam, quod haec omnia quae dicit Cardanus commode in
possit. Verae enim demonstrationes (quales Geometricae sunt) ex im­ nostrae secundae demonstrationis figura [Fig. 7.52] conspici possunt si per
mediatis causis fieri debent, ut Aristoteles docuit:11 Qui autem ex causis planum quidem K , nostrum ACEF planum intelligamus; per Circulum
remotis in Geometria demonstrationes conficiunt, sophistice quidem vero VXG, ipsum HGI apud nos Circulum. Cum itaque ita Conum et
demonstrant, paralogismosque committunt. Hec autem ad Vernerum dicta planum in linea AC eum tangens Cardanus construxisset, ut verba eius
sufficiant. explicant, volens in primis probare illud planum non posse tangere Coni
superficiem alibi quam in AC linea: incipit hoc indirecta demonstratione
180 / DIGRESSIO CONTRA ostendere, postea vero directe concludit supponens id quod a principio
CARDANUM. probandum suscepit. Ac demum haec eius demonstratio neque directa,
neque indirecta Geometrica, sed potius Chimerica mihi videtur. Nam
Hieronymus vero Cardanus Mediolanensis in libro sextodecimo de directa quidem demonstratio Geometrica directe semper arguendo et ea
Subtilitate Problema de quo loquimur demonstravit eo demonstrandi modo quae vera sunt supponendo, propositum ex eius causis concludere
cui nos in secunda nostra demonstratione maiorem perfectionem dona­ debet: indirecta vero Geometrica demonstratio supponens stantim
vimus. Quanvis autem Cardanus ibi dicat se velle uti demonstratione Rabbi a principio contrarium eius quod quaeritur, arguensque semper in­
Moysis Narbonensis exponentis dictum Rabbi Moysis Aegyptii nihilo­ directe iuxta secundum Hypotheticarum Aristotelis ratiocinationum mo­
minus Demonstratio Cardani a prima praecipuaque Rabbi Moysis demon­
dum, deducit tandem nos ad aliquod inconviens, quod suppositionem
stratione tantum differt, quantum nostra secunda demonstratio a prima
fuisse falsam indicat, eiusque contrarium, nempe Quaesitum, verum esse
discrepat. Nam prima praecipuaque Rabbi Moysis quidem demonstratio
demonstrat. At haec Cardani demonstratio supponit quidem mox a prin­
(ut inferius manifestum fiet) eadem quasi est cum nostra prima et cum
cipio contrarium eius quod quaeritur (cum dicat: Quod si potest tangat
Verneri demonstratione, Cardani vero demonstratio secundae nostrae
in G, etc.). Deinde directe semper arguens ad nullum deducit incommo­
demonstrationi, necnon ultimo ipsius Rabbi Moysis exemplo similis est.
dum, sed Quaesitum denique directe concludit illis verbis: Quamobrem
In1 suae itaque demonstrationis initio peccat Cardanus, quoniam volens
cum illa linea iam tangat Circuli peripheriam in linea AC, cadet ex dem­
probare (exempli gratia in secundae nostrae demonstrationis figura [Fig.
onstratis ab Euclide in tertio Elementorum extra circunferentiam Circuli
7.52]) quod planum ACEF non potest tangere Coni superficiem alibi quam VXG, igitur non tanget. Quae porro verba nullum absurdum continent,
in linea AC, petit principium, atque idem per idem probat. Ut autem quod
sed propositum directe concludunt. Ni2 forsan dicat aliquis ad hoc incon­
dicimus magis perspicuum fiat, audiamus eius verba. Inquit itaque Car­ veniens hanc demonstrationem deducere, quod eadem plana superficies
danus: Sit igitur Conus ABCD: nunc triangulum nullum (quanquam ibi eandem Conicam superficiem in eodem signo G prius tangere supponatur,
nullo perperam legatur) secantem intelligo, sed per ABD intelligo con- postea vero non tangere concludatur. Huic autem3 dictum volo, quod

11 mg.: Lib. Posteriorum.

2 mg. : Obiectio.
mg.: Cardani defectus primus. :i mg. : Responsio.
420 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S F R O M B A R O Z Z T S AD M IRAN DUM PROBLEMA 421

hoc admitteretur, si ilia ultima conclusio quae suppositioni oppugnat a igitur Circulus sit in una superficie, erunt puncta contactus plani K et
Quaesito diversa esset: quoniam autem eadem cum Quaesito est, non peripheriae Circuli illius in una recta linea, ex demonstratis in undecimo
182 possumus dicere ipsam esse incommodum ad quod deducitur. Nam / in­ Elementorum·, hic demonstrat unum quod ad concludendum Quaesitum
commodum ad quod omnis indirecta Geometrica demonstratio deducit maxime confert, quod scilicet puncta contactus plani K et circunferentiae
diversum a Quaesito semper esse debet: alioquin illa demonstratio ex ducti circuli (quae in nostra figura sunt puncta G, H ) sint in una recta
indirecta in directam transiret, nugatioque in eius principio fieret. Ex­ linea iacente tum in plano K , tum in plano ipsius circuli. Sed verba eius
empli4 gratia volens Geometra probare quod si trianguli duo anguli aequales obscure diminuteque hoc explicant, sic enim illa verba intelligi debent:
inter se fuerint, latera etiam quae sub aequalibus angulis subtendunt aequa­ Cum igitur Circulus sit in una superficie·, hoc est, cum circulus ille ductus
lia invicem erunt; atque non potens hoc commode per demonstrationem unum sit planum et ipsum K alterum planum, quae duo plana ex sup­
directam probare, per indirectam ostendit; et supponit quidem latera esse positione in ipsis contactus punctis se secant. Erunt puncta contactus
inaequalia, quod est Quaesito contrarium, ac demum ratiocinando deducit plani K et peripheriae, etc·, hoc est, erit communis eorum planorum sec­
ad hoc absurdum quod pars sit aequalis toti, quod quidem absurdum tio una recta linea transiens per illa contactus puncta per tertiam pro­
idem cum Quaesito non est, sed longe diversum, ut patet; neque sup­ positionem libri undecimi Elementorum Euclidis. Talis meo quidem iudicio
positioni illi oppugnat quae statim in principio Quaesito contraria posita debet esse verborum illorum sensus. Hucusque autem bene procedit in­
fuit (quoniam idem cum Quaesito esset, duo enim eidem contraria esse directa demonstratio, ut ex nostra secunda demonstratione coniici potest.
non possunt) sed illi communi sententiae adversatur quae ait, omne totum Postea vero subiungit: Quamobrem cum illa linea iam tangat circuli pe-
est maius sua parte. Quemadmodum igitur in hac indirecta demonstra­ ripheriam in linea AC, cadet ex demonstratis ab Euclide in tertio Ele­
tione inconveniens ad quod deducitur a Quaesito diversum est, sic etiam mentorum extra circunferentiam Circuli VXG, igitur non tanget illum in
in omnibus aliis demonstrationibus indirectis esse debet: alioquin id quod puncto G. Haec sunt illa verba quae totam hanc demonstratiunculam
diximus sequeretur. Si nanque in iam dicta demonstratione ad contrarium evertunt atque corrumpunt, quaeque petitionem principii continent.
primae suppositioni incommodum (ut fecit Cardanus) deduceretur quod Eorum enim talis est sententia: Cum ostensum quidem sit unam esse
idem cum Quaesito est, nempe latera sub aequalibus angulis subtendentia rectam lineam communem istorum planorum sectionem quae transit per
invicem aequalia esse, nonne haec potius esset directa Quaesiti demon­ illa duo puncta in quibus planum K tangit conicam superficiem: illa autem
stratio? quid igitur opus esset a principio contrarium Quaesiti supponere, recta linea iam tangat circuli circunferentiam in linea AC, cadet per deci-
si directa demonstratione illud concludi posset? nonne manifesta com­ mamoctavam et sextamdecimam propositionem tertii libri Elem. Eucl.
mitteretur nugatio? Quod itaque nullo pacto huiuscemodi demonstratio extra circunferentiam ipsius circuli. Quamobrem neque ipsa recta linea
in Geometria fieri possit, bonis Geometris perspicuum est. Quanvis5 ipse neque planum K , in quo ipsa est, tanget circuli circunferentiam in signo
Cardanus in sua Logica (quam manuscriptam ipse nobis ostendit) dicat G; atque idcirco neque etiam Conicam superficiem in qua iacet circuli
hunc esse quendam pulcrum demonstrandi modum, appellarique Chrysip­ circunferentia in ipso G signo tangere potest. Haec itaque est perfecta
peum seu Cornutum: sed ipse viderit an chimericus potius quam Chrysip­ verborum illorum sententia, quae directe Quaesitum videtur concludere et
peus sit. Quod vero Cardanus talem faciat demonstrationem, quae etiam chimeri/cam illam (quam superius diximus) demonstrationem conficere.
fuit causa ut committeret petitionem principii, ex eius verbis manifestum Fortasse autem directa haec Quaesiti conclusio bona esset auferendo illam
est. Ait enim: Dico primo hanc superficiem planam non posse tangere Quaesito contrariam suppositionem in principio positam, ni (quod peius
Coni superficiem alibi quam in linea AC: hoc est illud quod probandum est) in illis ultimis verbis principium peteret. Verum iccirco nullo modo
proponit, quod scilicet superficies plana apud ipsum nominata K , in nostra admittenda est. Cum enim dicit rectam illam lineam iam tangere circuli
vero figura ACEF, non potest tangere superficiem conicam alibi quam circunferentiam in linea AC, tunc nimirum petit principium, quoniam
in linea AC. Quod si potest, tangant in G nunc aggreditur indirectam nequaquam illucusque probavit lineam illam tangere circuli circunferen­
183 probationem, et / statim a principio supponit Quaesito contrarium, quod tiam, quod nihilominus tanquam iam probatum inferre videtur. Idem
scilicet plana illa superficies non solum in linea AC tangat conicam super­ autem hoc est ac si supponeret planum K non tangere Coni superficiem
ficiem, ut Quaesitum dicebat, sed etiam in alio ipsam tangat signo extra alibi quam in ipsa AC linea, quod utique probandum ab initio sibi propo­
lineam AC iacenti, ut, causa exempli, in signo G. Et duco circulum PG suerat. Quod vero idem sit lineam illam in plano K iacentem tangere
aequidistantem basi BCD; hoc construit ut demonstrationi deserviat. Cum circulum in linea AC, ac si dicamus planum K non tangere Coni super­
ficiem alibi quam in linea AC, non est cognitu difficile. Si enim recta
4 mg. : Exemplum. linea in plano K iacens tangat in linea AC circulum in superficie conica
5 mg. : Logica Cardani non extat impressa. descriptum, necessario rationibus superius dictis tum ipsa linea, tum
422 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S F R O M B A R O Z Z F S AD M IRAN D U M PROBLEMA 423

planum K , in quo ipsa iacet, extra circuli circunferentiam cadent; nec quamvis ab ipsius plani K cum superficie conica contactu exeant, nihilo­
tangent conicam superficiem nisi in linea AC, in qua ipsam tangere sup­ minus possunt quandoque tangere circulum, quandoque vero secare, se­
ponuntur. At6 si quis forte dicat, cum supponatur planum K tangere coni­ cundo scilicet modo secundi modi secandi: non tangent autem nisi unico
cam superficiem in lineaAC, necnon circulorum omnium basi parallelorum tantum eo modo, quem diximus. Cum11 autem tribus his modis recta linea
in ipsa descriptorum circunferentias; ex hoc sequi rectas etiam lineas ab secare possit circulum, illam tantum proprie Geometrae circulum secare
ipsa AC recta linea in plano K ductas tangere tum Coni superficiem, tum dicunt quae primo secandi modo secat. Ita igitur probanda est haec con­
circulorum in ea sic descriptorum circunferentias: huic respondeo,7 quod sequentia, quemadmodum nos eam in secunda nostra demonstratione pro­
licet planum K tangat Coni superficiem et circulorum in ea descriptorum bavimus; aliter paralogismus committitur, quandoquidem ex contactu ip­
circunferentias in linea AC, non ob id tamen necessarium est ut rectae sius lineae et ex propositione tricesimasexta tertii lib. Elem. Eucl. infertur
etiam omnes lineae ab ipsa AC in plano K deductae tangant tum Coni paulo inferius in ipsa Cardani demonstratione (ut etiam in nostra secunda
superficiem, tum dictorum circulorum circunferentias: nam Coni quidem 187 / demonstratione conspicere licet) quaedam consequentia maxime neces­
superficiem necessario semper tangent, circumferentias vero circulorum saria ad concludendum Quaesitum, quippe quae non verificatur nisi in
quandoque tangere, quandoque etiam secare possunt, ut manifestum lineis unico tantum illo modo uti diximus Circulum tangentibus. Hoc autem
est. Haec igitur dicta sint ad Cardani falsam principiumque petentem animadversione dignum esse censui, quoniam Cardanus hic, quemadmodum
demonstratiunculam, quam nos in secunda nostra demonstratione cum etiam plerique alii nesciunt distinguere quaenam apud Geometras rectae
indirecte, tum directe instauravimus. lineae Circulum tangentes, quaeque Circulum secantes esse debeant; atque
Rursus8 autem paulo inferius paralogismum committit Cardanus his definitionem secundam tertii eiusdem Elem. penitus non intelligunt, ut ibi
verbis. Et ducantur recte LT et OM in superficie K (quamvis ibi cor­ nos in Commentariis nostris in Euclidem plenius adnotavimus. Haec etiam
rupte, H legatur) quae contingent circulos QLP et XOV, quia ducuntur ad secundum Cardani defectum dicta sufficiant.
185 ex loco contactus. In figura secundae nostrae / demonstrationis intelli- Extat12 autem tertius quoque defectus in ipsa Cardani demonstratione
gantur lineae quidem LT et OM esse ipsae N S, HR: circuli vero QLP et prope finem, ubi habet haec verba: Sed MN maior est TR (quia si du­
XO V , ipsi NM O , H G I. Sententia igitur horum verborum talis est: et ducan­ ceretur per N superficies aequidistans, ipsa [quanvis mendose legatur,
tur (ut in nostra iam dicta figura) rectae lineae NS et HR, quae con­ ipsum] caderet infra R , sed melius legatur extra R , aliter occurreret K , quia
tingent circulos NMO et HGI quia ducuntur ex loco contactus ubi scilicet Diameter QP est minor XV, et superficies Circulorum sunt aequidistantes)
planum ACEF tangit circunferentias circulorum NMO et HGI. Volens igitur ST maior est GM. Verba haec primo obscure admodum atque con­
itaque Cardanus probare id quod superius supponebat dum petebat princi­ cise dicta sunt; secundo nihil concludunt. Quare primum eorum sententiam
pium, nempe lineas NS et HR tangere circulorum illorum circunferentias, explicabimus. Deinde quod nil concludant ostendemus. Omnia autem quae
dicit quod tangunt quia ex loco contactus ducuntur. Mihi autem male a nobis dicentur in figura secundae nostrae demonstrationis inspiciantur;
videtur deduci haec consequentia: ex loco contactus ducuntur, ergo tangunt accipiendo scilicet planum K ipsius Cardani pro nostro ACEF plano, et
circulorum circunferentias, non est enim necessarium ut ipsas tangant, lineam MN ipsius pro linea RP nostra, et punctum N eius pro puncto
sed possunt etiam eas secare, utputa si ad lineam AC perpendiculares P nostro, et lineam TR ipsius pro SQ nostra, et punctum R eius pro
non fuerint, vel si etiam productae Conum secuerint. Nam rectae lineae puncto Q nostro, et lineas ST, GM ipsius pro lineis MS, GR nostris, et
circulos tangentes, de quibus hic Cardanus loquitur, quas etiam Euclides Circulum QP ipsius pro Circulo NO nostro, et Circulum VX eius pro
definit in tertio libro Element, definitione secunda, illae sunt quae in eodem Circulo/// nostro. His declaratis, dico quod verborum Cardani tale sensum
186 plano cum circulo iacentes cum circulum tangant, si / producantur cir­ est. Cum ostendisset ipse lineam MS (et loquor nunc in nostra figura)
culum non secant. Animadvertendum9 autem est quod recta linea dupli­ eam habere rationem ad lineam GR quam habet linea RP ad lineam SQ,
citer circulum secare potest: vel10 quando eius planum secat in eodem volens probare RP esse maiorem ipsa SQ, ut ex hoc concluderet etiam
et ipsa plano iacens, vel quando eius planum secat in alio ipsa existens ipsam MS maiorem esse quam GR, probat hoc illis verbis. Sed MN maior
plano; hocque dupliciter, vel intra circunferentiam, vel in circunferentia, est TR (quia si duceretur per N superficies, etc.), quae verba in nostra
quemadmodum in praesentia lineae, de quibus Cardanus sermonem habet, figura sic legantur: Sed RP maior est SQ (quia si duceretur per punctum
P planum Parallelum plano ACEF, ipsum utique ductum planum caderet
6 mg.: Dubitatio.
extra punctum Q: aliter occurreret ipsi ACEF plano, quia dimetiens Cir­
7 mg. : Solutio. culi NO est minor dimetiente Circuli HI, et plana Circulorum ipsorum
8 mg.: Cardani defectus secundus.
9 mg. : Notandum. 11 mg.: Tribus modis Recta linea circulum secare potest.
10 mg.: Quot modis Recta linea circulum secare possit. 12 mg.: Tertius Cardani defectus. The brackets in this sentence are Barozzi’s.
424 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S FR O M B A R O Z Z F S AD M IRAN D U M PROBLEMA 425

188 sunt parallela) igitur / MS maior est quam GR. Sic legi debent Cardani caderet in extremitate vel intra extremitatem dimetientis circuli minoris,
verba in nostra figura, quorum (si ita legantur) clara erit sententia. Inquit occurreret lineae AC. Nam si per terminos duarum linearum parallelarum
enim quod RP maior est quam SQ, hac scilicet ratione, quia si duceretur et inaequalium rectae producantur lineae: illas ad partem minoris parallelae
per punctum P planum unum parallelum plano ACEF, ipsum necessario 190 concurrere est necesse. Quod Theorema verissimum est, et / ab Euclidis
caderet extra punctum Q; cadens autem extra punctum Q, secabit se extra Elementis dependet, demonstraturque a Vitellione in sextadecima pro­
Conum cum EFKL plano ipsius Hyperboles, et communis eorum sectio positione primi libri suae Perspectivae. Cum igitur lineae illae in vertice
erit per tertiam propo. undecimi lib. Elemen. Eucl. recta linea ipsi FE Coni sibi coinciderent, necessario plana quoque illa duo in quibus ipsae
paralella (/) per sextamdecimam propositionem eiusdem, cui quidem paral­ sunt ibidem sese tangerent. Hoc itaque pacto ratio Cardani concluderet
lelae occurret ipsa SQ si extra Conum producatur, quoniam SQ, et RP quod scilicet plana illa sibi occurrerent propterea quod circulorum di­
parallelae sunt per eandem sextamdecimam cum ex suppositione in paral­ metientes inaequales et parallelae sunt, quamvis hoc ad rem non esset.
lelis sint planis: quae autem cum una parallelarum coincidit, cum altera At quoniam planum illud non ducitur per extremitatem dimetientis cir­
etiam coincidet si in infinitum producantur per ultimam definitionem et culi maioris sed per punctum P, idcirco ratio non concludit. Dimetientium
tricesimam propositionem 1. lib. Elemen. Eucl. Quare cum iam dicta enim inaequalitas et parallela positio non est causa quod planum ductum
parallela occurrat ipsi SQ extra Conum productae, fiet unque unum paral- per punctum P plano ACEF parallelum eidem ACEF plano occurrat
189 lelogrammum cuius latera erunt ipsa parallela et SQ produ/cta, et SR et (quandoquidem planum ductum per punctum P potest ita produci versus
RP. Cum autem parallelogrammorum latera opposita sint aequalia per 34 Coni Verticem, ut non tangat etiam ipsas dimetientes; tamen si in signo Q vel
proposit. eiusd. 1. Elemen. ergo SQ producta tota erit aequalis ipsi RP. intra ipsum caderet cum plano ACEF concurrat). Sed vera huius concursus
Igitur per nonam Com. Sent, eiusdem primi pars ipsius productae, ipsa 191 causa est linearum RP et SQ inaequalitas et paral/lela positio. Quod13 si dicat
videlicet SQ, erit minor quam RP. Hoc modo probat Cardanus lineam aliquis ex inaequalitate dimetientium dependere linearum RP, SQ inaequali­
RP esse maiorem quam SQ. Quod autem planum ipsum quod per punctum tatem atque iccirco Cardanum ex inaequalitate dimetientium arguere: huic
P ducitur parallelum ipsi ACEF plano necessarie cadat extra punctum Q, respondeo14 quod tunc fieret Petitio principii, quoniam quod probare vult
sic probat Cardanus. Si extra punctum Q non caderet, sed in ipso Q, vel Cardanus est inaequalitas linearum RP, SQ quam si immediate vel per aliud
intra ipsum, occurreret ipsi ACEF plano, non potest autem ipsi occurrere medium magis remotum supponat, dubio procul principium petit, idemque
cum parallelum ipsi ducatur (parallela enim plana sunt per octavam de­ probat per idem. Talis enim esset eius argumentatio. Dico quod linea
finit. 11 lib. Elem. Eucl. quae sibi non coincidunt). Ergo patet quod extra RP est maior SQ, quia si duceretur per signum P planum parallelum
cadet. Quod vero iam dictum planum plano ACEF occurreret si caderet plano ACEF, caderet extra signum P; aliter sequeretur hoc absurdum
in puncto Q, vel intra ipsum, hinc probat: quia dimetiens Circuli NO est quod dictum planum occurreret eidem plano ACEF, cui parallelum ducitur,
minor dimetiente Circuli HI, et plana eorundem Circulorum parallela sunt. quia dimetiens circuli NO est minor dimetiente circuli///, et plana ipsorum
Hoc itaque est verborum sensum. Verum probatio haec maxime peccat, circulorum parallela sunt. Hoc est quia linea RP est maior quam SQ,
quoniam nil concludit, quod quidem facile cognoscitur. Quando enim in qua quidem probatione nil vitiosius. Nullo modo igitur haec Cardani dem­
ultima probationis particula probat planum quod per punctum P ducitur onstratio servari potest, quoniam vel nihil concludit vel principium petit,
plano ACEF parallelum occurrere ipsi ACEF plano si cadat in puncto ut iam ostendimus. Quapropter non est eo modo probandum lineam RP
Q, vel intra ipsum, quia dimetiens Circuli NO est minor dimetiente Cir­ esse maiorem linea SQ quo Cardanus id probare conatur: sed ita ut nos
culi HI, et plana ipsorum Circulorum sunt parallela: tunc nil mihi con­ in secunda nostra demonstratione demonstravimus. Haec itaque de tribus
cludere videtur, nam si planum quod ducitur plano ACEF parallelum quoque insignioribus demonstrationis Cardani deliquiis dicta sint.
duceretur per extremum dimetientis Circuli maioris, caderetque in ex­
tremo dimetientis Circuli minoris, vel intra eius extremum, tunc sequeretur
/ DIGRESSIO CONTRA
quod necessario planum illud occurreret ACEF plano, propterea quod
ORONTIUM.
dimetiens Circuli NO est minor dimetiente Circuli HI, et plana eorundem
Circulorum sunt invicem parallela. Nam dimetientes ipsae cum sint in Orontius autem Finaeus in libello suo de Speculo ustorio Problema
planis parallelis et in plano trianguli per axem Coni (ut patet per quartam nobis propositum demonstravit eo demonstrandi modo quem nos in tertia
petitionem huius), erunt per sextamdecimam propositionem undecimi libri nostra demonstratione sub perfectione universalique doctrina redegimus.
Elementorum Euclidis invicem parallelae. Cum autem sint etiam inaequa­
les, hinc necessario sequeretur quod communis sectio plani trianguli per 13 m g Obiectio.
axem Coni et ipsius plani quod duceretur plano ACEF parallelum, si 14 mg.: Responsio.
426 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S F R O M B A R O Z Z I S AD M IRAN DUM PROBLEMA 427

Nam eius quidem demonstratio cum maximas habet imperfectiones, tum partem illam Problematis quae ait lineas ipsas in infinitum productas nun­
particulatim in Cono tantum rectangulo propositum demonstrat: cum ta­ quam sibi coincidere; probat eam his verbis: Quanto magis igitur AD
men in omni Cono verum sit. Demonstrat autem Orontius praesens Prob­ et FG lineae in continuum producentur ad partes quidem D et G, tanto
lema non solum de duabus lineis, recta et Hyperbolica, in eodem plano propiores evadent: et nihilominus eas tandem convenire est impossibile,
iacentibus, verum etiam de duabus dictis lineis non in eodem plano, sed in utpote, quae in planis consistunt invicem parallelis, ex ipsa Constructione
una Coni superficie existentibus, quod nos quoque ad calcem tertiae nos­ et semper tantum ad minus invicem distabunt quanta est linea recta
trae demonstrationis subiunximus. In his itaque duabus suis demonstra­ utrique praedictarum superficierum perpendicularis. Utraque igitur pro­
tionibus praeter multa deliquia, et infinitas constructionum consequen­ positionis pars verissima relinquitur. Haec sunt Orontii verba, quae per
tiarumque rationes omissas (ut ex nostra tertia demonstratione quisque se clara sunt; sed maximam falsitatem continent. Non est enim verum
coniicere potest), quae fortasse tolerari possunt, tres potissimum errores quod semper tantum ad minus illae duae lineae invicem distabunt, quanta
commisit, qui nullo modo tolerandi sunt. Primus error1 talis est. Volens est linea recta utrique praedictarum planarum superficierum perpendicu­
Orontius probare lineam NR (et loquor in figura nostrae tertiae demon­ laris, ut ait Orontius; quoniam recta linea utrique praedictorum planorum
strationis [Fig. 7.55]) maiorem esse linea PS, ex quo porro tota illa de­ perpendicularis nullo modo potest esse ipsarum linearum distantia, cum non
pendet demonstratio. Probat illud ex eo, quod aequales rectae lineae ab ambas sed alteram tantum earum tangat. Quo nam pacto igitur duae illae
inaequalibus circulis inaequales secant circunferentias, minorem quidem a lineae semper tantum ad minus invicem distabunt, quanta est linea recta
maiori, maiorem vero a minori. Quod autem hoc suum Theorema verum utrique duorum planorum in quibus illae lineae sunt perpendicularis cum
sit, sic ille confirmat. Quoniam (inquit) plus incurvatur minor quam ipse ipsa perpendicularis ipsarum linearum distantia esse minime possit? Ex­
maior circulus. Quantum autem ratio haec debilis inanisque sit hinc in­ empli gratia in figura Tertiae nostrae Demonstrationis [Fig. 7.55], quomodo
tuendum est: quoniam nos in superioribus ostendimus quod aequales rec­ 194 linea FG inflexa a recta AD tantum ad minus distabit, quan/ta est recta
tae lineae ab inaequalibus circulis inaequales auferunt circunferentias non linea MS utrique, tum Hyperbolis, tum per axem trianguli plano perpen­
solum minorem a maiori et maiorem a minori, sed etiam maiorem a maiori dicularis, cum tamen ipsa MS ipsam AD non tangat, ideoque earum line­
et minorem a minori circulo, in minoribus scilicet atque maioribus cir­ arum distantia esse non possit? quaelibet enim duarum linearum dis­
culorum segmentis; et tamen circulus minor semper plus incurvatur quam tantia terminis suis ambas ipsas attingere debet. Quod autem recta linea
maior. Ea igitur ratio nulla est. Quare Theorema Orontii non universe utrique illorum planorum perpendicularis non possit nisi alteram ipsarum
verum est atque indemonstratum remanet, cum eius ratio non concludat duarum linearum attingere, sic in eadem nostra figura probetur. Si ab aliquo
et consequenter lineam NR linea PS maiorem esse indemonstratum re- pun[cto] lineae AD erigatur recta linea ad rectos angulos plano ADE, ut
193 linquitur. Cum autem ex / hoc tota dependeat Problematis demonstratio, ipsaAT, erit per tertiam definitionem undecimi lib. Elem. Eucl. ad rectos
manifestum est Orontium potius nugari quam demonstrare. Nos vero ut 195 angulos dimetienti circuli conicae basi paralleli per pun/ctum illud transi-
demonstrationem hanc exacte traderemus, prius Theorema hoc ad uni­ entis. Unde per 16 prop, tertii lib. eorundem extra ipsum LNO circulum
versalem doctrinam redegimus, universeque demonstravimus; ex cuius cadet, ideoque non tanget inflexam FG lineam (si enim tangeret, intra
demonstratione Corollarium manifestum excerpsimus, lineam scilicet NR circulum caderet per secundam propositionem eiusdem tertii) et nihilo-
esse maiorem PS. Nam harum linearum inaequalitas potius est causa in­ secius utrique ipsorum planorum perpendicularis est, quoniam duobus
aequalitatis circunferentiarum quam e converso. Quandoquidem in dem­ planis parallelis existentibus quae alteri eorum perpendicularis est; reliquo
onstranda circunferentiarum inaequalitate prius apparet nobis harum rec­ etiam perpendicularis erit, ut patet ex 16 prop, undecimi et 29 primi et
tarum linearum inaequalitas, ex qua postea circunferentiarum quoque tertia definitione undecimi lib. Elemen. Eucl. utque probat Vitellio in 23
inaequalitas concluditur, ut in nostra illius Theorematis demonstratione prop, sui primi libri. Idem autem ostendetur de omnibus aliis lineis a
quisque conspicere potest. Is itaque sit primus, insignisque Orontii error. recta AD plano ADE et plano Hyperboles perpendicularibus ductis. Prae­
Secundus2 autem est idemmet cum secundo Verneri, quem errorem om­ terea si ab aliquo inflexae FG lineae puncto erigatur ad rectos angulos
nes, quos vidi huius rei Autores commiserunt, praeter Cardanum. Ter­ plano Hyperbolis recta linea quae (ratione iam dicta) plano etiam ADE
tius3 vero est huiusmodi. In fine suae primae demonstrationis, in qua perpendicularis sit, quemadmodum ipsa LR necessario cadet intra LNO
demonstrat Problema propositum de duabus lineis, recta et inflexa, in una circulum, quoniam per quintam petitio, primi lib. Element. Eucl. lineae
Coni superficie et non in eodem plano consistentibus: volens probare NR in partes R occurrere debet, cum anguli RLN et RNL duobus rectis
minores sint, anguli enim RLT et RNT per Constr. et per tertiam defin.
1 mg. : Orontii primus error. 11 lib. eorundem Elem. recti sunt. Cum igitur linea LR utrique plano
2 mg. : Orontii secundus error. perpendicularis intra circulum cadat, manifestum est quod non contingit
3 mg.: Orontii tertius error. lineam AD. Similiter autem ostendi potest quod omnes aliae lineae quae
428 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S FR O M B A R O Z Z T S AD M IRAN DUM PROBLEMA 429

a linea FG inflexa plano Hyperbolico et plano ADE triangulari perpen­ causam dicit sed quandam potius probabilem imaginationem. Quod nanque
diculares ductae sunt ipsam AD non tangent; quod erat probandum. Cum Circuli versus Coni basin maiores semper fiant, latusque Hyperbolis per­
itaque verum sit quod rectae lineae utrique illorum duorum planorum petuo et gradatim versus basim ascendat verum et necessarium est. Non ex
perpendiculares non possunt nisi alteram ipsarum AD, FG linearum at­ hoc tamen linearum continua appropinquatio necessario sequitur. Quid
tingere, perspicuum est quod neque etiam ipsarum distantiae esse pos­ enim si quis rectam illam lineam quae a centro Hyperboles oritur, quam­
sunt. Quare dictum Orontii falsum omnino relinquitur, quod nempe duae que in superioribus affectioni iam dictae succumbere demonstravimus,
illae AD et FG lineae semper tantum ad minus invicem distabunt, quanta non debitis conditionibus protrahat? utrum continue lateri Hyperbolis per­
est linea recta utrique praedictorum planorum perpendicularis. Quum enim petuo et gradatim versus Coni basim ascendenti appropinquet, et nunquam
lineae LR et NT sint utrique plano perpendiculares, nullae aliae ab eisdem coincidat? nonne posset etiam ita protrahi ut quantum inflexa illa linea
L, N punctis duci possunt lineae quae eisdem planis perpendiculares sint continne (!) ascendit, tantum haec, scilicet recta illam effugiat ab eaque
per 13 prop, undecimi libri eorundem Elemen. At LR et NT omnesque recedat? non est igitur causa huiusce affectionis necessaria illa quam dicit
eiusmodi lineae utrique plano perpendiculares nullo modo linearum AD et Peletarius; sed contingens quaedam ac probabilis coniectura, quam alienae
FG distantiae esse possunt; non distant igitur semper tantum ad minus vero huiuscemodi rationes a Geometria sint, audiamus Aristotelem as­
invicem, quanta est linea recta utrique plano perpendicularis: imo semper serentem2, simile esse a Rhetorico demonstrationes exigere, et Mathe­
maiori adinvicem intervallo distabunt quam sit linea recta utrique ipsorum matico probabiliter disputanti assentiri: necnon in Platone Simmiam
planorum perpendicularis. Si enim a puncto L ad rectam AD lineam una dicentem. Quoniam3 ex appa/rentibus demonstrantes vanos esse scio.
196 perpendicularis ducatur, erit per 19 pro/positionem primi libri Elemen. Multum itaque deficit Peletarius in iam dicti membri demonstratione. At in
Eucl. maior quam LR utrique plano perpendicularis, quoniam per tertiam reliquo membro demonstrando maximum etiam passus est deliquium.
defin. undecimi lib. eorundem angulum rectum subtendit, eum scilicet qui Probat4 enim illud eo modo quo Cardanus. Cum autem ostendere velit
ad R signum est. Idem etiam verum erit de omnibus aliis lineis ab FG linea planum in quo est illa recta linea qua Hyperboli semper magis magisque an­
inflexa ad rectam AD perpendicularibus ductis. At ipsae perpendiculares nuere et nunquam illi occurrere debet non tangere Coni superficiem nisi in
(ut in superioribus dictum est) minimae sunt distantiae ipsarum FG, AD una recta linea iam dictae rectae lineae parallela: tali utitur ratione. Intel-
linearum; plus ergo semper invicem distabunt ipsae AD , FG lineae quam ligatur (inquit) plana superficies super Cono iacens secundum longitudinem,
sit recta linea utrique duorum iam dictorum planorum perpendicularis. quae quidem superficies unica sui linea Conum tanget; constat enim ipsa
Quamobrem patet etiam tertius Orontii error. Haec autem ad Orontium infinitis lineis rectis et Conus infinitis circulis: linea vero recta sive secet
quoque dicta sint. circulum sive tangat, eum in unico puncto secat et tangit. Haec est ratio
Peletarii, per quam demonstrare credit planum illud nullibi tangere Coni
197 / DIGRESSIO CONTRA superficiem quam in illa recta linea. Quod autem ratio haec nil concludat,
PELETARIUM. sic ostendemus. Primo quidem cum dicit: Linea vero recta sive secet cir­
culum sive tangat, eum in unico puncto secat et tangit, falsum dicit. Nam
Jacobus vero Peletarius in Commentario suo de Contactu linearum rem nulla recta linea circulum secans in uno tantum puncto eum secare potest.
hanc de qua laboramus tetigit. Et primum quidem iuxta Priscorum viam Si enim in circuli circunferentia duo quaelibet puncta suscepta fuerint,
rem ipsam breviter demonstrare tentavit, et usus est modo demonstrandi recta linea quae ipsa puncta coniungit tota intra circulum cadit per se­
Cardani, sed breviter quidem ac imperfecte. Nam1 alterum Problematis cundam propositionem tertii libri Elementorum Euclidis: talisque linea
membrum, nempe linearum continuam appropinquationem, quodam ap­ iuxta Euclidis doctrinam dicitur secare circulum; quae si etiam in infinitum
parenti signo, debilique potius coniectura quam Geometrica ratione vel ex altera vel ex utraque parte producatur, semper secans circulum
ostendit. Cum enim (inquit ipse) latus Hyperbolis perpetuo et gradatim vocabitur. Passim enim Euclides, cum de recta linea circulum secante
ascendat versus Coni basin, propterea quod Circuli quibus Conus ipse facit mentionem, semper de illa recta linea intelligit quae intra circulum ab
constat continue maiores fiunt versus basim ipsam; recta autem linea ad uno circunferentiae puncto ad aliud transit. Exempli gratia in tricesimase-
quam debet latus Hyperboles approximari in suo plano immutabilis maneat: cunda propositione tertii libri eorundem Elementorum cum dicit: Si
procul dubio (inquit) ipsa inflexa linea secundum augumentum (!) Cir­ Circulum tetigerit aliqua recta linea, a contactu autem ad Circulum per­
culorum continue etiam propior fiet ipsi rectae lineae. Hoc pacto membrum ducatur quaedam recta linea Circulum secans: anguli quos ad contingen-
hoc demonstrat Peletarius. Quantum autem ratio haec imbecilla sit, a Geo­
metricisque rationibus aliena, nemo est qui non videat. Nullam enim rei 2 mg.: Aristoteles 1. Ethic, c. 3.
mg.: Plato in Phaedonem.
mg. : Primus Peletarii error. 4 mg.: Secundus Peletarii error.
430 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S F R O M B A R O Z Z I S AD M IRANDUM PROBLEMA 431

tern facit aequales erunt iis qui in alternis Circuli segmentis existunt an- scilicet Hyperboles. Praeterea7 Geminus antiquissimus in Geometria,
199 gulis. Falsum igitur dicit Peletarius, cum asserat rectam omnique laude dignus scriptor (referente8 Proclo in libro secundo Com­
200 lineam in unico puncto Circulum secare. Quod si etiam mentariorum in primum librum Euclidis Elementorum Commentario
secans circunferentiam vel Circulum improprie talis recta linea vocari quarto) lineas quidem bifariam dividebat, in simplices nempe et mistas:
admittatur, dico quod ratio Peletarii nil concludit. Intelligatur enim et simplices quidem, in rectas et circulares; mistas vero in planas et solidas:
(quemadmodum ait ipse) planum super Conicam superficiem iacens et planas quidem in sibi coincidentes, ut Cyssoides et in eas quae in infinitum
secundum longitudinem. Si itaque planum hoc unica sui linea tangere producuntur, ut Helices: solidas vero in eas quae circa Solida sine ulla
Conum ea ratione probetur, quia scilicet planum ipsum infinitas in se ipsorum Solidorum sectione describuntur, quales sunt Helices circa
recipere possit rectas lineas, et Conica superficies infinitos parallelos Cir­ Sphaeram, vel Conum, vel Cylindrum descriptae; et in eas quae ex Soli­
culos, recta vero linea sive secet Circulum sive tangat, eum in unico puncto dorum sectione oriuntur, ut tres Conicae Sectiones, Parabole scilicet,
secat et tangit: dico quod nihil concluditur. Nam omnes quidem rectae Hyperbole, et Ellipsis. Ecce quam dilucide Geminus et Proclus declarant
lineae quae Circulos illos tangunt in unico puncto tangent, atque extra illius lineae naturam esse non rectam, neque circularem, sed ex his mistam.
Conum totae cadent per sexta[m]decimam propositionem tertii libri Ele­ Aristoteles9 etiam in primo de Coelo tres ait esse motus rectum, circu­
mentorum Euclidis et eius corollarium. Non omnes autem rectae lineae larem, et tertium ex his mistum; quoniam (inquit) tres sunt etiam lineae,
quae Circulos illos secant, quanvis in unico etiam puncto secent, totae extra recta, et circularis, et ex his mista. Cum autem linea illa de qua loquimur
Conum cadent: sed possunt esse quaedam rectae lineae quae in unico neque recta, neque circularis sit, sed Hyperbole, necessario ex genere
puncto Circulos illos eo modo improprie secent, et tamen totae intra conum mistarum erit. Ioannes10 etiam Vernerus Problema de quo agimus sic
cadent; et planum in quibus illae sunt Conum secabit. Non concludit proposuit: Duas producere lineas alteram rectam, alteram inflexam, quae
igitur ratio haec: rectae lineae Circulum in Coni superficie iacentem tan­ Hyperbole Coni Sectio est, quae quanto amplius producuntur, eo magis
gentes, vel secantes in unico tangunt vel secant puncto; ergo planum in quo vicissim appropinquant, nunquam coincidentes, etiam si in infinitum
sunt illae rectae lineae unica sui linea Conum tanget. Potest enim etiam non producantur. Similiter Hieronymus Cardanus11 sic Problema proponit:
tangere sed secare, ut iam diximus. Aliter autem hoc probandum est, Duas invenire lineas in eodem plano quarum altera erit recta, reliqua
quemadmodum nos tum directe, tum indirecte in secunda nostra demon­ latus Hyperboles, quae semper sibi invicem magis approximabuntur et
stratione, id demonstravimus. Duos itaque iam dictos commisit errores nunquam se tangent. Omnes itaque tam Prisci quam Recentiores Autores
Peletarius in demonstrando proposito Problemate breviter, atque concise lineam hanc Hyperbolem ac propterea ex genere mistarum esse explicant,
iuxta Antiquorum demonstrationem. Cum autem hoc modo rem ipsam praeter12 Calcagninum, qui revera deceptus est; quoniam in Epistola quam
demonstrasset; subiungi quod ea inventio Antiquorum est acuta quidem, et scripsit ad Iacobum Zienglerum ambas iam dictas lineas rectas esse ait.
Geometrica prorsus: sed quae non explicet ex quo genere linearum sit ea Male igitur Peletarius Antiquorum inventionem mentemque percepit, cum
quae in eodem plano rectae lineae semper magis ac magis appropinquat, 202 dicat eos genus illius lineae non explicasse. Verum in / duabus illis demon­
nunquam ipsi coincidens. Non nullis enim (inquit ipse) videri possit recta, strationibus, in quibus ipse ortum, rationem, et naturam illius lineae ob
propterea quod rectissime procedere videtur in superficie Coni, quod oculos ponere inquit, in multos etiam insigniores errores ellapsus est, quos
Caelius Calcagninus putavit. Quamobrem (inquit) locus postulat ut lineae ut facile possimus ostendere verba eius in medium adducemus, quae sunt
illius ortum, rationem, naturamque ob oculos ponam. Subiunxitque huiusmodi . . .
demum duas illas imperfectas demonstrationes, quas nos superius perfec­
tione donavimus; per easque cum propositum Problema duobus modis [Extract VII, pp. 209-49]
demonstrare, tum dictae lineae naturam manifestare voluit. In5 hac autem / LIBELLI RABBI MOYSIS
parte toto coelo errare mihi videtur. Cum enim dicit Antiquorum inven- NARBONENSIS
201 tionem non expli/care ex quo genere linearum sit illa iam dicta linea, DILUCIDATIO.
magnopere hallucinatur. Nam6 Apollonius quidem in duodecima proposi­ PROEMIUM.
tione primi libri Conicorum pulcherrime ortum et formam et naturam pro-
Ostensis iam atque declaratis Autorum defectibus reliquum est ut dilu-
priamque eius affectionem explicat, quibus omnibus a circulari et recta
cudemus Rabbi Moysis Narbonensis libellum in Italica lingua scriptum,
distinguitur, inter mistasque lineas collocatur, et vocatur Hyperbole: ut
etiam ipse Peletarius sibi contradicens fatetur cum dicat: Neque dubium 7 mg.: Geminus.
est quin ipsa ex earum sit genere quas ex Apollonio proponunt, latus 8 mg.: Proclus.
9 mg.: Aristoteles primo de Caelo tex. 5.
10 mg.: Io. Vernerus.
5 mg.: Tertius Peletarii error. 11 mg.: Cardanus.
e mg. : Apollonius. 12 mg.: Caelii Calcagnini error.
432 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S FROM B A R O Z Z L S A D M I R A N D U M P R O B L E M A 433

Mantuaeque impressum anno a Natalibus Christi M.D.L, cui titulus est: Haec9 patet ex vicesimasecunda propositione libri sexti eorundem
Opus1novum Geometricum ad demonstrandum quomodo super una plana Elementorum.
superficie duae lineae possint exire quae procedentes semper invicem ap­
propinquent, nunquam tamen sibi occurrant. In eo2 itaque Libello ipse PROPOSITIO TERTIA,
Rabbi Moyses post quoddam breve proemium, per quod intentionem suam THEOREMA III.
proponit, se nempe demonstraturum Problema illud admirandum in Geo­
metria, quippequod nos iam undecim variis modis demonstravimus Si10 aliquot rectae lineae proportionales ab aliquot totidem numero rectis
decem3 et octo propositiones in medium affert, quibus totum propositum lineis proportionalibus multiplicentur: rectangula ab illis contenta propor­
absolvit, demonstratque primum rem ipsam haud dissimili demonstratione tionalia erunt. Similiter si aliquot proportionales numeri alios totidem
a nostra prima superius allata; deinde quodam etiam exemplo quod magis numeros proportionales multiplicent: qui producuntur proportionales
sensu percipi potest rem apertius declarat. Ex illis autem decem et octo erunt.
propositionibus novem quidem sunt Quaesito conferentes, reliquae vero 212 / Sint11 quatuor rectae lineae A , B, C, D [Fig. 7.58] proportionales,
Quaesito propriae ipsumque demonstrantes. Talis quidem est ipsius Rabbi scilicet ut A ad B sic C ad D; et quatuor aliae E, F ,G ,H proportionales,
210 Moysis in eo Libello intentio, libellique / divisio. Qui4 porro maxime ob­ nempe ut E ad F sic G ad H\ et multiplicetur A ab E, et fiat rectangulum
scurus est tum quoniam ex Hebraico sermone in Italicum male fuit tralatus: IKL: et B ab F, et fiat MNO: et C a G, et fiat PQR: et D ab H, et fiat STV.
tum quia multis in locis mendose legitur: tum demum quod Autor ipse Dico12 quod sicut IKL rectanguli ad ipsum MNO sic PQR ad STV. Cum13
propositiones suas non demonstrat rationibus Geometricis, sed duntaxat enim ratio quidem IK ad MN eadem est cum ratione PQ ad ST, ratio vero
proponit eas, exemplisquae numerorum confirmat: adde etiam quia ipse KL ad NO eadem cum ratione QR ad TV ex suppositione: erit ratio com-
quaedam omisit quae necessario declaranda demonstrandaque erant. 213 posita ex rationibus PQ ad ST et QR ad TV eadem rationi / compositae ex
Quare5 Libellum hunc nobis dilucidare volentibus opus est propositiones rationibus IK ad MN et KL ad NO per primam communem sententiam
illas latinitate fideliter donare, et mendis expurgare, Geometricisque ra­ huius. At rectangulum etiam IKL ad MNO per vicesimam tertiam proposi­
tionibus demonstrare, ac demum ea adiicere quae ab Autore praetermissa tionem sexti libri Element. Euclidis eandem habet rationem compositam
fuere. Sit igitur prima Propositio huius modi. ex ratione ipsius IK ad MN et ipsius KL ad NO : ergo IKL rectangulum ad
MNO compositam habet rationem ex rationibus PQ ad ST et QR ad TV
211 / PROPOSITIO PRIMA, per undecimam propositionem quinti libri eorundem Elementorum. Sed
THEOREMA I.
eandem habet etiam rationem per iam dictam vicesimamtertiam proposi­
Si6 trianguli uni lateri quotlibet rectae lineae parallelae duo reliqua eius tionem ipsum PQR ad STV. Igitur per eandem undecimam quinti ratio
latera secantes ducantur: triangula partialia in eo facta sibiinuicem et toti rectanguli IKL ad rectangulum MNO eadem est rationi ipsius PQR ad ip­
similia sunt. sum STV rectangulum, quae14 est prima propositionis pars. Similiter
Probatur7 haec Propositio per secundam partem vicesimaenonae pro­ autem si numeri tum lineis, tum superficiebus adsignentur; secunda15
positionis primi et quartam propositionem et primam definitionem sexti quoque pars demonstrabitur per primam communem sententiam huius
libri Elementorum Euclidis. semel sumptam et quintam propositionem octavi et undecimam quinti libri
Elementorum Euclidis bis sumptas. Numeri nanque linearum A, B, C ,D , et
PROPOSITIO SECUNDA, E, F ,G , H , erunt latera numerorum planorum areas superficiales rectan-
THEOREMA II. gulas denotantium. Utraque igitur propositionis pars vera et perspicua est.
Si8 aliquot rectae lineae proportionales fuerint, quadrata etiam earum Si16 aliquot itaque rectae lineae proportionales fuerint, et reliqua ut supra,
proportionalia erunt. Et si quadrata aliquot proportionalia fuerint, latera quod oportebat demonstrare.
quoque ipsorum proportionalia erunt.

1 mg.: Titulus libri. 9 mg.: Demonstratio.

2 mg.: Propositum Rabbi Moysis 10 mg.: Propositio.

3 mg.: Divisio libri. 11 mg.: Expositio.

4 mg.: Obscuritatis causae. 12 mg.: Determinatio.

5 mg. : Propositum Dilucidatoris. 13 mg.: Demonstratio primae partis.

6 mg.: Propositio. 14 mg.: Conclusio primae partis.

7 mg.: Demonstratio. 15 mg.: Demonstratio secundae partis.

8 mg.: Propositio. 16 mg.: Conclusio universalis.

434 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S F R O M B A R O Z Z T S AD M IRANDUM PROBLEMA 435

PROPOSITIO QUARTA, 215 / PROPOSITIO SEPTIMA.

THEOREMA IIII. THEOREMA VII.
In praesenti propositione applicat Rabbi Moyses affectionem prae­ S P fuerint duae rectae lineae quarum altera quidem secta sit ut in
cedentis Theorematis quibusdam rectis lineis proportionalibus quae in praecedenti propositione proponitur, altera vero ita secetur ut totius ip­
duobus simul iunctis triangulis sunt. Non indiget autem probatione, quia sius secundae lineae quadratum ad quadratum quod fit ab insecto priorum
per praecedentem patet. Nos vero ne tot demonstrationibus et figuris primae lineae segmentorum et ab altero secundorum tanquam ab una linea
legentium mentem conturbemus: nullum huius propositionis exemplum eam habeat rationem quam habet quadratum alterius segmentorum se­
subiicimus, idemque in prima et secunda propositione fecimus, faciemus- cundae lineae ad rectangulum contentum a tota prima linea et iam dicto il­
que in multis aliis sequentibus. Satis enim erit si percepta ex Elementis lius priori segmento: habebit etiam eandem rationem excessus quadrati
Euclidis earum veritate, ipsas demum praecipuo nostro proposito ap­ totius secundae lineae supra iam dictum quadratum alterius suorum seg­
plicaverimus. mentorum ad excessum quadrati ab insecto priorum primae lineae seg­
mentorum et ab altero secundorum tanquam ab una recta linea facti supra
rectangulum quod a tota prima linea et eodem eius priori segmento com­
214 / PROPOSITIO QUINTA.
prehenditur.
THEOREMA V. Sit24 AB quidem recta linea [Fig. 7.60] secta ut in praecedenti proposi­
Si17 ab aliquo puncto dimetientis Circuli ad circunferentiam recta linea tione proponitur in signis C, D, recta vero EF ita secetur in signo G ut
ad angulos rectos ducatur: erit quod a dimetientis partibus ab illa perpen­ quadratum totius EF ad quadratum ipsius DB eandem habeat rationem
diculari abscisis continetur rectangulum aequale quadrato ipsius per­ quam habet quadratum alterius segmentorum lineae EF, verbi gratia ip­
pendicularis. sius GF, ad rectangulum contentum ab AB, BC. Dico25 quod eandemmet
Haec18 probatur per tricesimam primam prop, tertii, et per corollarium quoque rationem habebit excessus quadrati lineae EF supra quadratum ip-
octavae propositionis sexti et per primam partem decimae septimae pro­ 216 sius GF ad excessum quadrati / lineae DB supra rectangulum ab AB, BC
positionis eiusdem sexti libri Elementorum Euclidis. comprehensum. Haec26 etiam patet ex decimanona propositione quinti lib.
Elementorum Euclidis. Verum quod in Constructione huius Theorematis
praecipitur, scilicet, ut recta linea EF ita secetur in signo G ut quadratum
PROPOSITIO SEXTA. ipsius EF ad quadratum ipsius DB eandem habeat rationem quam habet
THEOREMA VI. quadratum ipsius GF segmenti ad rectangulum ab AB, BC contentum:
Si19 recta linea secetur utcunque, deinde alterum ex eius segmentis per facile fieri potest per quoddam Problema quasi simile primo nostro Prob­
medium secetur: rectangulum a tota et insecto priorum segmentorum lemati in principio huius operis demonstrato, quod tale sit. Dato Parallelo-
contentum superatur a quadrato quod fit ab eodem priori segmento, et ab grammo rectangulo et duobus quadratis: invenire tertium quadratum quod
habeat eandem rationem ad datum rectangulum quam habet alterum
altero posteriorum segmentorum tanquam ab una linea, quadrato alterius
posterioris segmenti. datorum quadratorum ad reliquum. Quod quidem Problema eodem modo
eisdemque Euclidis propositionibus construitur, ac demonstratur quemad­
Sit20 recta linea AB [Fig. 7.59], quae secetur primum utlibet in signo C,
modum illud iam dictum in principio demonstratum: utendo insuper Corol­
deinde alterum segmentorum ipsius, utpote AC, secetur per medium in
lario quartae propositionis quinti et nona propositione sexti libri Ele­
signo D. Dico21 quod rectangulum ab AB, BC comprehensum superatur a
mentorum Euclidis. Adnotandum27 autem est quod iam dicta huius Theore­
quadrato ipsius DB, quadrato ipsius DC, vel ipsius DA. Hoc22 enim patet
matis Constructio tres Casus sortita est. Aut enim secanda EF recta linea
ex sexta propositione secundi libri Elementorum Euclidis, praesensque
propositio parum ab illa differt. maior est quam AB data recta linea iam secta, aut minor, aut ipsi aequalis.
Et quilibet horum trium Casuum duo membra potest habere. Nam alterum

17 mg.: Propositio.
1Kmg.: Demonstr. 23 mg.: Propositio.
19 mg.: Propositio. 24 mg.: Expositio.
20 mg.: Expositio. 25 mg.: Determinatio.
21 mg.: Determinatio. 26 mg. : Demonstratio.
22 mg.: Demonstratio. 27 mg.: Casus Constructionis Theorematis huius.
436 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S FR O M B A R O Z Z T S AD M IRAN D U M PROBLEMA 437

rectae EF lineae segmentum, cuius quadratum eandem habere debet ra­ in quarta propositione primi libri eorundem et ab aliis; a nonnullis autem
tionem ad rectangulum ab AB,BC contentum quam habet quadratum totius tanquam Petitio supponitur. Tertia autem pars facile probari potest. Quod
EF ad quadratum ipsius BD, duobus modis ab ipsa EF linea abscindi enim dimetientes basi propinquiores remotioribus maiores sint probatur
potest; vel ex parte E ut sit EG, vel ex parte F, ut sit FG ; in quibus omnibus per secundam partem vicesimaenonae propositionis primi et quartam
Casibus eadem est Constructio atque Demonstratio. Hoc28 autem quod propositionem sexti et nonam Com. Sent, eiusdem primi libri Elemen.
praecipitur commode fieri potest, quoniam tertium ipsum inveniendum Eucl. et nonam Com. Sent, huius: vel per easdem propositiones, vicesimam-
quadratum est semper minus ipso quadrato dato, a cuius latere secando nonam primi et quartam sexti, et nonam Com. Sent, primi, et per quartam-
abscisio iam dicta facienda est, ut patet ex suppositione et decimaquarta decimam propositionem quinti libri eorundem. Hoc34 autem probato, patet
propositione quinti libri Elementorum Euclidis et prima communi sententia etiam circulos basi propinquiores circulis a basi remotioribus esse maiores
huius. per tricesimam definitionem huius. Tota igitur haec propositio perspicua est.
Propositiones hucusque demonstratae Quaesito conferentes sunt:
217 / PROPOSITIO OCTAVA, sequentes autem Quaesito propriae erunt.
THEOREMA VIII.
PROPOSITIO DECIMA,
Si29 recta linea secetur utcunque, quadratum totius superat quadratum THEOREMA X.
alterutrius segmentorum rectangulo bis a segmentis comprehenso una cum
quadrato reliqui segmenti. Si35 iam dicti trianguli per axem in Cono facti latus alterum versus Coni
Haec30 ex quarta propositione secundi libri Elementorum Euclidis om­ verticem in directum producatur, et ab eius extremitate extra Conum
nino manifesta est, ab illaque parum discrepat. existente ad conicae basis dimetientem recta linea ducatur secans reliquum
trianguli latus et omnes rectas lineas basi parallelas: erit ratio rectanguli
contenti a partibus cuiuscunque parallelae ad rectangulum comprehensum
PROPOSITIO NONA, a tota partibus ipsis conterminali versus Coni verticem extensa et parte ip­
THEOREMA IX. sius conterminalis intra Conum existente sicut rectanguli comprehensi a
219 partibus cuiuscunque alterius parallelae ad rectangulum / contentum a tota
Si31 Conus a Vertice ad basim plano secetur: communes plani et super­ similiter ipsis conterminali et eius parte intra Conum existente. Et rectan-
ficiei conicae sectiones una cum basis conicae dimetiente triangulum recti- gula quae continentur a partibus parallelarum basi propinquiorum erunt
lineum faciunt vocatum triangulum per axem. Cuius duo latera supra Coni semper maiora rectangulis quae a partibus parallelarum a basi remotiorum
basim consistentia si rectae lineae basi trianguli parallelae secuerint, continentur.
omnes illae rectae lineae erunt dimetientes circulorum in superficie conica Sit36 iam dictum triangulum per axem in Cono factum ABC [Fig. 7.61],
circunferentias habentium. Atque ex eis tum dimetientibus, tum tum (/) cuius latus AC versus Coni verticem in directum producatur usque ad D,
circulis basi propinquiores quidem maiores remotioribus sunt. et a puncto D ducatur DE secans ipsam quidem BC basim trianguli sive
Tres32 haec propositio partes habet. Prima, quod Coni sectio a vertice conicae basis dimetientem in signo E: reliquum vero trianguli latus AB in
usque ad basim triangulum sit, quod triangulum per axem definivimus. signo F. Sintque ipsi i?C basi parallelae GH et IK, quas secet ipsa DE in sig­
Secunda, quod rectae lineae basi trianguli parallelae duo eius latera se­ nis L, M. Dico37 itaque rectangulum a GL, LH contentum ad rectangulum
cantes sint dimetientes circulorum in superficie conica circunferentias quod a D L , LF continetur eandem habere rationem quam habet rectan­
habentium. Tertia, quod dimetientes et circuli basi conicae propinquiores gulum ab IM , MK ad rectangulum a DM, MF comprehensum; et quod
maiores remotioribus sint. Harum33 igitur trium partium prima quidem rectangulum ab IM, MK contentum maius est rectangulo a GL, LH com­
218 sensui patet, et a plerisque / qui de Conicis scripsere tanquam Petitio prehenso. Quum38 enim in triangulo quidem BEF ipsae GL, IM basi BE
ponitur; ab Apollonio autem in propositione tertia primi libri Conicorum et parallelae sint, in triangulo vero DEC ipsae LH, MK basi EC similiter sint
a nonnullis aliis demonstratur. Secunda vero probatur ab eodem Apollonio parallelae: erit triangulum FGL simile triangulo FIM, et triangulum DLH

28 mg.: Notandum.
29 mg.: Propositio. mg.: Conclusio.
:!0 mg.: Expositio. mg.: Propositio.
31 mg.: Propositio. mg.: Expositio.
32 mg.: Expositio. mg.: Determinatio.
33 mg.: Demonstratio. mg.: Demonstratio primae partis.
438 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S FRO M B A R O Z Z T S A D M I R A N D U M P R O B L E M A 439

triangulo DM K, per primam propositionem huius Dilucidationis bis sump­ ad angulos rectos erectae lineae. Unde ipsa etiam linea ad rectos angulos
tam. Sicut igitur GL ad LF sic IM ad MF, et quemadmodum FIL ad LD ita erecta basi propinquior remotiore maior erit.
220 KM ad MD, per primam de/finitionem sexti libri Elementorum Euclidis bis 222 / Sit44 eadem figura quae in praecedenti propositione (Fig. 7.61], et a
sumptam. Quare si quatuor quae in FIM triangulo proportionales ostensae punctis L & M intelligantur duae rectae lineae L N , MO ad rectos angulos
sunt lineae in quatuor quae in triangulo DMK similiter sunt ostensae pro­ erectae plano trianguli ABC conicae occurrentes superficiei in signis A , O
portionales, unaquaeque in suam correspondentem, multiplicentur, prima [Fig. 7.62]. Dico45 quod ratio quadrati lineae LN ad rectangulum a D L , LF
scilicet in primam, et secunda in secundam, et tertia in tertiam, quartaque contentum est sicut ratio quadrati ipsius MO ad rectangulum a D M , MF
in quartam: fient rectangula ab ipsis comprehensa invicem proportionalia; comprehensum. Intelligantur46 duo plana conicae basi parallela Conum
nempe sicut primum ad secundum ita tertium ad quartum per tertiam secantia, quorum communes sectiones cum plano trianguli ABC sint per
propositionem huiusce Dilucidationis. At primum quidem est quod a GL, tertiam propositionem undecimi libri Elementorum Euclidis rectae lineae
LH : secundum vero, quod a DL, LF: tertium autem, quod ab IM, MK: GLH et IMK, communes vero eorundem planorum et conicae super­
quartum denique, quod a DM, MF continetur. Igitur39 sicut quod a GL, ficiei sectiones sint per quartam petitionem huius circulorum circunferen-
LH ad id quod a D L ,L F ita quod ab IM, MK ad id quod a DM, MF. Patet tiae, quae porro transibunt per signa N , O, quoniam ipsae L N , MO in
itaque prima propositionis pars. Et nunc quidem in proposito nostro eodem sunt plano cum ipsis GH, IK per secundam propositionem eiusdem
praecipuo demonstravimus id quod Rabbi Moyses in duobus simul iunctis undecimi, ipsae autem GLH et IMK erunt per eandem quartam peti­
triangulis frustratorie in quarta propositione declaravit. Vanum est enim tionem huius dimetientes eorum circulorum. Quoniam47 itaque rectae
in Geometria bis idem repetere. Secunda40 vero pars propositionis per­ lineae LN et MO ex suppositione ad rectos sunt angulos plano trianguli
spicua est ex tertia communi sententia huius. Partes nanque parallelarum ABC: ergo per tertiam definitionem eiusdem undecimi Elementorum ipsis
basi propinquiorum partibus remotiorum maiores sunt unaquaeque sua etiam GLH et IMK dimetientibus ad rectos sunt angulos. Quare per quin­
correspondent!, ipsa scilicet IM maior quam GL et ipsa MK maior quam tam propositionem praesentis Dilucidationis quadrata ipsarum L N , MO
LH. Quod41 autem hoc verum sit duobus illis modis probari potest quibus aequalia sunt rectangulis a GL, LH et ab IM, MK contentis. Igitur per
in praecedenti propositione dimetientes propinquiores basi maiores esse primam partem septimae propositionis quinti libri Elementorum Euclidis
221 remotioribus probatum fuit. Utraque igitur propositionis / pars clara iam ratio quadrati lineae MO ad rectangulum a DM, MF contentum est sicut
est. Si42 ergo iam dicti trianguli per axem in Cono facti latus alterum versus ratio rectanguli ab IM, MK contenti ad idem a DM, MF contentum rectan-
Coni verticem in directum producatur, et reliqua ut in propositione. Quod 223 gulum. / Similiter eadem erit ratio quadrati LN ad rectangulum a DL, LF
demonstrandum erat. comprehensum, quae etiam est rectanguli a G L ,L H comprehensi ad idem
iam dictum rectangulum. Est autem per praecedentem propositionem
PROPOSITIO UNDECIMA. ratio eius quod a G L, LH ad id quod a D L ,L F continetur sicut eius quod ab
THEOREMA XI. IM, MK ad id quod a DM, MF comprehenditur. Ergo per undecimam prop,
quinti libri eorundem Elem. bis sumptam eadem erit ratio quadrati lineae
Si43 in recta linea ab extremitate producti lateris trianguli per axem ad
LN ad id quod a DL, LF, quae est quadrati lineae MO ad id quod a DM,
Coni basin (ut praecedens propositio proposuit) deducta aliquot puncta
MF comprehenditur rectangulum. Quae48 quidem est prima pars proposi­
intra Conum accipiantur, ab eisque rectae lineae plano ipsius trianguli ad
tionis. Secunda vero pars patet ex quinta huius Dilucidationis et ex secunda
rectos angulos erigantur Conicae occurrentes superficiei: erit ratio quadrati
parte praecedentis prop, et ex prima et secunda parte septimae prop,
uniuscuiusque ipsarum ad rectos angulos erectarum ad rectangulum con­
eiusdem quinti Elem. et ex nona Com. Sent, huius: vel ex eadem quinta
tentum a tota sua conterminali versus Coni verticem extensa et parte ipsius
prop, et secunda parte praecedentis et ex septima Com. Sent, huius: vel ex
conterminalis intra Conum existente sicut ratio quadrati cuiuslibet alterius
eadem quinta et secunda parte praecedentis et prima parte 14 prop. quin,
ad angulos rectos erectae ad rectangulum a tota similiter sua conterminali lib. Elemen. Eucl. Cum49 enim rectangulum a G L ,L H contentum aequale
et eius parte intra Conum existente comprehensum. Et quadratum ad
sit quadrato ipsius LN, et rectangulum ab IM, MK comprehensum quadrato
angulos rectos erectae propinquioris basi maius erit quadrato remotioris

44 mg.: Expositio.
39 mg.: Conclusio primae partis. 45 mg.: Determinatio.
40 mg. : Secundae partis demonstratio. 46 mg.: Constructio.
41 mg.: Conclusio secundae partis. 47 mg.: Demonstratio primae partis.
42 mg.: Conclusio totius propositionis. 48 mg. : Conclusio primae partis.
43 mg.: Propositio. 49 m g.: Demonstratio secundae partis triplex.
440 A R C H IM E D E S IN T H E M ID D L E A G E S E X T R A C T S FROM B A R O Z Z FS A D M I R A N D U M P R O B L E M A 441

ipsius MO per iam dictam quintam prop. Dilucidationis, ipsum autem quod Reponatur igitur hic figura illa [Fig. 7.51], quae etiam sequentibus proposi­
ab IM, MK continetur maius sit contento a G L ,L H per secundam partem tionibus deserviet, et omnia clara erunt.
praecedentis; necessario quadratum etiam ipsius MO quadrato ipsius LN
maius erit, propter iam dictas propositiones et Communes Sententias. 225 / PROPOSITIO TERTIADECIMA.
Quae50 est secunda pars propositionis. Ex51 hac autem secunda parte, et ex THEOREMA XIII.
Secunda Com. Sent, huius patet etiam tertia propositionis pars. Tota Cunctis57 ita se habentibus ut in praecedenti propositione, si quadratum
igitur haec propositio perspicua est. Si52 itaque in recta linea ab extremitate parallelae quae tota extra Conum est ad suae conterminae usque ad biparti­
producti lateris trianguli, et reliqua, ut in Theoremate. Quod53 demon­ tam sectionem pervenientis quadratum eandem habuerit rationem quam
strasse oportuit. Adnotandum54 autem est quod tertiam huius propositionis 226 habet quadratum factum ab interna par/te cuiuslibet reliquarum paral­
partem tacuit Rabbi Moyses tanquam manifestam. Quoniam autem absque lelarum ad rectangulum comprehensum a tota eiusdem internae partis con-
ea propositum praecipuum concludi non potest (ut inferius constabit), terminali usque ad extremitatem producti lateris trianguli per axem per­
idcirco eam nos adiunximus ac demonstravimus. Quanvis autem tota veniente et eius conterminalis parte intra Conum existente: quadrata
praesens propositio eadem sit cum propositione secunda trium in principio quoque totarum reliquarum parallelarum ad quadrata suarum conter­
huius operis positarum, nihilominus placuit hic quoque eam ponere atque minalium usque ad bipartitam sectionem pervenientium eandem habebunt
demonstrare ne ordinem Elementarem huius Dilucidationis discerperem, rationem quam habent quadrata internarum partium parallelarum ad iam
et praesertim cum demonstratio haec aliquantulum ab illa diversa sit. dicta rectangula.
Propositio58 haec facile probari potest per primam partem undecimae
propositionis huius Dilucidationis et per praecedentem et per undecimam
224 / PROPOSITIO DUODECIMA. propositionem quinti libri Elemen. Euclidis toties sumptam, quoties
THEOREMA XII. oportuerit; si hic quoque praecedentis propositionis figura [Fig. 7.51]
Si55 a signo in quo linea recta ab extremitate producti lateris trianguli per repetatur. Notandum59 autem quod in hac propositione supponitur quod
axem ad Coni basim deducta reliqu[u]m trianguli latus secat quaedam recta fiat quadratum lineae GH ad quadratum lineae GM sicut quadratum lineae
linea plano trianguli ad rectos angulos erigatur parallela illis quae intra 227 LK ad rectangulum ab EL, LG conten/tum, quod quidem fiet per primam
Conum (ut proponit praecedens propositio) eidem plano ad rectos angulos huius operis propositionem, coadiuvante Corollario quartae propositionis
erectae sunt, atque pars ipsius lineae ab extremitate deductae extra Conum quinti libri Elementorum Euclidis.
iacens per medium secetur, et a puncto huiusce sectionis ducatur recta PROPOSITIO QUARTADECIMA,
linea secans extra Conum omnes iam dictas parallelas in directum pro­ THEOREMA XIII.
ductas: quadrata totarum earum parallelarum productarum ad quadrata
In hac Propositione concludit Rabbi Moyses alteram praecipui quaesiti
suarum conterminalium ad bipartitam (corr. ex bipertitam) usque sec­
partem (nempe duarum in eodem plano describendarum non coincidentium
tionem se extendentium eandem rationem habebunt.
linearum continuam appropinquationem) iuxta numerorum exempla quae
Praesens propositio nulla prorsus indiget declaratione; quoniam posita
in superioribus propositionibus ipse posuit. Quoniam autem nos nullum
hic primae nostrae demonstrationis secunda figura [Fig. 7.51], facile potest
exemplum in numeris dedimus, cum propositum nostrum sit Geometricis
ex prima et secunda huius Dilucidationis propositionibus demonstrari.
tantum rationibus Quaesitum demonstrare: idcirco nulla habita huius
Cum56 enim per iam dictam primam Dilucidationis propositionem sit ut
propositionis tanquam frustratoriae (qualis etiam quarta fuerat) exposi­
GH ad GM sic LO ad LM, et QR ad QM, et si quae essent aliae parallelae:
tione, ad sequentes nobis transeundum est quae propositum nostrum
manifestum est per secundam eiusdem Dilucidationis quod sicut etiam
Geometrice concludent.
quadratum ipsius GH ad quadratum ipsius GM ita quadatum ipsius LO ad
quadratum ipsius LM, et quadratum ipsius QR ad ipsius QM quadratum. PROPOSITIO QUINTADECIMA,
THEOREMA XV.
50 mg.: Conclusio secundae partis. Omnibus60 eo modo iacentibus quo propositio tertiadecima proposuit:
51 mg.\ Demonstratio tertiae partis. excessus quadratorum totarum parallelarum partim intra partim extra
52 mg.\ Conclusio eiusdem.
53 mg.: Conclusio universalis. 57 mg.: Propositio.
54 mg.: Notandum. 58 mg.: Demonstratio.
55 mg.: Propositio. 59 mg.: Notandum.
56 mg. : Demonstratio. 60 mg.: Propositio.
442 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZI S AD M IRAN DUM PROBLEMA 443

Conum iacentium supra quadrata suarum internarum partium eam habe­ supra quadratum LK ad excessum quadrati LM supra rectangulum conten­
bunt rationem ad excessus quadratorum linearum ipsis parallelis conter- tum ab EL, LG, idest per primam partem praesentis propositionis, ad
minalium usque ad bipartitam illam sectionem supra rectangula contenta quadratum GM habeat rationem sicut quadratum LO ad quadratum LM per
a totis conterminalibus usque ad extremitatem Producti triangularis lateris primam partem praecedentis: quadratum autem LO ad quadratum LM eam
pervenientibus et partibus earundem conterminalium internis quam habent habeat rationem quam quadratum GH ad quadratum GM per duodecimam
228 qua/drata totarum ipsarum parallelarum ad quadrata ipsarum conter­ propositionem huius Dilucidationis: ergo per undecimam propositionem
minalium usque ad bipartitam sectionem se extendentium: nec non eam eiusdem quinti Elementorum excessus quadrati LO supra quadratum LK
quam habent quadrata internarum partium parallelarum ad iam saepe dicta eam habebit rationem ad excessum quadrati LM supra rectangulum con­
rectangula. tentum ab E L ,L G , idest ad quadratum GM, quam habet etiam quadratum
Duas haec etiam propositio partes habet, quarum61 prima quidem patet ex GH ad idem GM quadratum. Igitur per primam partem nonae propositionis
tertiadecima et septima propositione huius Dilucidationis: secunda vero eiusdem quinti Elementorum excessus quadrati LO supra quadratum KL
probatur per primam partem huius propositionis et per tertiamdecimam aequalis est quadrato lineae G H . Idem quoque de excessibus quadratorum
huius Dilucidationis, et per undecimam propositionem quinti libri Ele­ caeterarum parallelarum totarum supra quadrata suarum internarum
mentorum Euclidis, ut in proxima figura declarari potest. partium ostendetur. Praeterea65 quoniam per secundam / partem prae­
cedentis propositionis excessus quadrati LO supra quadratum KL ad ex­
229 / PROPOSITIO SEXTADECIMA. cessum quadrati LM supra rectangulum comprehensum ab EL, LG, idest
THEOREMA XVI. ad quadratum GM, eandem habeat rationem quam habet quadratum KL
ad iam dictum rectangulum: quadratum autem LK ad iam dictum rectan­
Manentibus cunctis ut in praecedenti propositione: necessario excessus gulum eam habet rationem (per suppositionem tertiaedecimae propositionis
quo iam dicta rectangula a conterminalium usque ad bipartitam sectionem huiusce Dilucidationis) quam habet quadratum GH ad idem GM quadratum:
se extendentium quadratis superantur nil aliud est nisi quadratum rectae erit per eandem undecimam et primam partem nonae prop, quinti lib. Elem.
lineae quae in ipsa conterminali inter lateris triangularis sectionem et bi­ Eucl. excessus quadrati LO, quo superat quadratum LK , aequalis quadrato
partitam divisionem recipitur. Unde excessus quo quadrata internarum GH. Idem autem hoc quoque secundo modo de caeteris etiam quadratorum
partium parallelarum a totarum parallelarum quadratis exceduntur aequa­ parallelarum totarum supra quadrata suarum internarum partium exces­
lis est quadrato primae parallelae, quae tota extra Conum iacet. Necnon sibus demonstrabitur. Patet igitur utroque modo secunda propositionis
isti omnes excessus quadratorum totarum parallelarum supra quadrata pars. Tertia66 vero ex secunda iam demonstrata et ex prima communi
suarum internarum partium invicem aequales sunt [Fig. 7.51]. sententia primi libri Elementorum Euclidis manifesta est. Tota igitur haec
Praesens propositio tres habet partes, quarum prima quidem patet ex propositio perspicua relinquitur. Manentibus67 igitur cunctis, et reliqua ut
sexta propositione huius Dilucidationis. Cum62 enim ipsaEL utcunque in
supra. Quod demonstrandum erat.
signo G secta sit, alterum autem ipsius segmentum, nempe GE, per medium
in signo M sit divisum, proculdubio per ipsam sextam propositionem
rectangulum a tota EL et insecto eius segmento LG contentum superatur a PROPOSITIO DECIMASEPTIMA.
quadrato ipsius ML ex eodem priori GL segmento et altero secundorum THEOREMA XVII.
segmentorum, nempe GM , constantis, quadrato eiusdem GM lineae Continens alteram partem demonstrationis duodecimae.
idemque in caeteris etiam ostendetur. Secunda63 autem probatur per Cunctis68 similiter ut in praecedenti iacentibus, excessus quadratorum
primam partem praesentis, et praecedentis, et per duodecimam huius totarum parallelarum supra quadrata suarum internarum partium nil aliud
Dilucidationis, et per undecimam propositionem quinti libri Elementorum est nisi duplum rectanguli ab interna et externa parallelarum parte contenti,
Euclidis, et primam partem nonae propositionis eiusdem quinti: Vel per una cum quadrato externae partis earundem. Unde quadrata externarum
secundam partem praecedentis, et suppositionem tertiadecimae proposi­ partium basi Conicae propinquiorum quadratis externarum partium a basi
tionis huius Dilucidationis, et eandem undecimam quinti, et primam remotiorum minora sunt. Nec non ipsae externae parallelarum partes basi
230 partem nonae propositio/nis eiusdem. Cum64 enim excessus quadrati LO 81*4 propinquiores externis partibus a basi remotioribus sunt minores.

81 mg.: Demonstratio. 65mg.: S e c u n d a D e m o n str a tio s e c u n d a e p artis.

β2 mg.: Demonstratio primae partis. 66mg. : T e r tia e p artis D e m o n str a tio .
63 mg. : Demonstratio secundae partis duplex. 67mg.: C o n c lu s io u n iv e r sa lis .
84 mg.: Prima Demonstratio secundae partis. 68mg. : Propositio.
444 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZTS A D M IR A N D U M PR O B LE M A 445

232 / Tres haec quoque partes habet, quarum prima69quidem ex octava huius tionibus et vicesimaprima definitione huius manifestum est) ab eodem
Dilucidationis omnino clara est. Secunda vero facile demonstrari potest plano Conum secante, eo magis sibi appropinquant, quanto longius versus
per secundam partem tertiae propositionis illarum trium quas in operis Coni basim in eodem iam dicto plano producuntur. Quod quidem sic
initio demonstravimus, peracta scilicet Constructione. Intelligatur70 enim breviter concludatur. Quoniam KO maior est quam PR, ut fuit ostensum,
duplum rectanguli ab LK, KO Contenti [Fig. 7.51], quod est per primam igitur et KT minima distantia puncti Carecta Mi? maior est quam PV minima
propositionem secundi libri Elementorum Euclidis rectangulum a KO et distantia puncti P ab eadem MR recta linea. Haec enim in nostra prima
dupla ipsius KL contentum, quadrato ipsius KO sic esse adiunctum ut demonstratione satis superque demonstrata fuere. Idem77 autem de omni­
proponit illa iam dicta tertia propositio: necnon duplum rectanguli a QP, bus etiam aliis inflexae GKPID lineae a recta MR minimis distantiis os­
PR contenti, idest rectangulum a PR et dupla ipsius PQ contentum, eodem tendetur. Alterum igitur Quaesiti praecipui membrum clarum est. In78 quo
modo quadrato ipsius PR adiectum esse. Cum71 itaque duo haec aggregata sane demonstrando Rabbi Moyses paralogismum illum committit quem
rectangula per primam partem huius propositionis sint excessus quadra­ etiam omnes, quos legimus huiusce rei Autores praeter Cardanum com­
torum LO et QR supra quadrata KL et PQ, quippe qui excessus per tertiam miserunt: quandoquidem ex ipsarum KO et PR, caeterarumque similium
partem praecedentis propositionis invicem aequales esse demonstrati sunt: inaequalitate, duarum linearum, inflexae scilicet GKPID et rectae MR,
cumque linea PQ maior sit quam KL per tertiam partem undecimae pro­ continuam versus Coni basim appropinquationem concludit. Demonstrato
positionis huius Dilucidationis, ideoque dupla etiam lineae PQ maior sit igitur altero praecipui Quaesiti membro, ad reliquum nobis pergendum
quam dupla lineae KL per quintamdecimam propositionem quinti libri est, quod in sequenti propositione demonstrabitur.
Elementorum Euclidis: igitur per secundam partem illius iam commemo­
ratae tertiae propositionis quadratum lineae KO maius est quadrato lineae
P R . Idemque de caeteris etiam externarum partium quadratis demonstrari / PROPOSITIO DECIMAOCTAVA.
potest. Patet72 itaque secunda etiam propositionis pars. Cuius utique THEOREMA XVIII.
demonstrationem obscure admodum et caliginose Rabbi Moyses declarat Continens reliquam partem demonstrationis duodecimae.
quoniam73 tertiam illam propositionem a nobis in superioribus demon­ Omnibus79 eodem modo ut in praecedenti dispositis, illae iam dictae
stratam ipse sicco pede praeteriit, ex qua nimirum tota huius secundae lineae, Hyperbolica scilicet et recta in eodem ambae plano existentes, et
partis demonstratio dependet. Ex hac autem secunda propositionis parte iam quo magis producuntur, eo magis sibi invicem proximantes, ut in prae­
demonstrata et ex secunda communi sententia huius tertia74 quoque proposi­ cedenti demonstratum est: si etiam in infinitum protrahantur, nunquam
tionis pars explorata relinquitur. Quare tota propositio candore lucescit.
coincident.
Cunctis75 igitur, etc. ut supra. Quod erat demonstrandum. Verum enim-
/ Sit80 hic quoque figura superiorum propositionum posita [Fig. 7.51].
vero cum ex tertia parte huius propositionis pateat externas parallelarum Dico81 itaque duas lineas, inflexam scilicet GKPID et rectam MR in eodem
partes conicae basi propinquiores externis earundem partibus a basi plano iacentes, et quanto magis versus Coni basin producuntur, eo magis
remotioribus esse minores, idemque de cunctis externis eiusmodi partibus sibi[i]nvicem proximantes: si etiam in infinitum protrahantur, nunquam
in infinitum ostendi possit: habemus iam alterum praecipui nostri Quaesiti
sese contingere. Nam82 si fieri potest tangant se in aliquo signo, exempli
membrum veluti demonstratum, quod scilicet duae lineae in eodem plano gratia in signo I, a quo ad GF rectam lineam ducatur (ut in prima nostra
iacentes quanto longius producuntur, eo magis sibi invicem propiores demonstratione docimus) IN recta linea perpendicularis in planum trian­
fiant. Nam76 recta MR linea, quae a bipartita illa sectiobe (/) extra Coni guli ABC et parallela ipsis GH et KL et PQ rectis lineis. Erit igitur per
233 superficiem protracta / fuit secans omnes illas parallelas extra Conum primam partem undecimae huius Dilucidationis ratio quadrati lineae IN
productas, in eodemque plano cum dicta recta linea iacentes: necnon in­ ad rectangulum ab EN, NG comprehensum sicut ratio quadrati lineae KL
flexa GKPID linea latus Hyperboles factae (ut ex superioribus construc­ ad rectangulum ab EL, LG contentum. Sed ratio quadrati lineae KL ad
rectangulum ab E L , LG contentum per tertiamdecimam huiusce Dilucida­
69 mg.: Primae partis demonstratio. tionis est sicut ratio quadrati lineae LO ad quadratum lineae LM. Ergo per
70 mg.: Secundae partis Constructio.
71 mg.: Demonstratio secundae partis. 77 mg. : Conclusio alterius partis Quesiti praecipui.
72 mg.: Conclusio secundae partis. 78 mg.: Rabbi Moysis defectus secundus.
73 mg.: Rabbi Moysis defectus primus. 79 mg.: Propositio.
74 mg.: Tertiae partis demonstratio. 80 mg. : Expositio.
75 mg.: Conclusio universalis. 81 mg.: Determinatio.
78 mg.: Demonstratio alterius partis Quaesiti principalis. 82 mg.: Constructio, et Demonstratio reliquae partis Quaesiti praecipui.
446 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZFS A D M IR A N D U M PRO BLE M A 447

undecimam propositionem quinti libri Elementorum Euclidis ut quadratum tates extra Conum iacentes rectam ducere lineam in eodem plano cum
ipsius IN ad rectangulum ab EN, NG contentum sic quadratum ipsius LO 237 ipsa Hyperbolica iacentem, quippe quae tanto eidem i Hyperbolicae
ad quadratum ipsius LM. Quadratum autem ipsius LO ad ipsius LM quad­ propior fiet, quanto longius ambae eodem in plano versus Coni basim una
ratum eandem habet rationem quam ipsius IN quadratum ad quadratum cum toto ipso Cono producentur: nunquam tamen ipsam tanget, etiam si in
ipsius MN per duodecimam huius Dilucidationis: igitur per eandem un­ infinitum protrahantur. Verumtamen quoniam exemplum hoc ipsius Rabbi
decimam quinti ut quadratum IN ad rectangulum ab EN, NG sic etiam idem admodum obscure, perplexe, conciseque ab ipso declaratur; et praesertim
IN quadratum ad quadratum ipsius MN. Quare per secundam partem quia Lemma quoddam Geometrica nimirum demonstratione indigens
nonae propositionis eiusdem quinti Elementorum quadratum lineae MN sicco pede praeterire videtur, in quo tota huius exempli vis consistit: id­
aequale est rectangulo ab EN, NG contento, quod est absurdum: quoniam circo haec ultima pars a nobis leviter praetereunda non est, sed maxime il­
revera quadratum ipsius MN superat rectangulum ab EN, NG contentum lustranda atque dilucidanda. Pulcrum siquidem exemplum hoc est, et
quadrato lineae GM, ut in sextadecima Dilucidationis huius demon­ sensu perceptibilem rei veritatem Geometricis rationibus confirmatam ob
stratum fuit. Non tangit igitur recta linea MR inflexam GKPID in signo /. oculos per quandam imaginationem expeditissime ponit. Age igitur illud
Similiter autem ostendetur quod neque etiam in alio puncto dictae lineae Lemma prius demonstremus, quod ipse Rabbi indemonstratum reliquit.
sese tangere possunt. Nullibi ergo se contingent si etiam in infinitum Sit autem Theorema huiuscemodi.
producantur. Omnibus83 itaque eodem modo, etc. ut in propositione. Quod
demonstrasse oportuit. Verum84 hoc etiam demonstrato, universum iam Lemma Sequentis Demonstrationis.
praecipuum nostrum Quaesitum luce clarius est. Hactenus igitur totam THEOREMA.
Rabbi Moysis demonstrationem iam dilucidavimus, ad perfectionemque Si85 quotlibet circuli concentrici in eodem plano designati fuerint, in
Geometricam quoad fieri potuit redegimus, quae quidem erit duodecima eisque duae rectae lineae parallellae ducantur, altera quidem tangens ul­
praecipui nostri Problematis demonstratio. timum interiorem circulum et se[c]ans reliquorum circulorum circun-
ferentias, altera vero a communi centro exiens secansque omnes circu-
238 lorum circunferentias; et a punctis ubi linea / a centro exiens circunferen-
236 / DILUCIDATIO ULTIMAE tias secat ad reliquam parallelam perpendiculares rectae ducantur lineae:
PARTIS LIBELLI erunt ipsius rectae lineae interiorem ultimum circulum tangentis partes
RABBI MOYSIS inter ipsas perpendiculares et circulorum circunferentias receptae quo
Quae continet aliam sensu magis perceptibilem praecipui Pro- magis a iam dicto contactu remotae, eo quidem minores.
lematis Demonstrationem, quae erit Propositi Sit86 circulus AB [Fig. 7.63], cuius centrum C, et intra ipsum alius DE
nostri xiii, et ultima. circulus habens idem centrum, et intra hunc tertius circulus FG prioribus
Post ipsam autem iam dilucidatam demonstrationem volens ipse Rabbi concentricus, et a dato puncto G (quodcunque sit) ducatur per decimam-
Moyses ostendere quodam etiam sensu magis perceptibili exemplo quod septimam propositionem tertii libri Elementorum Euclidis recta linea
causa huiusce admirandi effectus in praefatis lineis supra Conum descrip­ tangens datum circulum FG in ipso signo G, et secans circunferentiam
tis non aliunde provenit nisi ab ipsa tumosa montosaque Coni rotunditate: quidem DE in signo E, circunferentiam vero AB in signo B, quae per se­
reducit per imaginationem Conum in plano, in quo reperit quasdam a cundam petitionem primi libri eorundem producatur ultra signum B inter­
conica superficie distantias, quae quanto magis Conus crescit basim ver­ minate, et per punctum C ducatur per tricesimamprimam propositionem
sus, tanto minores fiunt; et nunquam tamen adeo decrescunt quod nullae eiusdem primi libri Elementorum parallela ipsiBG, quae sit CFDA, et a sig­
prorsus evadant. Quo quidem in plana superficie ostenso, reducit iterum nis A , D, F erigantur per undecimam propositionem eiusdem primi ad
per excogitationem Conum ipsum cum omnibus designatis lineis, ut rectos angulos ipsi AC ipsae AH, DI, FK rectae lineae, quae productae
iacebat in plano, ad corpus conicum, in cuius conica superficie ostendit secent ipsam GB ex parte B interminatam in signis Η, I, K. Secabunt
easdem iam dictas permanere distantias continue versus Coni basim enim eam necessario cum secent ipsi parallelam AC ratione saepe superius
decrescentes et nunquam evanescentes. Postea docet per extremitates dicta. Erunt itaque ipsae AH, DI, FK perpendiculares super ipsam GH per
tertiam partem vicesimaenonae propositionis et decimam definitionem
ductarum distantiarum Cono infixas curvam, seu Hyperbolicam lineam, in
quodam plano designare, et per alteras earundem distantiarum extremi- 239 primi libri Elementorum Euclidis, cadentque per sextamdecimam / pro-

83 m g Conclusio eiusdem reliquae partis. mg.: Propositio.

84 mg.: Conclusio universalis. mg.: Expositio.
448 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZTS A D M IR A N D U M PRO BLEM A 449

positionem tertii lib. eorundem extra circulorum circunferentias quas tan­ torum, quod utique est maximum inconveniens, quoniam revera con­
gunt per Corollarium eiusdem sextaedecimae. His ita expositis dico87 tentum ab LI, BH minus est contento ab MH, HB per primam proposi­
lineam EI minorem esse ipsa GK, et ipsam BH ipsa EI. Si88 enim ita non tionem sexti libri eorundem Elem. Accepta enim BH pro communi alti­
fuerit, sit si fieri potest EI aequalis ipsi GK, vel maior quam ipsa; et pro­ tudine, quemadmodum basis MH basi LI per nonam Comm. Sent, primi
ducatur per secundam petitionem primi libri eorundem ipsa HG in partem eorundem maior est ita etiam contentum ab MH, HB contento ab LI, BH
G quousque secet ex altera parte circunferentias circulorum ipsius quidem maius erit per nonam Com. Sent, huius. Non est igitur BH ipsi £7 aequalis.
DE in signo L, ipsius vero AB in signo M. Quoniam89 igitur FK et KG Sit modo maior quam ipsa. Erit itaque rectangulum ab MH, HB compre­
tangunt circulum FG, aequales sunt per tricesimamsextam propositionem hensum maius rectangulo ab LI,IE comprehenso per tertiam Comm. Sent,
tertii et primam Communem Sent, primi libri Elementorum Euclidis et per huius. Hoc autem fieri non potest, quoniam paulo ante duo iam dicta
secundam Com. Sent, huius. Est autem FK aequalis ipsi DI, necnon ipsi rectangula aequalia invicem esse ostensa sunt. Non est ergo BH maior
AH, per tricesimamquartam propositionem eiusdem primi Elementorum. quam EI: at ostensum est quod neque etiam aequalis ipsi esse potest: de
Ipsae enim FK, DI, AH invicem parallelae sunt per secundam partem necessitate igitur minor quam ipsa erit. Quare demonstratum est ipsam BH
vicesimaeoctavae propositionis primi libri eorundem, et ipsa igitur GK ip­ ipsa EI, ipsamque EI ipsaGX, minorem esse. Idem autem eodem modo de
sis DI et AH aequalis est per primam Comm. Sent, eiusdem primi bis caeteris quoque huiuscemodi rectis lineis, si etiam infiniti essent con­
sumptam. Si ergo EI sit aequalis ipsi GK, erit etiam aequalis ipsi DI per centrici descripti circuli, ostendi potest. Patet90ergo propositum indirecte.
eandem primam Com. Sent. Est autem rectangulum ab LI, IE contentum Possumus autem et directe breviter idem sic demonstrare.91 Quoniam
aequale quadrato ipsius DI per iam dictam tricesimamsextam tertii. Igitur quadratum ipsius GK quadrato ipsius FK per tricesimamsextam prop, tertii
per eandem secundam Comm. Sent, huius, et per primam Com. Sent, et primam Com. Sent, primi libri Elementorum Euclidis aequale est, atque
eiusdem primi Elemen. quod ab LI, IE continetur aequale est quadrato idcirco quadrato etiam ipsius DI per secundam Comm. Sent, huius et ean­
ipsius EI, totum scilicet suae parti per tertiam propositionem secundi libri dem primam Com. Sent, primi. Est autem quod etiam ab LI, IE con­
Elementorum Euclidis, quod fieri non potest per nonam Com. Sent, primi tinetur rectangulum per eandem tricesimamsextam tertii aequale eidem ip­
libri eorundem. Non est igitur EI aequalis ipsi GK. Sit modo maior quam sius DI quadrato: igitur per eandem primam Comm. Sent, rectangulum ab
ipsa, erit etiam maior quam DI per secundam partem septimae proposi­ LI,IE contentum quadrato ipsius GK erit aequale. Sed quadratum ipsius EI
tionis quinti libri Elementorum Euclidis, et nonam Com. Sent, huius. minus est per tertiam propositionem secundi libri Elementorum Euclidis
Quare per secundam Comm. Sent, huius quadratum lineae EI maius erit rectangulo ab LI, IE contento: ergo per secundam partem septimae pro­
quadrato lineae DI. Cum autem rectangulum ab LI, IE contentum aequale positionis quinti libri eorundem et nonam Communem Sententiam huius
sit quadrato ipsius DI, erit per primam partem eiusdem septimae quinti, et quadratum ipsius EI minus est quadrato ipsius GK. Quare per secundam
nonam Com. Sent, huius dictum rectangulum minus quadrato ipsius EI, Communem Sententiam huius linea EI quam linea GK minor erit. Rursus
totum videlicet sua parte, per eandem tertiam secundi, quod utique ab­ contentum ab LI, IE rectangulum aequale est ei quod ab MH, HB con­
surdum priore peius est. Non est ergo EI maior quam ipsa GK: ostensum tinetur rectangulo (ut superius ostensum est) et per tertiam propositionem
autem fuit quod neque etiam ipsi aequalis: necessario igitur minor quam secundi libri Elementorum Euclidis, quod quidem ab LI, IE continetur
ipsa est. Praeterea si BH minor non est quam EI, sit primo ipsi aequalis. aequale est contento ab LE, EI, simulque ipsius EI quadrato: quod vero ab
Quoniam itaque rectangulum quidem quod ab LI, IE aequale est quadrato MH, HB ei quod ab Μ Β,ΒΗ una cum quadrato ipsius BH. Habemus ergo
240 ipsius DI per tricesimamsextam / propositionem libri tertii Elemen. Eu­ duo rectangula, alterum ab LE, EI et alterum ab MB, BH contenta, quae
clidis, quod vero ab ΜΗ, HB similiter quadrato AH per eandem, quadratum duobus quadratis, linearum scilicet EI et BH, ita adiungi possunt ut unum
autem AH quadrato DI per tricesimamquartam propositionem primi libri quidem cum ipsis commune latus habeant, duo vero duobus in directum
Elemen. Eucl. et secundam Comm. Sent, huius est aequale: igitur per iacentia: atque duo haec aggregata invicem aequalia sunt, unum autem
primam Com. Sent, eiusdem primi Elementorum bis sumptam rectangulum unius rectanguli latus ex in directum iacentibus, ipsum nempe EL minus est
ab LI, IE contentum aequale erit contento ab ΜΗ, HB rectangulo. Sed per nonam Communem Sententiam primi libri Elementorum Euclidis uno
contentum ab LI, IE contento ab LI, BH est aequale per suppositionem et ex eisdem alterius rectanguli lateribus, utpote ipso BM. Igitur per se­
tertiam Com. Sent, huius: ergo et contentum ab MH, HB aequale erit cundam partem tertiae propositionis in principio huius libri demonstratae
contento ab LI, BH per primam Communem Sent, eiusdem primi Elemen­ quadratum lineae BH quadrato ipsius EI minus est. Unde per secundam

87 mg.: Determinatio.
88 mg.: Constructio. 90 mg.: Conclusio indirectae demonstrationis.
89 mg. : Demonstratio indirecta. 91 mg.: Demonstratio directa.
450 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZLS AD M IRAN DUM PROBLEMA 451

Communem Sententiam huius linea quoque BH quam ipsa Ei minor erit. propositionem tertii libri Elemen. Euclidis: ita nunc etiam iam dictae
Similiter autem de quibuscunque etiam aliis huiuscemodi lineis idem distantiae eaedemmet quae in plano erant permanentes continue versus
ostendetur. Perspicuum92 igitur est propositum etiam directe. Si93 itaque Coni basim imminuentur, nunquam tamen prorsus delitescent, quoniam
quotlibet circuli concentrici in eodem plano designati fuerint, et reliqua ut nunc quoque ipsa qA H ,D I,F K lineae ipsis AC et GH parallelis perpendicu­
242 superius. Quod demonstrandum erat. / Hoc94 itaque Theorema proposito lares sunt et circulos tangunt, totaeque extra ipsos cadunt. Si itaque ipsae
nostro maxime necessarium ipse Rabbi praetermisit. Cum enim dixisset BH, EI, GK in Cono iterum extructo distantiae essent minimae per quas
distantiam EI minorem esse distantia GK, similiterque ipsam BH ipsa EI, et recta GH linea a punctis B, E, G in conica superficie iacentibus distare pos­
sic de caeteris eiusmodi: hoc porro nulla alia ratione probavit nisi quia sit: nimirum haberemus propositum iuxta hanc quoque imaginariam, sensu
(inquit ipse) maiores circuli minorem habent incurvationem, quae quidem perceptibilem, exemplaremque demonstrationem. Nam imaginaremur
ratio peius concludit quam illa Orontii ratio ab incurvatione circulorum planum quoddam ab ipsa GH linea Conum versus exurgere, ipsamque
ipsa quoque suscepta, quam in superioribus tanquam nil concludentem superficiem conicam penetrare, quod utique planum cum per Coni ver­
redarguimus. ticem non transeat, neque basi conicae parallelum sit, designaret nimirum
per quintam petitionem huius in conica superficie sub recta GH linea
DECIMATERTIA, ET ULTIMA quandam inflexam lineam, scilicet communem dicti secantis plani et
PRAECIPVI PROBLEMATIS conicae superficiei sectionem per signa B, E, G transientem, quae in
DEMONSTRATIO. eodem esset plano cum ipsa GH, propiorque ei fieret in signo E quam in
signo G, et in signo B quam in E signo, et sic in caeteris usque in infinitum,
Cum95 a nobis proximum Lemma Geometrice iam demonstratum sit,
si una cum toto Cono produci intelligantur. At quoniam ipsae porro distan­
proposito nostro nunc illud applicantes dicimus. Quod si Conum aliquem
tiae BH, EI, GK non sunt minimae quibus GH recta linea ab ipsis B ,E ,G
imaginemur esse supra basim suam compressum, totumque in plano iacen-
signis distare possit, ut in superioribus a nobis satis superque demonstra-
tem: tunc quidem omnes circuli qui in eius conica superficie basi paralleli
244 tum fuit, idcirco hic quoque opus est mi/nimas reperire distantias in eisque
erant in uno eodemque plano iacebunt, idem commune centrum circun-
propositum concludere ne paralogismus (quem superius diximus) commit­
stantes, quod utique prius Coni fastigium erat, latus vero Coni erit recta
tatur. Fortasse enim ad hoc respiciens ipse Rabbi cum dixisset duci tres
linea a centro communi concentricorum circulorum exiens, omnesque
illas lineas AH , DI, FK invicem parallelas et ipsis AC, GH perpendiculares:
circulorum ipsorum circunferentias ex altera parte secans, quemamodum
subiunxit, vel sint etiam remotiores a circulis quam ipsae perpendiculares
linea CA in superiori figura [Fig. 7.631, quae etiam hic tota repetenda est
quales scilicet in sequenti secunda figura [Fig. 7.64] sunt ipsae AL, DM,
praeter lineam GLM, ut cuius hoc loco nullus sit usus futurus. Si96 igitur
FN, quae quidem invicem parallelae sunt, non tamen ipsis AC, GL paral­
Conum hunc ita compressum, totum que in uno plano prostratum, ad suam
lelis perpendiculares: sed a circulis remotiores quam ipsae AH , DI, FK
pristinam montosam tumosamque rotunditatem restitutum esse intel­
perpendiculares putans fortasse Rabbi Moyses distantiasB L , EM, GN esse
lexerimus, et eius latus, nempe ipsam AC a basi usque ad verticem iam
in Cono rotundo minimas illas distantias quas quaerimus: vel has quidem
ascendisse, proculdubio statim inveniemus etiam rectae quidem AC paral­
esse maiores (ut etiam in plano sunt) ipsis BH, EI, GK distantiis, ipsas vero
lelam, ipsam scilicet GH, simul cum perpendicularibus AH, DI, FK cir­
BH, EI, GK esse dictas minimas. Nam in ipsis etiam BL, EM, GN verum
culos tangentibus, elevatam in aerem extra Conum et remotam a punctis
est dicere quod distantia EM minor est quam GN, et BL quam EM. Cum
B ,E ,G conicae superficiei iuxta distantias BH, EI, GK quanvis non mini­
enim ipsae AL, DM, FN invicem parallelae sint, sunt autem ipsae etiam
mas. Quae97 profecto distantiae quemadmodum prius in Cono supra basim
AC, GL. Igitur per tricesimamquartam propositionem et primam Com­
suam toto compresso continue versus ipsam basim, scilicet ultimum
munem Sententiam primi libri Elementorum Euclidis bis sumptas LM
exteriorem AB circulum, imminuebantur (ut demonstratum fuit in proximo
aequalis est ipsi HI, et MN ipsi IK. Quare ablatis communi HM et com­
243 Lemmate), nec tamen unquam puncta Η, I, K / vel alia eis similia cir­
muni IN: erit per tertiam Communem Sententiam eiusdem bis sumptam
culorum circunferentias tangere poterant, quoniam lineae AH , DI, FK et
HL aequalis ipsi IM, et IM aequalis ipsi KN . Si ergo ipsis BH, EI, GK iam
omnes eis similes semper extra circulos cadebant per sextamdecimam
demonstratis inaequalibus distantiis ipsae HL, IM, K N aequales adjun­
92 mg.: Conclusio directae demonstrationis. gantur distantiae: erit per quartam Communem Sententiam eiusdem tota
93 mg.: Conclusio totius. 245 distantia BL minor quam tota EM, et similiter tota EM / minor quam tota
94 mg.: Rabbi Moysis defectus tertius. GN. Quod erat demonstrandum. Hoc modo igitur ipsae Rabbi ostendit
95 mg. : Expositio. concluditque propositum. Verum98 hic magnopere animadvertendum est
96 mg. : Constructio.
97 mg. : Demonstratio. 98 mg.: Conclusio ipsius Rabbi imperfecta in non minimis distantiis.
452 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZFS A D M IR A N D U M PRO BLEM A 453

quod ipsae quae a nobis quaeruntur minimae distantiae non sunt neque ip­ plano (ut in superioribus figuris) ipsis BH, EI, GK non minimis distantiis
sae BH , EI, GK, uti diximus: neque ipsae BL, EM, GN, ut" forte Rabbi continue decrescentibus et nunquam evanescentibus quae etiam in rotundo
putavit. Hae namque nec in plano nec in rotundo Cono minimae possunt Cono eandem subeunt affectionem, cum in extruendo Cono eaedemmet
esse distantiae, cum ipsis BH, EI, GK non minimis in utroque Cono distan­ permaneant, nec ullam varietatem suscipiant: si voluerimus virtute
tiis maiores in utroque Cono sint: in plano quidem, ut patet per nonam istarum ipsas quoque minimas in extructo Cono rotundo reperire distan­
Communem Sententiam primi libri Elementorum Euclidis: in rotundo vero tias idem patientes, ducantur a signis, B, E, G (et quicquid nunc dicam in
quia obtusiorem subtendunt angulum, contentum scilicet a recta GL et Cono rotundo intelligendum est [Fig. 7.66]) ad rectam GH lineam perpen­
qualibet ipsarum BH, EI, GK distantiarum, ut in rotundo Cono inspicienti diculares per duodecimam propositionem primi libri Elementorum Eu­
manifestum est. Non sunt itaque minimae distantiae neque ipsae BH, EI, clidis, quae sint BL, EM, GN quae profecto erunt minimae quas quaerimus
GK neque ipsae BL, EM, GN: neque ullae aliae quae commode in plano distantiae per decimamnonam propositionem primi libri eorundem; cum
possint ostendi. Quum etenim revera earum extremitates inter signa/?, H omnes aliae rectae lineae ab eisdem B, E, G punctis ad eandem GH rec­
e t £ ,/ e tG ,K caderent, quales sunt in sequenti tertia figura [Fig. 7.65] ipsae tam lineam ex quavis parte ductae maiorem per tricesimamsecundam
BL, EM, GN·, fieri vero nequaquam potest per sextamdecimam proposi­ propositionem eiusdem primi libri Elemen. subtendant angulum, nempe
tionem tertii libri Elem. Eucl. ut inter ipsas AH, DI, FK perpendiculares rectum. Hae itaque sic repertae minimae distantiae, quoniam per secun-
et circulorum circunferentias ullae rectae lineae cadant quae ipsas minimas 248 dam partem / vicesimae octavae propositionis eiusdem primi invicem
distantias in recta GH linea distinguant: idcirco male in plano ipsae verae parallelae sunt, et una cum parallelis BH, EI, GK non minimis distantiis et
minimae distantiae ostendi possunt. Ni forte quis exempli causa inter ip­ cum recta linea GH (! NH) triangula claudunt similia per secundam partem
sas AH, DI, FK perpendiculares et circulorum circunferentias quasdam vicesimaenonae et tricesimaesecundae propositionis et tertiam Com­
rectas lineas falso designatas supponat, ut nos in praesenti figura fecimus, munem Sententiam primi et quartam propositionem et primam defini­
246 ut verum / minimarum distantiarum locum quem in rotundo Cono sortitae tionem sexti libri Elementorum Euclidis: igitur102 quemadmodum EI minor
sunt, in plano etiam ostenderemus. Haec99100 autem fuit causa quod ipse est quam GK, et BH minor quam EI, sic etiam EM minor erit quam GN, et
Moyses male minimas ipsas ostenderit distantias. Nam si ipsas quidem LB quam EM per nonam Communem Sententiam huius. Ac demum haec
ΒΗ ,ΕΙ, GK pro minimis accepit distantiis, supervacaneum erat ipsas etiam continua decretio similiter de omnibus aliis huiuscemodi minimis distantiis
BL, EM, GN (et loquor nunc in praecedenti secunda figura [Fig. 7.64]) demonstrari poterit, nec unquam tales distantiae minimae ita imminui pos­
continue decrescentes et nunquam delitescentes ostendere. Satis enim es­ sunt ut recta GH (/ NH) linea ipsam inflexam in Coni superficie iacentem
set per ipsas minimas distantias propositum concludere. Si vero ipsas tangat. Si enim inter dictas duas lineas, rectam scilicet et inflexam, maiores
BL, EM, GN distantias accepit (ut credo) tanquam minimas, dupliciter illae distantiae BH, El, GK semper sunt, ut superius ostendimus, tanto
deceptus est: primo quoniam ipsae tum in plano tum in rotundo Cono magis illae minores distantiae erunt, atque idcirco nunquam ille duae lineae
nequaquam minimae sunt, ut iam ostendimus: secundo quia si etiam essent propius semper sibi in eodem plano accedentes sese contingere poterunt,
minimae, paralogisimum maximum ipse committit dum per hasce, scilicet etiam si in infinitum versus Coni basim una cum toto Cono protractae
BL, EM, GN distantias concludit intentum; quandoquidem non eaedem- fuerint. Quoniam autem haec quae diximus aegre in plano conspici pos­
247 met ipsae / permanent in Cono rotundo, quae in plano Cono sunt, quod sunt, extruatur Conus ille quem superiores figurae in plano compressum
utique ipsis BH, EI, GK accidebat, sed variae fiunt: atque propterea non esse ostendunt, et fiant omnia sicut diximus, propositumque liquebit haud
recte concludit ipse Rabbi cum ex earum continua in plano diminutione et dissimiliter quam in nostra secunda demonstratione.
inevanescentia in rotunda etiam Cono eas continue decrescere, et nun­ Postremo103 vero docet Rabbi Moyses duas iam dictas lineas expeditis­
quam aboleri ostendit. Quum itaque minimae distantiae per quas recta sima quandam (/) via in eodem plano designare, quae talis est. Fiat Conus
GH a punctis B, E, G distare potest commode in plano ostendi non pos­ ex Rapa vel quadam alia tractabili facileque sectili materia, in quo ducatur
sint: alio quodam artificio in rotundo Cono reperiendae sunt, atque ex primum a vertice usque ad basim recta linea, qualis est ipsa ADFC [Fig.
ipsarum BH, EI, GK non minimarum (corr. ex minimatum) distantiarum 7.66]: deinde secetur Conus plano quopiam parallelo superficiei planae
tum in plano, tum in rotundo Cono perpetua versus Coni basim decre- 249 trianguli per axem / cuius trianguli latus est ipsa AC recta linea; planum­
tione, nullaque exinanitione, in ipsis veris minimis distantiis in rotundo que illud secans Conum sit verbi gratia Papyrus; post modum designetur in
Cono repertis propositum nostrum concludendum est. Repertis101 igitur in ipso Papyro linea BEG iuxta communem ipsius plani et conicae super-

99 mg.: Notandum. 102 mg.: Conclusio perfecta in minimis distantiis.

100 mg.: Error ipsius Rabbi. 103 mg.: Expeditissima via in Cono rotundo describendi duas lineas in eodem plano, nun­
101 mg. : Modus reperiendi minimas distantias in Cono rotundo. quam coincidentes, et semper sibi magis appropinquantes.
454 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZFS A D M IR A N D U M PRO BLE M A 455

ficiei sectionem, quae nimirum erit inflexa, Hyperbolicave linea, rationibus posse ut coincidant, licet in infinitum producantur; et unam istarum
superius dictis; postea ducantur a punctis A , D, F rectae linea e AH, DI, FK linearum esse rectam, alteram vero curvam.
ad angulos rectos ipsi AC, quae tangant circulos in signis A, D, F: et pro­ Supponamus2 hoc unum suppositum dicentes demonstrationem huius
ducantur quousque occurrant in Η, I, K punctis plano secanti Conum extra Theorematis explicari corpore solido in figura conica formato, cuius de-
conicam superficiem producto; et ducatur recta linea transiens per Η, I, K 260 Iscriptio sit huiusmodi. U t3 sit latum ab inferiori sui parte circulari, et in
puncta; ac demum inveniantur (ut docuimus) minimae distantiae G N , EM, angustum paulatim tendat, quousque terminetur in punctum. Et pars4lata
BL, quaesitumque factum erit. Recta enim HK et inflexa GEB semper istius Coni vocatur Basis. Et pars altera angusta sive acuta dicitur Caput.
propiores fieri et sibi nunquam occurrere superioribus rationibus demon­ Animadverte centrum circuli Basis esse e directo Capitis seu Verticis Coni
strabuntur, eruntque ambae in eodem Papryo plano designatae. Hoc modo per lineam rectam, ut si produxeris lineam a centro circuli basis ad ver­
ipse Rabbi consulit expedite propositum assequi posse. Veruntamen104 ticem sive caput Coni ad Apicem ipsius terminabitur. Et haec5 linea pro­
(ut mihi videtur) neque Rapa ipsa, neque lignum, neque aliud corpus tensa a centro Basis ad apicem Coni vocatur Columna, idest Axis, sive
opacum satis commodum esset ad extruendos Conos ipsos et faciendas in recta linea perpendicularis; quia veluti columna astat super centro Basis.
eis debitas sectiones et protrahendas lineas tum rectas, tum circulares, tum Et idcirco linea recta illa que nobis apparet in hoc solido corpore Cono
etiam mistas, punctaque omnia sectionum literis alphabeticis obsignanda, vocato, illa sane est linea quae dividit ipsum Conum in duas partes aequales.
propositi demonstrandi gratia. Nam ipsorum corporum opacitas esset Cuius divisio6 a puncto capitis ipsius Coni incipit et in punctum circun-
nobis impedimento, ut non omnes quas libuerit lineas protraheremus, vel ferentiae circuli ipsius basis terminatur. Quapropter hoc conicum corpus
protractas uno oculorum intuitu simul conspiceremus. Ni forte fortuna in duas aequales partes dividitur describendo hanc lineam praedicto
corpora illa conica diaphana atque omni ex parte conspicua essent; utpote modo. Verum linea curva7erit quaecunqae linea describetur prope rectam
Vitrea, vel Crystallina, vel cuiusdam alius eiusmodi materiae transparen- lineam dictam vel ad sinistras, vel ad dextras, quae dividet Conum in par­
tis. Quod tamen esset factu admodum difficile propter earum materiarum tem dimidio minorem, et necessario erit curva, cum non transeat per
fragilitatem. Melius105 igitur erit Conos ipsos construere lineis aeneis, vel punctum Apicis ipsius Coni, quia cum apponitur super figuram Hyper-
ferreis, vel etiam (si quis vellet) argenteis, et aureis: doctrina siquidem bolicam necessario magis distat a linea recta in principio ipsius Coni
haec adeo digna et nobilis est ut argento atque auro et alia (si qua esset) quam pars illius altera quae est ad basim, veluti infra demonstrabimus.
praeciosiori materia Coni ipsi confici mereantur. Nam si talibus materiis Sequitur ergo hanc lineam dividentem Conum in partem dimidio minorem
Coni extruantur, dubio procul puris materialibus lineis omnes in eis debitae esse curvam quia ducitur, sive accommodatur super figuram Hyper-
sectiones: necnon lineae rectae, circulares, et mistae, oculis uno intuitu bolicam. Et hoc sensui palam est. Probemus ergo sic dicentes: Sup­
repraesentari: punctaque omnia sectionum et linearum extrema literis ponamus8 inter hasce duas lineas, rectam scilicet et curvam, intercedere
alphabeticis super parvissima facta ex Papyro quadrata describi, ac demum superficiem unius palmi, et haec superficies sit aequalis quantitate palmi in
in cono tenaci quadam materia debitis locis applicari, vel etiam (quod omnibus coni lateribus ad faciem Coni. Hoc est ut semper superficies inter
melius est) in ipsa eadem metalli materia exculpi poterunt. Tot autem pro hasce duas lineas praedictas per distantiam unius palmi tum ad caput ip­
Dilucidatione quoque libelli Rabbi Moysis Narbonensis a nobis dicta suf­ sius Coni tum ad basim intercedat. Et dicamus manifestum esse distantiam
ficiant. maiorem inter hasce duas lineas intercedentem esse versus caput Coni.
Nam quo magis accedit linea illa curva versus caput, eo magis augebit
[Extract VIII, pp. 259-69] flexum in arcus formam; et quo magis flectetur informam arcus, augebitur
/ . . . . Modo reliquum est ut demonstrationem superius dictam ipsius quae inter ipsas intercedit distantia. Et erit chorda istius arcus maior quam
Rabbi Samtou expositoris iam dicti operis appositam in commentario partes chordarum caeterarum quae fient ad basim. Et probatio ipsius haec
eiusdem septuagesimi tertii capitis dicti Operis [Director dubitantium], erit. Quod si fingamus produci perpendicularem a dicta recta linea ver­
qualiscunque sit ex Hebraico in Latinum sermonem fideliter conversam sus lineam curvam, et applicamus super dictam lineam curvam perpen­
hic subiungamus, nostrumque de ipsa iudicium afferamus. dicularem parvam, quae attingat uno sui extremo lineam perpendicularem
a dicta linea recta protensam, et altero sui extremo attingat lineam curvam:
Rabbi Samtou Demonstratio.
Dari1 duas lineas inter quas a principio sui ortus sit distantia deter­ 2 mg.: Expositio.
minata, et quo magis procedant, sibi invicem proximiores fieri: et fieri non 3 mg.: Conus quid sit.
4 mg.: Basis, et sumitas Coni que sint.
104mg.: D e m a te r iis C o n o r u m e x tr u e n d o r u m .
5 mg.: Axis Coni qui sit.
6 mg. : Lateris Coni, et Trianguli per Axem Coni descriptio.
105 mg. : M a te r ia e c o m m o d a e ad e x tr u e n d o s C o n o s.
7 mg.: Hyperboles descriptio.
1mg.: P r o p o sitio . 8 mg.: Constructio, Determinatio, et Demonstratio primae partis
456 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZFS A D M IR A N D U M PRO BLE M A 457

et fingamus extendi chordam a principio illius lineae perpendicularis super duabus illis lineis demonstrat, de quibus etiam Orontius et Innominatus
261 lineam rectam appositae; a capite ipsius, quo coniungiltur cum linea Autor eam obscure, imperfecteque demonstrarunt. Quorum Demonstra­
recta; et haec extendatur versus caput perpendicularis parvae lineae super tionem nos in superiori nostro Opere in fine tertiae nostrae praecipui
lineam curvam constituta, qua parte curvae coniungitur. Sequetur ergo Problematis Demonstrationis instauravimus atque dilucidavimus. Sunt
absque dubio chordam maiorem esse in hac parte prope Coni sumitatem autem duae illae lineae latus hyperboles et latus Trianguli per Axem Coni,
quam in caeteris Coni partibus ad basim, eo quod diameter sive sub­ ambae in eadem conica superficie, sed non in eodem plano iacentes.
tendens angulum laterum quae sunt lineae descriptae super lineis dictis Demonstratio vero ipsius Rabbi Samtou non est eadem cum iam dicta
est maior quam in caeteris partibus inferioribus, quia est diameter sive Orontii et Innominati Autoris Demonstratione, sed ab eis omnino diversa.
subtendens angulum duorum maiorum laterum. Et manifestum est dia­ Quam eo quidem in loco nostri Operis posuissemus, si instauratione digna
metrum, sive subtendentem maiorum laterum maiorem esse quam dia­ nobis visa fuisset. Quoniam autem omnium de hac re demonstrationum
meter laterum minorum. Quod si produxerimus dictis lineis similes lineas imperfectissima est, et multis obscuritatibus, defectibus, superfluitatibus,
versus basim ipsius Coni, tunc erit chorda illa minor; quia linea perpen­ gravissimisque erroribus undequaque referta, idcirco noluimus eam inter
dicularis parva quae ducitur minor quam illa recta linea existit, et magis nostras Demonstrationes interserere; sed hoc in loco potius eam transferee
propinquae lineae illi curvae, quia quo magis tendit versus basim, eo magis voluimus cuiusmodi in Hebraico sermone legitur, ac imperfectiones eius
minuit curvitatem. Quod si apponamus chordam super dictas perpendicu­ breviter ostendere. Primum itaque cum iam dictam admirandam proposi­
lares parvas versus basim positas super dictam lineam rectam et curvam, tionem in modum Theorematis ipse Rabbi proposuerit, eius expositionem
quae chordae ducantur a principio unius perpendicularis in alterum, tunc aggrediens declarat breviter, et satis obscure, et impropriis Geometriae
chorda haec erit minor quam chorda illa quae facta fuit versus Coni terminis, ac phrasibus (ut legenti Geometriae non ignaro perspicuum erit)
sumitatem, quia latera harum linearum sunt parva, et hac de causa nec es se quasdam definitiones videlicet Coni; et Basis, et Summitatis, et Axis eius;
erit ut sit diameter sive subtendens minor. Et idcirco erit chorda minor necnon Lateris Coni, et Trianguli per Axem; ac demum Hyperboles. Quas
quam chorda versus caput Coni producta. Quapropter ex hoc patet omne nos in principiis superioris Operis nostri dilucide tradidimus ac declara­
magis curvum maiorem habere chordam: et quo minus fuerit curvum, vimus. Postquam autem iam dictas definitiones tradidit, mox demonstra­
minorem. Idcirco quo magis fuerit distantia inter duas illas lineas, eo tionem suam aggreditur, ibi [Probemus ergo sic dicentes, etc].113 Quae
maior in illa linea curva erit curvitas. Et quo magis illa linea curva versus quidem quam obscura, confusa, atque imperfecta sit, licet cuilibet in Geo­
basim ibit, rectitudinem aliquantum emulabitur. Et huic argumentum erit, metria versato Lectori perspicuum esse facile poterit: nihilominus quas­
quod linea illa curva versus Coni caput magis distat a linea recta quam dam eius obscuritates, confusiones, imperfectiones, falsitatesque ad-
caeterae partes illius lineae curvae versus basim. Si supposueris enim notabo. Primum itaque Demonstratio ipsa obscurissima est, quoniam
produci lineas perpendiculares a linea illa recta in curvam illam lineam, 263 nulla / ipse Rabbi figura usus est. Et quanvis in principio Expositionis
videbimus sensu ipso, et ratione quoque, lineam curvam magis distare a dixerit se supponere quod Demonstratio huius Theorematis corpore
recta prope Coni sumitatem quam prope basim. Quare quo magis pro­ Conico explicanda sit, nihilominus debebat eam saltem in figura plana
cedant iam dictae lineae per Conum versus basim, eo propiores inter se corpus conicum representante declarare, quemadmodum caeteri omnes
fient, quia curvitas inter ipsas intercedens minuetur: et9 nunquam coin­ Autores fecerunt. Verum ut defectus, et imperfectiones, erroresque
cident, quoniam inter ipsas superficies permanens palmi distantia e re­ gravissimos huiusce Demonstrationis facile possimus ostendere, neces­
gione intercedit, sicuti supposuimus. Quod10 si coincidant, falsum seque­ sarium nobis est in figura conica plana demonstrationem ipsam explicare.
tur, esse superficiem illius lineae parvae curvae aequalem superficiei Quicquid autem in ipsa plana conica figura dixerimus, in Cono rotundo
lineae maioris rectae. Et Deus nos ab erratis vendicet. corporeo intelligantur. Sit11 igitur Conus Rectus ABC [Fig. 7.67], in quo sit
Latus quidem Trianguli per Axim recta AD linea; Latus vero Hyperboles
EFG sit curva linea EF. Ait ergo Rabbi [Supponamus inter hasce duas
62 / Sequitur Commentarius Francisci Barocii
lineas, rectam scilicet et curvam, intercedere superficiem unius palmi, et
Haec est itaque Rabbi Samtou Demonstratio, qua conatur ipse quoque haec superficies sit aequalis quantitate palmi in omnibus Coni lateribus ad
demonstrare dari duas lineas, alteram rectam, et alteram curvam, quae quo faciem Coni, hoc est, ut semper superficies inter hasce duas lineas praedic­
magis producuntur, eo proximiores sibi invicem fiant: et nunquam coin­ tas per distantiam unius palmi tum ad Caput ipsius Coni, tum ad Basim
cidant, etiam si in infinitum protrahantur. Quam porro affectionem de intercedat]. Quod autem his verbis dicere vult, tale est. Supponamus hasce

9 mg.: Demonstratio secundae partis. lla The brackets here and below on marg. pg. nos. 263, 265, 267, and 269 are Barozzi’s.
10 mg.: Conclusio. 11 mg.: Expositio.
458 ARCHIMEDES IN THE MIDDLE AGES EXTRACTS FROM BAROZZTS AD M IRAN D U M PROBLEMA 459

264 duas proposi/tas lineas, rectam scilicet AD et curvam EF, esse designatas caput Coni. Nam quo magis accedit linea illa curva versus caput, eo magis
a communibus sectionibus superficiei conicae et duorum planorum sibi augebit flexum in arcus formam; et quo magis flectetur in formam arcus,
invicem parallelorum secantium Conum, alterum quidem eorum per Coni augebitur quae inter ipsas intercedit distantia. Et erit chorda istius arcus
verticem, faciendo Triangulum per Axem cuius latus est AD recta linea, maior quam partes chordarum caeterarum quae fient ad basim.]. Haec est
alterum vero non per verticem, faciendo Hyperbolem cuius latus est ipsa eius Determinatio, cuius sensus (quamvis ex eius verbis aegre eliciatur)
curva EF linea. Et distent haec duo parallela plana ab invicem unius palmi talis est. Dicimus distantiam maiorem esse inter hasce duas lineas, scilicet
distantia. Ecce igitur quomodo candidissimam hanc suppositionem ipse 266 AD et EF, interiectam versus caput sive summitatem Coni quam / versus
Rabbi verbis obscurissimis falsisque verborum sensibus proposuit. Vide­ basim ipsius, ut proposuit prima pars Theorematis; quam videtur con­
tur12 enim dicere quod supponatur inter illas duas, rectam scilicet et cur­ firmare prius quadam probabili non Geometrica ratione desumpta a maiori
vam, lineas interiectam esse superficiem unius palmi ad faciem Coni, ipsius curvae lineae incurvatione. Deinde vult probare idem alia ratione
nempe in ipsa externa conica superficie. Quod porro falsum est, quoniam Geometrica desumpta ab inaequalitate quarundam chordarum subtenden­
latitudo superficiei conicae interiectae inter iam dictas duas lineas, non tium quosdam arcus ipsius inflexae Hyperbolicae lineae. Quam14 inaequa­
debet esse, neque supponi, eiusdem semper quantitatis in omnibus Coni litatem determinans proposuit illis verbis [et erit chorda istius arcus maior
partibus (vel, ut ipse improprie ait, lateribus) tam versus summitatem, quam partes chordarum caeterarum quae fient ad basim.]. Hoc est chorda
quam versus basim ipsius: cum revera superficies ipsa conica inter illas arcus Hyperbolicae lineae versus summitatem Coni existens erit maior
duas lineas interiecta, latior versus Coni summitatem quam versus basim quam caeterae chordae arcuum Hyperbolicae lineae versus Coni basim
265 ostendenda sit; si / nimirum ipsae duae lineae magis sibi semper versus existentes. Quamvis ipse male, et confuse loquutus chordas ipsas in­
Coni basim appropinquari debeant, ut proponitur. Si autem intelligit super­ feriores, vocaverit partes caeterarum chordarum. Hoc autem perperam,
ficiem planam inter duas illas propositas lineas interiectam unius palmi et obscure, atque confuse proposuit; quoniam adhuc non declaravit quae­
latitudinis, esse planam superficiem, quae secet planum Trianguli per nam istae arcuum inaequales chordae sint, et quomodo reperiantur; ex
Axem in latere ipsius Trianguli et planum Hyperboles in recta linea quarum inaequalitate propositum demonstraturus est. Inaequalitatem
secante duobus in punctis latus ipsius Hyperboles; quippequae plana autem harum chordarum demonstrat construens,15 atque arguens hoc
superficies procedat per lineas rectas ductas ab aliquibus duobus punctis modo, hisceque verbis [Et probatio ipsius haec erit. Quod fingamus pro-
ipsius lateris Hyperbolici perpendiculares ad propositam iam dictam rec­ 267 duci / perpendicularem a dicta recta linea versus lineam curvam] quae
tam lineam, sive Latus Trianguli per Axem Coni. Hoc etiam supponi non scilicet sit in figura a nobis apposita, ipsa recta HI linea [et applicemus
debet, quoniam falsum est, et contrarium ei quod est praecipue demon­ super dictam lineam curvam perpendicularem parvam quae attingat uno sui
strandum, nempe iam dictas perpendiculares per quas latitudo ipsius extremo lineam perpendicularem a dicta linea recta protensam et altero sui
planae superficiei procedit fieri semper eo minores quo basi propinquiores extremo attingat lineam curvam]. Quae quidem parva perpendicularis,
sunt. Verum neque etiam de plana superficie intelligere potest, quae sit quam superius chordam vocavit, sit ipsaIK perpendiculariter erecta super
perpendicularis sive erecta ad duo iam dicta parallela plana Conum secan­ ipsam HI. [Et fingamus extendi chordam a principio illius lineae perpen­
tia; quoniam licet de hac superficie plana verum sit, quod eius latitudo sit dicularis super lineam rectam appositae a capite ipsius quo coniungitur
eiusdem quantitatis, utputa unius palmi in omnibus suis partibus, nihilose- cum linea recta, et haec extendatur versus caput perpendicularis parvae
cius verba eius nullo modo de hac superficie intelligi possunt. Cum haec lineae super lineam curvam constitutae qua parte curvae coniungitur.]
superficies non intercedat inter iam dictas duas propositas lineas, quoniam Quae demum erit recta HK linea, quam ipse improprie chordam nominat,
ipsas ambas attingere minime potest, sed inter iam dicta duo parallela plana cum nullius arcus chorda sit. Potius enim ipsa IK parva perpendicularis ab
intercedit quae ipsas duas propositas lineas efficiunt. Quare nullo pacto ipso vocata chorda hic quoque quemadmodum superius vocanda erat, quia
illa verba accommodari possunt ut veritatem exprimant; licet ipse seipsum chorda arcus Hyperbolici est. Sed ipse (ut inferius etiam videbitur) con­
exponens eandem falsitatem repetat, cum dicat ipsam unius palmi sup­ fundit hosce Geometricos terminos, et eas rectas lineas, quae revera
positam superficiem inter illas duas lineas praedictas intercedere. Quum chordae sunt, non vocat chordas; eas autem quae chordae non sunt appel­
autem hoc modo suppositionem hanc proposuisset, sequitur determinans, lat modo chordas, modo diametros, modo subtendentes angulum, modo
primam Theorematis sui partem, inquiens13 [Et dicamus manifestum esse subtendentes latera, vel laterum. Hisce ita constructis, probat16 proposi-
distantiam maiorem inter hasce duas lineas intercedentem esse versus

14 mg.: Determinatio.
mg.: Primus error ipsius Rabbi Samtou. 15 mg.: Constructio.
mg.: Determinatio primae partis Theorematis. 16 mg.: Demonstratio primae partis.
 EXTRACTS FROM BAROZZI S A D M I R A N D U M PROBLEM A 461
460 ARCHIMEDES IN THE MIDDLE AGES

turn hoc modo, videlicet quod si fecerimus aliud triangulum inferius simile invicem coinciderent, quod suppositioni oppugnat: infert ipse quoddam
ipsi HIK triangulo iam facto, quale est ipsum LMN\ sequetur ipsam HK aliud absurdum a iam dicto diversum, his verbis [Quod si coincidant, fal­
esse maiorem ipsaZJV quia subtendit angulum rectum contentum a duobus sum sequetur, esse superficiem illius lineae parvae curvae aequalem super­
HI, IK lateribus maioribus quam duo latera LM, MN continentia angulum ficiei lineae maioris rectae.], quorum sane verborum (meo quidem iudicio)
rectum subtensum a linea ZJV. Quod autem la tera///,/X lateribusLM, MN nullus est sensus qui proposito applicari possit. Et ego ingenue fateor me
maiora sint, ex hoc confirmat, quia cum iam probasset illa probabili, nequaquam intelligere quaenam21 sit superficies illius lineae parvae curvae,
frivolaque maioris incurvationis ratione rectam HI maiorem esse recta neque illa superficies lineae rectae maioris, cui aequalis concludatur illa
LM: probat similiter eademmet maioris flexus ratione rectam quoque IK minoris lineae parvae curvae superficies. Prorsusque demum (ut hoc im­
maiorem esse recta MN. Ex quo probat etiam rectam HK esse maiorem perfectionum huiusce Demonstrationis sigillum sit) nescio quid sibi vir iste
rectaLN: ex quarum inaequalitate demum concludit duarum propositarum iam dictis verbis voluerit. Ipse viderit an Deus eum ab erroribus vendicarit,
linearum maiorem appropinquationem versus basim Coni quam versus quemadmodum in fine huiusce suae Demonstrationis illum depraecatus
Apicem. Haec est tota Demonstratio primae partis Theorematis ipsius est. Haec autem de hac etiam Demonstratione eiusque imperfectionibus a
Rabbi Samtou, explicata ab ipso multis obscuris, impropriis, confusis, per- nobis satis superque dicta sint.
plexisque verbis, ac supervacaneis repetitionibus (ut quilibet lector in ipsa Commentarii Francisci Barocii Finis.
eius litera cognoscere poterit) ab illis verbis [Sequetur ergo absque dubio
268 chordam, etc.] usque ad illa verba [quia cur/vitas inter ipsas intercedens
minetur]. Quae17 quidem Demonstratio gravissime peccat primum in hoc,
quod cum probasset iam illa sua probabili ratione perpendiculares rectas
lineas HI, LM, quae sunt minimae duarum propositarum linearum distan­
tiae, esse inaequales, non concludit ex earum inaequalitate propositum,
sed frustra probat etiam linearum H K, LN non minimarum distantiarum
inaequalitatem, ex qua male concludit propositum, sicuti etiam omnes
Autores fecerunt praeter Cardanum. Secundo18peccat haec Demonstratio,
quia non demonstrat nisi illa frivola et probabili ratione incurvationis
maioris inaequalitatem linearum IK, MN, ex qua concludit inaequalitatem
ipsarum H K, LN: ita ut tota haec Demonstratio in ipsa nullius momenti
ratione consistat. Et quod19 peius est peccat etiam magnopere in hoc, quod
non docet quo artificio ipse duae IK, MN chordae parvae inaequales
in ipsa Hyperbolis curva linea reperiendae sint, ut Geometrice et non proba­
biliter earum inaequalitas demonstrari possit. Verum20 cum ita primam sui
Theorematis partem ipse Rabbi demonstrasset, concludit etiam secundam
269 ex suppositione, quod scilicet non coincidant sibi invi/cem ipsae duae
propositae lineae quoniam (inquit) inter ipsas permanens superficies palmi
distantia e regione (hoc est aequidistanter) intercedit, sicuti supposuimus.
Quod ita intelligendum est uti superius diximus, perinde ac si dixisset,
quoniam inter ipsa parallella Hyperbolis et Trianguli per Axem plana
permanet semper eadem palmi distantia, sicuti supposuimus. Alioquin
falsum hic quoque quemadmodum superius diceret. Postea volens ex hoc
inferre absurdum sive inconveniens quod sequeretur si coinciderent; quod
scilicet duo iam supposita parallela plana Hyperbolis et Trianguli per Axem

17 mg.: Secundus Error ipsius Rabbi Samtou.

18 mg.: Tertius Error eiusdem.
19 mg.: Quartus Error eiusdem.
20 mg.: Demonstratio secundae partis. 21 mg.: Quintus Error ipsius Rabbi Samtou.