JULYAUGUST 2OOO us $7 50/cAN $10 50

;-tsffi
'**4*T*@**,*
ot?6*a
@
B-=ea
u%%q
B€.
Eg
€^:

-?-:1.1
;. l.t

,j
:i,, t: ^@-
ra!:':j
@
u':t
Springer

+
=
i=tr-=iirii=::::-=_:=;
_:i::
:i:i.i::::
i=:: :i+
:i= ::l
=:ii::.:: :ii-:=::=
'i:.:= i=j=i=::::='i':= Wffi\
GALLERY C

8,,, .

t#+i.:fii=:iffi€1:+i{i ,:i;:-, -;- =- ; j!.!.i;.:-;:-..:.1,,,1..aj==E

Oil ot t*ttrl, 15 7,,€ x I 8 31.1, Cil I ol L,.lsil \\'illian1 nil.l BLil1it.(. Cltrtl /.r Cdr brr. Jr. C lr0[] I.,.ir.

The Mounted Acrobats (1825 or after) American

THE THREE-RING CIRCUS HAS DELIGHTED posed points on the ring that the above horse is circling
I audiences for over a century/ but not everyone enyoys where a continuous function takes equal values. We're
watching animals perform under the Big Top. There are sure you'll never look at certain circus acts the same
two diametrically opposed points of view as to whether way again after you have explored this theorem in greater
or not the use of animals in the circus is a cruel prac- detail. Turn to page 16 to learn more about the "Borsuk
tice. Coincidentally, there are also two diametrically op- theorem."
-Ulam
 A
JULY/AUGUST 2OOO
TU VOLUME 10, NUMBER 6

FEATU R ES
#*;Crut-t.i 4 Tite Work
Sell-similar tno$aic$
by N. Dolbilin

10 Musical Notes
Catr[inu tlp olt rays altd ttuaue$
by A. Stasenko

10 Opposing Points
T[e Bonsulr-Ulam lheol'sln
by M. Krein and A. Nudelman

nn
Cover afi by Vasily Vlasov ll l-amous l-lgures
"What came first, the chicken or the AlUehraic altd IrattscEltdsltlal ltulnIBrs
egg?" is a question that has tran- by N. Feldman
scended time. But our fowl-feath-
ered friend on the cover seems more
concemed with certain transcenden-
tal numbers than debating his ori-
gin. If you'd like to find out more
DEPARTMENTS
about the topic that has just passed
this chicken's 1ips, turn to Algebraic
and Transcendental Numbers on
3 Bl'ainleasBr$ 4l AI t[B Blacfiloand lll
The enigmatic magnetic force
page 22.
l[ lloul Do You Fiuune?
44 ln lhe lal
2l llappeninUs Can you carry water
in a sievel
Bulletin board

28 l(aleidoscope 46 lookin0 Baclr

Ernst Abbe and "Carl Zeiss"
Ceometric surprises

30 Physios Conlesl 50 Cl'isscl'oss Science

Indcxed in Mttgttzine Article
Suntnt orie s, A c: aclentic Ab s ft act s,
Batteries and bulbs
Il [n$ul8ts, llinls & s0lutions
Ac ttdentic clt, 35 Al I[s Black[oard I
S
Se ar V o c at ictn r:tl
ettc.lt, MasterF ILE, and (]ener ctl Sharing a point 55 lnlol'malirs
Stience Sottrce. Availerble ir-r Chores
microforrl, electronic, or paper 38 At lhs Blailhoal'd ll
format from lJniversity Micro- The little house on the
fih-ns International. tundra

OUAlllTUll4/[ONIINIS
 QntLntunt iISSN 1048-8820) is
published bir.r.ronthly by the Na-

OUANTUM
tional Scier-rce Tcachers Association in coo1r-
cration \4.ith Sprrngcr-Verlag New York, Inc.
Volun're 1 1 ((r issucs) rvlll be published in 2000-
2001. Quantrurr contains authorlzed English-
language translations frorn Kvalt, a physics TH,!€,,,,.,']M,A-AZr'tN:,Et'.O.F M'A,.T.H AN D. SC lE NC E
ancl rlathcrlatics mzrgazine published by
lAQr,,,,,r,,', BtLre.ttt'Morcorv. Rusrial, a. fr oubtioatlon of the, Ntilianel. Sgienee Teachers Assaeiatian (/VSIA)
[\ ncll:rs originrl rnatcriaI irr trglish.
Editorial offices: NSTA, 1840 Wilson Boule- ':,,.. :, 1::',::l:::.::... ',:..:..:.:'.:..:, .;,..;y1.:C66.iune,li:O.ti..tyith.
vard, Arlington VA 22201-.3000; telephone:
l--100; clcctr,,itic tnail: A the American Association of Physics Teachers (AAPT)
,-01' 24
& the National Councit af Teachers of Mathematics (NCTM)
.iuantum,Tn:tir.org, Production ofiices:,ffiq,
\l,rin!.cr-Vcrlrtg New \url<. Irtc., l-i ,.-€)-! Natlonal Science Teacbcrs 4ssoqiaiiqn..r's
' l.'r" ':
qP L.;r;;.':1iill',:1,,$ha:,,niii!lfuWth9
Fittlr Ar errrc, Ncw Yurl< NY 100 l0-
7858.

Perioclicals postagc paid at Nerv York, NY, and

Publisher
, Executive Director, NSTA
additior-ra1 mailing of{ices. Postmaster: send
address change s to. QLtall tLln7 | Springer-Verlag Associate Publisher
Nerr' York, h-rc., fournirl Fulfilhnent Services Sergey S. Krotov, Director, Quantum Bureau,
De;.av1L.r-r,, P. O. Box 2485, Secaucus Nf Prqf e,qSof r.o.-f,:?.rhrsics, Mqseow State:.Ijniversity
0?016-llS-.. Copyright O 2000 NSTA. Printcd :..,.'li:,r r' :.:1.'l

in U S.-\
Subscription Inlorrnation:
\r.,r':,:, -,,r,.: -. ,r: Student rate: $ 18; Per-
sor-r...i:.,.:. rr rr!rrllentl: $25. This rate is
ar ar .:.r,: : r-r.r'. -.ir,ra1 subscribcrs for per-
sLrn.'.- -r:. :- . :r ::r Sprrr1ger-Verlag New
Y,,,t.,. I::: '...,,=:: :-.-.: :r personal check of
charq. S:--.-:-:'-: :-> -11': .rlltCred rvith pre-
f11\lIall: l1-' -:,.:-:--::' naL rate: $56.
Srn.sl; I..,-. l:.:-. -'- : . Lates ir-rclude

,r. l'\:-- . -:- -.'--.rntion

Irlrlrl.et i- llll.'1. . ....-_ ..-.. :. -.

gttr rvith nexr l.tthlr.hc.l r---.- -- ':.*:-
nriry b( rcrlLl(stcdl. Bttll< t.ttc-: : -:...- -
r[c a\irililhle. Mrrrl order,rnJ 1.tr':':-:.: :
Sprlnger-Verlag New York, Inc., i.,-::--.
Fuliillmelrt Scrvices Department, PO B. r Special Prajects Coordinator
2485, Secar-rcus, Nl 07094-2485, US-{ Kenneth L. Roberts
Telephonc: 1t800) SPRINGER; {ax: (2011
Lditortat Advisar
14x --i i0i j e-nriri I : rrr.t.r'r \',/ 5prirgcr-rt ) .Corn.
Timothy Weber
C)Lttside Nortlt Arnerica: Personal rate:
Please contact Springcr-Verlag Berlin at Editarial Consultants
[email protected]. Institutronal rate Yull' Danilov, Senior Researcher, Kurchatov Instltute
is USS57; airmail delivery is US$18 addi- Yevgeniva Motozova, Managing Editor, Quatrtutn Bureau
tionai (al1 rates calculzrted in DM at the ex- I nternation al Can sultant
change rate currcnt at the tilne of purchasel. Edward Lozansky
SAL (Suriace Airmail Listed) is mandatory
for fapan, Inclia, Australia, and New Zealand. -:-. -. i-r ' 'al:aser, Mawll:;Srr,nn"r-Vertas)
Customers should ask for the appropriate
pnce list. Orders may be placecl through your i7.lsory Board
bookscller or directly through Springcr- Bemard \'. h-houn Executire Oiricer', AAPT
Vcrlag, Postiach 3l 13 40, D-106:13 Bcrlin,
fohn A. Thorpe. Erecutive Director, NCTM
Cerrnany. George Betzsenyi, Proie..-r l==:i::.s oi Mathematics, Rose-Hulmat-r lnstitute
-\dr-ertising:
: l.--'-n logl'I\
F. : -.'
: : : : :::rtli\ cs : iWashrngton ] Paul Kur-rtzler
Arthur Eisenkraft, Sciel:. D=::.r-lent Charr, Fox Lane High School, NY
- : - : - '-r \lv Karen |ohnston, Ptoiesse-r :,: !::::-s. \lrth Carolina State University, NC
YorkJ Joe Antrno i212)
rllargaret |. Kenney, Pr..:.==:: .:.: ]1:.j:cnatics, Boston Collegc, MA
-1(r -'-'' : - Probst, Sprin.eer-\-er1ag
Alexander Soifer, Profes..rr .,--,: Ji::ie=arics. Umversity oi Colorado-
Cr-r:--:, . : : '' -191 Berhn, Ceti'n.rrty, C, ,. :-.: ., - -;- CO
telepi' :. :'--.-0 teler M-111. Barbata I. Stott, Mathem;itir. T;::::rr' fur-erdale High School, LA
Printeti . :: Ted Vittitoe, Retrrej ?l:r'.:i. Teacher larrish, FL
'-: Peter \-unovich, Capital Area Sc:c-:e s. ,\lath Center, Lansing, MI
Vr.it u-
quantum

JllLi r.i.:' .::.

 BRAINTEASERS

Jusl lol' Ihe lun ol it!

8296
Picture pe{ect, A teacher put a pentahedron (a polyhedron with five
faces) on the table. Two of the faces were triangles, and the other three
were quadrilaterals. |ean drew a top rriew of this polyhedron (see the
figure). Is this drawing correct?

8297
Connect the dots.Is it possible to mark six points in the plane, and
connect some pairs of them with nonintersecting segments/ so that
every point is connected to four others?

8298
Polls apart. One hundred persons-chemists and alchemists-attended a
conference. They were asked a question: "Which group is more numer-
ous here, chemists or alchemists (not including yourself)?,, The first 50
persons said that alchemists were more numerous. Now, it,s known that
alchemists always lie and chemists always tell the truth. How many
chemists attended the conference?

8299
Vexing hexagons. Six congruent regular hexagons are given. Cut three of
them into two parts such that the nine parts obtained (three hexagons
and six "halves") can be used to compose an equilaterai triangle.

o
8300 -U
FD

Wrong-way mercury. A thermometer is quickly removed from molten tin. o

At first the mercury
goes up/ not down! Explain this strange phenomenon. o
J
o
f
c
a
ANSWERS, HINTS & SOLUTIOIVS O/V PAGE 54 x

O U A III T U ll,l/B RA I lll T EAS I R S

 W

$ell-similar lno$aics
When the whole is the sum of its parts

by N Dolbilin

OSAICS MADE OF PIECES Each of these smali triangles is simi- fixed point. Motions of a thircl typc,
that are similar to the whole lar to the original one. In this sense, rcflections in a line 1, have rnfinitely
are cailed self-similar. One of the triangle is a self-similar figure. many fixecl points: the points of the
the first famous examples of a A figure F is self-similar rf it can line I itself. Thus, some rigid mo-
self-similar mosaic was constructed be cut into several figures, F1, F2, ..., tions have no fixcd points, others
by the English physicist Roger F-, each of which is similar to the have a single fixed proint, and still
Penrose. After cluasicrystals were initial figure. Since every constitu- others hirve infinitel\' many.
discovered in 1984, Penrose's pat- ent { is similar to F, there exists a This situation changes if we con-
terns became a generally recognized simiiarity transformation h, that sidcr a sirnilarin' trll-Isformation h
model used for analyzrng their geo- takes F onto F, : hJF) : {. The coef- that is not a rigrcl motion. This hap-
metrical properties. The mathema- ficients of these similarity transfor- pens if the similarrty coefficient, k,
ticians |ohn Conway and William mations are not necessarily equal, is not equerl to I In this case, the
Thurston found new and unex- but ali are less than 1. transformation Jr has a uniclue fixed
pected relations of these mosaics to point.l
other fields of mathematics. Self- $imilal'ity ll'altslot'tnafiiolt$ By virtue oi thrs remarkable fixed
similarity plays an important role in Recall that a similailty transfor- point theorem, any simrlarity trans-
modern fields of mathematics such mation is a transformation, h, ol a formation of the plane with a coef-
as dynamic systems/ fractals, and plane (or space) such that the dis- ficient k * I has a unique fixed point,
quasicrystals. tance dlx, y) between any two points say, a pornt O. For this reasotr, any
x and y changes by the same num- similarity trans{ormation can be
Sell.$milar fiUul'es ber k; represcnted as a dilation2 (centered
It is well known that the medial at O with some coefficient k), fol-
dlr, yl : kd(h(xl, h(y)|, k > 0.
lines (the lines connecting the mid- lThis remarkable theorem holds not
points) of a triangle split the triangle In the case -k : 1, the similarity just for similarity trans{ormation, but
into four equal triangles (figure 1). transformation is a rigid motion. for a more general type of mapping: a
Some rigid motions-for example, contraction mapping. This is a
translationby a nonzero vector- mapping f(xl (or which d(*, yl <
move any point to a different point. kdlh(xl, hly)) tor a certairO < k < 1.
Others-for example, arctation g of 2A dilation (or homothecy, or
the plane about a point O by a cer- homothety) is a transformation with a
tain angle-leave certain points center O and a coefficient k, which
where they were. Such a point (for takes any point P onto a point P' such
which g(O) : O) is called a fixed that O, P, and P'are collinear, and
OP' = k. OP. fik is positive, O is
point of the transformation g. Thus, outside line segment PP'. fi k is
Figure 1 a rotation about a point has a single negative, O is outside this segment.

4 rtlIY/AttGU$r 2ooo
 lowed by a rotation about O through the case of the right triangle, trans- Sphinx. The sphinx figure (hexo-
a certain angle, and perhaps reflec- formation coefficients differ in abso- mino) consists of six equilaterul tri-
tion through a line that passes lute va1ue. angles (figure 5) and can be decom-
through point O. In particular, such Domino. The domino figure con- posed into four similar copies (figure
a transforniation can be a simpie sists of two equal squares. It can be 51.
dilation without anv rotational easily cut into four pieces similar to Problem 3. Find the fixed point
component. itself (figure 3). for the similarity transformations
that take the original sphinx onto
Examples ol sell-similar liuures each smaller figure.
The triangle. We have seen that
any triangle can be decomposed into Sell-similar liUut,es and mmaics
four similar triangles by its medial Consider a " good" seif-similar fig- "
lines. What similarity transforma-
Figure 3
ure F; that is, one that does not con-
tions map the originai tiangle ABC Problem 1. Find the fixed point tain holes. Some examples are a tri-
into these similar parts? Three of for each of the corresponding simi- angle, parallelogram/ or any other
these transformations are dilations larity transformations. polygon that can be decomposed into
centered at the vertices of the tri- Problem 2. Find another way to equal polygons, each similar to the
angle, with coefficient Il2. The cut the domino into four pieces, originai one. Then an entire plane
fourth is the dilation centered at the each similar to the original. can be tiled by copies of F, without
intersection of the triangle's medi- Chair. The chair figure (tromino) gaps or overlaps. A covering of the
ans, with coef{icient -1/2. This last consists of three equal squares. It entire plane with non-overlapping
transformation can also ]:e described can be cut into four similar copies of tiles is called a mosaic or tessella-
as a dilation3 centered at the inter- itseif: F, F2, Fs, and Fo (figure 4l.Let tion.If all the tiles of the mosaic are
section of the medians with positive congruent, it is called monodical.
coefficient lf 2, foTlowed by a rota- How can we obtain a monodical
tion about the same point by 180". mosaic from a self-similar figure F?
The right triangle. A right tri- There are several ways. One way
angle can be decomposed into two seems to be the simplest, but in fact
similar triangles (figure 2). Let us conceals a tricky point. Consider, for
example, a chair F of a certain size
(figure 7al and decompose it into

trtu
Figure 4 four small chairs, as shown in figure
h1, h2, hr, the similarity
and ho be
transformations that transfer the
ADC "big" chait into the corresponding
parts. They have a coefficient of I lZ.
Figure 2 The dilations .h, and hrhave their ab
rnap LABC onto L,ABD. Consider centers at the points A and D, re-
the similarity transformation g1 that spectively.
is the product of the dilation h, cen-
tered at A and coefficient k, J e&l
AC, {ollowed by a reflection about
the bisector of ZBAC.It is not hard
to see that
C
|TAABC) = LABD. Figure 7
In the same way, we can see that 7b. Now double this pictrue to make
the transformation g, that maps each sma1l chair equal to the initial
LABC into ABDC is the product of Figure 5
one (figure 7c). Then cut each of the
the dilation centered at C with co- four chairs into four smaller chairs
efficient k, = BCIAC and a reflection (figure 7dl and double the resulting
about the bisector of z,qcg. picture once more. Repeating this
Note one difference between this procedure, we obtain an infinitely
example and the preceding one. In expanding chair-shaped domain
sThis observation depends on the consisting of congruent chairs.
fact that the centroid {intersection of This procedure has a peculiar
the medians) divides each median in name: deflation-inflation. Inflation
the ratio 2:1. Figure 6 corresponds to the increase in the

OUAIIIUlll/IIATURI
 Thus, the second-leve1 mosaic is
uniquely determined. Since the sec-
ond-level mosaic possesses the same
properties as the first, the self-simi-
lar decomposition of the plane into
chairs is strongly hierarchical.

Pnopel'lhs ol $l'ongly hiel,ancfiical

lll0saffi
Strongly hierarchical moserics
have a number of peculiar properties
that differ from the properties of
a weakly hicrarchical rrrosaics.
Figure B Aperiodicity. A moszric for which
'---'r-*'
I
.i-- I

--1--- ._,-.. -. ---t--' at least one translation exists such

1
t
that it maps the mosaic onto itself
I

1..
a
is callcd ltt'rioclit'. A: \ve cittt 5ee ilt
I I
,Aa A
tlte.xlrlplc of rltc \r.[ttarc lllo\i]ic,
t$ :

:
u.cak1)- hlerarchrcal rlosaics can be
t
i
perroclic. Any tile oi the s.p.itrre mo-
I I
j I saic can be translated rnto an1' other
._._..1-..-.
a-'' \ _-..-)--__
i
I
I

t
tile togethcr with thc unurc rnosaic.
C1 b U The most important property o{
strongly hierarchical mo,saics is that
Flgure 9 they are ttpefiodic, Let us prove that
size of the tiles; deflation corre- are also called hierarchical, referring such a mosaic cannot be periodic.
sponds to the decomposition of large to the hierarchy that exists between Assume that a translarion t exists
tiles into smaller ones. Taken to the the tiles of preceding and succeeding which nraps tllc cntire llr',.Jic onto
limit, this process (which is called levels. This hierarchy can be strong itself. Then r moves a rrle F, onto
the d-process) results in a monodical or weak. With a strong hierarchy, some other tile Fr. Because tire next
mosaic. However, the transition to the mosaic of each succeeding level mosaic consisting ot second-levcl
the limit is the tricl<y point men- can be constructed from the tiles of tiles is uniclucly determrned, the
tioned above. As a matter ol fact, the preceding level in a unique way. translation I also maps the sccond-
what is the limit here? Apparently, With a weak hierarchy, the mosaic level mosaic onto itsclf, Again, by
the d-process results in a sequence of tiles can be joined to obtain the tiles virtue of the fact that the second-
domains that increase in size arrdare of the succeeding level in several level tiles are uniquell- integrated
covered with congruent tiles. How- different ways. into thc third-level tiles, the trans-
ever, this sequence is not a sequence The strong or weak nature of the lation t that maps the second-level
of fragments that grows as we add hierarchy is mainly determined by mosaic onto itself also transfers the
new tiles to the mosaic constructed the figure F itself. For example, a thircl-level mosaic onto rtseli, and in
in previous steps. Nevertheless, it square produces a weak hierarchy. general it maps the rnosarc of any
can be proved that the plane can al- Indeed, a mosaic composed of 7<th ler,el onto itself, for any k.
ways be tiled by self-similar poly- squares (figure 9a) can be integrated The tiles oi thc kth level are 27' 1

gons. into its second-level mosaic in dif- times brgger than the tiles of the frrst
A mosaic constructed from a self- ferent ways. That is, a given square ler.el. Thus ii the riles oi the first
similar figure F is called self-similar A can occur among the second-1evel
if tiles in different positions (figures 9b
ar'd9cl.
o the tiles of this mosaic (let's call
However, the chair, sphinx, and
them first-level tiles) can be joined
domino figures produce a strong hi-
into bigger tiles (second-level tiles)
erarchy. Consider a chair-figured tile
that are similar to the first-level tiles
and such that the second-level tiies
in its corresponding mosaic. To-
gether with three other chairs, it
also constitute a mosaic (figure 8a);
r this "sequential integration" can be constitutes a second-level chair, and
each first-level tile unicluely deter-
performed for any level (figure Bb). a
mines three other complimentary
For this reason, self-similar mosaics tiles. Figure 1O

JUI.Y/AUGUSI 2OOO
 level contain a circle of diameter d seif-simiiar mosaics composed of any finite fragment in a mosaic of
(figure 10a), the tiles of the kth level chair-shaped tiles. Moreover, there the "Chair" family also occurs in
contain a circle of diameter Zk-t . d. are even innumerably many of them. any other mosaic of this family, and
For a large enough k, 2*-l . d will be To be more precise, two (infinite) it even occurs infinitely many times
greater thah the length of the trans- plane mosaics are considered identi- in each mosaic. Thus, although all
lation vector t (figure 10b). This ca1 if one of them can be matched strongly hierarchical mosaics of the
means that the translation by t maps with the other by a rigid motion of same family differ globally, they
a circle of diameter 2k - 1 . d onto a the plane. Otherwise, these mosaics look identical locally.
circle that overlaps the original one. are considered different.
On the other hand, these circles Let us explain why an uncount- [onulay$ lno$aic$
must belong to different tiles o{ the able set of self-similar mosaics can Recall that a self-similar mosaic
kth level and thus cannot intersect. be obtained from the chairs. Decom- can be periodic. Strongly hierarchi,
Why do they belong to different pose a chair into four smaller chairs ca1 mosaics are aperiodic. However,
tiles? No bounded figure can be and assign a number 1,2, 3, or 4 to despite their aperiodicity, the tiles
mapped into itself by translation; each of them, as shown in figure 7b. must have a finite number of differ-
since different tiles of the kth-Ieve1 Let a chair occur in a bigger chair ent positions up to translation. For
mosaic do not overlap, the circles in- under number a, at the first stage of example, with dominos, every tile
side those tiles do not overlap either. the d-process. At the second stage, belongs to one of the two classes of
Thus, we have arrived at a contra- this chair occurs under number a, in parallel tiles. With chairs, there are
diction. So, we know that all the chair of the second 1eve1, and so four classes of tiles.
strongly hierarchical mosaics are on. Thus, the mosaic that grows Problem 4. How many classes of
aperiodic. from the given chair determines a parallel tiles are in the sphinx hier-
Periodic mosaics provide a good sequence consisting of the numbers archical mosaic?
model for crystals, while strongly hi- 1,2,3, and 4. The same mosaic can It would be interesting to dis-
erarchical mosaics play an impor- grow from any of its other tiles, but cover whether a mosaic consisting
tant role in the study of quasicrys- the sequence produced wil1be differ- of identical tiles exists such that
tals. In contrast to crystals, these ent. Since the mosaic is composed of these tiles have infinitely many dif-
structures are aperiodic. In particu- a countable set of tiles, and there ferent orientations . In lgg2Conway
lar, the famous Penrose patterns (fig- exist innumerable different se- suggested a self-similar, strongly hi-
lluences/ there exist innumerably erarchical mosaic with tiles of equal
many different self-similar mosaics triangles having infinitely many dif-
consisting of chairs. ferent orientations. The underlying
Because there are infinitely many idea is very simple. Consider a right
strongly hierarchical mosaics. that triangle with legs equal to I and Z
can be constructed from a given self- and a hyporenuse of /S . This tri-
similar ti1e, these mosaics cannot be angle decomposes into five equal
enumerated by natural numbers as and self-similar triangles (figures
the elements of a sequence can be. lZa,l}b). The acute angle of the tri-
However, they can be enumerated angle is a: arctarl (1/2). This decom-
by real numbers. position induces self-similar mosa-
Assume that all mosaics from the ics called Conway mosaics. It can
uncountable family "Chair,, have easily be seen that in a Conway mo-
aheady obtained names in the form saic, for any integer m and for any
of real numbers. Suppose we want to
Figure 11
make a family album of these mosa-
ure 11), the best-known model of ics. Every mosaic is infinite, and it
quasicrystalst are a direct generaliza- cannot be placed on a finite photo.
tion of strongly hierarchical mosa- Thus, the portrait of each mosaic
ics. inevitably captures only a small part
All Alike. Let's consider another of it. Therefore, a given mosaic can
peculiarity of strongly hierarchical have rnfinitely many portraits. Now
mosaics: The mosaic of each suc- assume that a photographer has cho-
ceeding level can be uniquely recon- sen a portrait of each mosaic for his
structed from the previous level. album, but he did not 1abe1 the pho-
Therefore, it may seem that a tos in time, and instead wrote them
strongly hierarchical mosaic-for in afterward at random. However
example, the chair-is determined extraordinary it may seem/ he did b
uniquely. However, there are many not get any wrong. The point is that
Figure 12

OUA|llIU[l1l/IIAIUBE
tile, there exists another tile that is prisms whose orientations are dis- saic. Note that in this construction,
oriented at the angle mu with re- tributed everywhere densely in the the prisms that are turned at angles
spect to the initial tile. Since the set of all possible orientations. The of 120" and 90' about mutually per-
angle o is incommensurable with 27r orientation of a polyhedron can be pendicular axes with respect to each
(see probletn 5), any two triangles determined using a triple o{, say, other are identical. Suppose we con-
oriented at the angle mcr with re- mutually perpendicular vectors that struct a self-similar mosaic consist-
spect to each other cannot be paral- are rigidly attached to this polyhe- ing of the Conway-Radine prisms,
lel. Therefore, triangle tiles in a dron. The property of being " eyery- using the inflation-deflation pro-
Conway mosaic occur in infinitely where dense" means that for any cess. Then this mosaic will contain,
many different orientations. orientation of the polyhedron P and together with each prism P, aII
It turns out that a1l Conway mo- for any arbitrarily small e > 0, there prisms that are turned with respect
saics are self-similar mosaics with a exists a three-dimensional tile Pu to P by various angles obtained by a1l
strong hierarchy. Therefore, there such that the angles between the possible combinations of the form "
are infinitely many of them, and all vectors of its triple and the vectors
of them include triangles with infi- of the P-triple are less than e. ofrl . ofi2 . ofia . ofrA. .''61
.6mn .6mn-l
61 62 61 62 62 t
nitely many different orientations. The right triangular prism of
Problem 5. Prove that the angle height 2, whose base is the right tri- where g, and grare turns by 120" and
atctan (l l2) is incommensurable angle with legs 2 and 2"13 and hy- 90' about mutually perpendicular
with ru-that is, the equation n potenuse 4 at its base (figure 13a) axes.
arctanlll2l: mnhas no solution in serves as the initial object. The base We can use the three mutually
natural numbers lm, nl. triangle is decomposed into four perpendicular edges that meet at the
Let's rephrase the statement of similar triangles, as shown in figure vertex of the right angle at the base
the main property of Conway's mo- 13b. Therefore, the initial prism can of the prism as the orientation triple.
saic: For any possible orientation of be decomposed into eight similar It is relatively simple to prove (using
Conway's original triangle, A, and prisms, as shown in figure 13c. The the fact that the rotational axes of g,
any small positive number e, there two prisms, A and B, on the upper and g, are mutually perpendicular)
exists a tile { that is "almost paral- floor constitute a regular triangular that the set of different orientations
1e1" to A, up to afl accuracy of e. That prism. Therefore, this pair (taken as is infinite. It is more di{ficult to es-
is, the angles between the corre- a whole) can be rotated by 120' and tablish that the set of orientations is
sponding sides of the triangles A and then returned to its piace (figure everywhere dense. The proof re-
{ are less than e. In other words, the 13d). The prisms C and D at the quires the use of group theory.
tiles' orientations are distributed ev- lower floor constitute a right-angled
erywhere densely in the set of al1 parallelepiped with a square face of "Chao$" and sell-similar tnosaiffi
possible orientations. 1 x 1. This pair can be turned by 90" The game "Chaos" provides an
Later on, Conway (together with and returned to its place {fig. 13d). unexpected and simple method of
Charles Radine) constructed a mo- As a result, we obtain the con- obtaining sel f-similar mosaics using
saic in space consisting of equal struction of a Conway-Radine mo- the computer.
Let's look at the rules of this
game. We choose an initial set of
transformations. For the present
case/ we take those that map the
Conway triangle onto its five con-
stituent triangles. We denote these
b transformations by ft,, h2, hs, ho, and
h, (figure L2b).Let a random number
generator produce numbers 1,2,3, 4,
and 5. Mark an arbitrary initial point
xo on the plane.
Step 1. Assume that the random
number generator produced the
number 2. Set x, = hr(xs).
Step 2. Assume that the random
number generator produced the
number l. Set x, : hr(x1).
Step n. Assume that the random
number generator produced the
C d number ctr,where 0..: 1, 2,3, 4, or
Figure 13 5. Then set x, = ho,-(r,-r).

I JUIY/[UOlJ8T 2OOO
 This procedure produces the se-
quence of points Conway$ fl'ollem
; Let's summarize our results. If a
X = lxg,x1 ,...,xn,...j. polygon is self-similar, then copies
of this polygon can tile the entire
,), (,.
After two or three thousand steps, p1ane. If the mosaic consisting of
we will see Conway's tdangle dis- these polygons is strongly hierarchi-
played. The reasons for the phenom- cal, it is aperiodic-for example, the
enon are not simple, and are a sub- self-similar mosaics consisting of
ject for another artiele. chairs. However, it would not be
In the general case, let F be a self- correct to think that one can com-
similar figure, F = fi [_J... [J F-, and pose only aperiodic mosaics of
hr, ..., h^be the similarity transfor- chairs. Figure 15 shows a simple
mations that transfer F into its con- periofic mosaic consisting of chairs.
stituent parts F1, ...t F_. The
"Chaos" game makes it possible to
obtain the figure F on the display by
using the transformations hr.
To obtain the mosaic, let us paint
the point xn=ho,(x,_,) depending
on the value of crr. Suppose we as-
sign green to 1, red to 2, blue to 3, or-
ange to 4, andgray to 5. We paint the
b
point xn =ho.,(*r_r) the color that
corresponds to the number crr. Then Figure 15 Figure'16
we obtain on the display the color Thus, self-similar polygons can It is not difficult to verify that
picture shown in figure 14a. This is make up periodic mosaics along space can be tiled with such bi-
the flrst fragment of the self-similar with aperiodic, strongly hierarchical prisms without gaps or overlaps.
Conway mosaic. If we choose to mosaics. The construction of such mosaics is
paint x, the color corresponding to )ohn Conway raised a question: predefined in many respects. If we
on we will obtain a more detailed
_ 1/ Does a polygonal or even curvilinear want to tile space with such
colored portrait of the same mosaic plane figure exist such that it can biprisms, first we must construct a
(figure 14b). We will obtain a still produce only aperiodic mosaics? In- layer of them (figure ITal.In such a
more detailed portrait if we paint x, terestingly enough, an affirmative Iayer, all the biprisms are parallel to
the color corresponding to x, _, (fig: answer has recently been discovered each other. Moreover, the layer is a
ure l4cl, and so on. in the form of the so-called
S chmidt-C onway-Danzer biprism

(figure 16a). This biprism can be

AM
glued from the pattern shown in fig-
ure 16b. Note that this development
of the polyhedron is not of the usual
type. It contains a rhombus, which
is not a f.ace of the polyhedron, but
rather an auxiliary element of the
construction.
F}* The biprism is constructed as fol-

IY$N lows. First, take the triangular prism

ABCAppl, whose lateral face b

b i rrr--\/ >--.- ABBrAtis a rhombus (with an acute

angle u). Then, attachto this lateral
Figure 17
periodic family of biprisms. Then all
face the same prisrn turned by 180" of space can be filled with such lay-
about the diagonal of the rhombus ers (figure 17al. Each succeeding
face. Note that the lateral edges of layer is obtained from the previous
the second prism are at the angle cx one by turning it about the axis per-
with the Tateral edges of the first pendicular to the plane of the layer
one. A pair of such prisms attached through an angle equal to the acute
to one another constitutes the de-
Figure 14 sired biprism. CONTINUED ON PAGE 20

OUAlllTUlt4/TIATllRI
CalchinU tlp on ray$ and tnlatlo$

A rhapsody on wavelengths and the Stefan-Boltzmann law

by Albert Stasenko

HY DO MUSICAL INSTRU- string. Each of these standing waves

ments produce musical contains an integral number of half- i=z^Ii/'"
lz)
sounds and not the disorderly wavelengths:
noise produced by, say, bang- along the y-axis (i :
1,2, ...).
ing a spoon against apan? Because ,=ri,i=1,2,3,...
t- ,Xi i-l (1) An interesting feature of these
musical instruments don't generate standing waves is that the wave-
random sounds of every frecluency. The longest wavelength rs 21(i = ll, lengths 1,, and )", can describe either
They emit sounds of only certain while all other standing waves have independent waves or the same
frequencies-so-called mono chr o - smaller wavelengths and larger i. wave traveling at some angle cr rela-
matic ( "single-colored" ) tones. The number i indicates how many tive to the x-axis (the angled solid
If the frecluency of a sound is v, half-wavelengths exist on the string. lines in figure 3). In the latter case,
the corresponding wavelength in Now what if, instead of a string,
the air is l, = vfv, where v is the we have a square plate with at area r21r21
speed of sound in the air. The of 7 x I (figure 2)? Then we can have "'l -- aorc, i "'il -_--
sino i '

length of a piano string, or a pipe in the following numbers of half-wave-

a pipe orgatt, determines the wave- Iengths along each axis: Whenever physicist encounters
a
lengths of the sound generated. Fig- " sirt" arrd "cos" in a formula, there
ure 1 illustrates this idea for a i=2L
)"i
is a keen desire to square them and
string. It shows three variants (or add the squares together:
modesl of standing waves on a along the x-axis li : 1,2,...) and
cos2 u+sin2 0 = I = L(i, * ir\.
41''
It's clear that this equation can be
satisfied by more than one pair of
numbers i, i. For example, the bro-
ken lines in figure 3 show another
Figure 1
wave described by the same equa-
L
l
C
tion: (U

,) .l
1'+1' = -(*)' = R2. (3) oE
Cd

_o
This is the equation for a circle of
Figure 2 Figure 3 radius R in the i, i-plane (figure 4).

10 JUI.Y/AUGUST 2OOO
 $e
., i," ,'
r'i: r'n ' t;
I ":I 'h1-
,,:&
;f
i\
i.
J 1,iar. _,],;,:le-, i'''I
,&
. tl!:',S
, r$r
iii! c
1-' 'I :t tn
_ .. 't*--
.' +..,;
|
q -. r+.
"*
-f!

&

i .du-
'' ",#
tu
1a

1 lq*o

-'"*

OUAIlIIUII4/IIATURI
 length 1). Sound waves can now tion" sorresponding to the tempera-
travel along three axes (x, y, and z). ture 7 of the cube's walls. At this
We need to add a new equation to temperature the walls emlt and ab-
the system (2): sorb the same amount of energy per
ffi",
N k=2j-
1
unit time. Every cubic centimeter of
the space inside the cube is perme-
lk=1,2,...1,
iu 7, ated by electromagnetic waves trav-
and equation (3) becomes eling in every direction. These are
waves of every sort-ultraviolet
waves, visible light, infrared radia-
,)
1- +l-,) + R- =
11 -(*)" =
n)
_11
tion... Of course, all have wave-
lengths less than 21.
This is the equation for a sphere If such a "stove" is heated only to
Figure 4
in the i, i, k-coordinate system (fig- room temperature, it will be a very
However, the abscissa and ordinate ure 5). This space is also "granu- weak "radio station// that mainly
in this plot are integers; therefore, lized" and has a minimum volume gives off " w arrr," (infrared) radiation.
the area in this plot is " granuhzed," AV,ni,, : Li. Li. Aft = I . I . I : 1. An open-hearth furnace, on the other
and its minimum value is AS-in = Ai Therefore, one-eighth of a sphere of hand, heated to about 1,000 K pro-
. Ai = 1 (the shaded square). The ra- radius R (guess why we consider duces not only infrared radiation but
dius of this circle isn't measured in only one-eighth of it?)-the "first visible light as well. The wave-
meters-it belongs to the realm of sglavlftt the following iengths of the electromagnetic waves
dimensionless numbers. How many
-sontains
number of such "granules": in this rafige vary from a fraction of
such square "granules" could be a micron to a few microns/ so the dis-
placed in one quarter of the circle? (ys)(+na3la) tance between adjacent spectral lines
N=
(Why only one quarter? Because the
numbers i and i are positive.) Scien-
+(*)' $) (with wavelengths \ and \. r) given
by equation (1) is very small. There-
tists call this quarter-circle the "first The smaller l, is, the larger N is. fore, the set of wavelengths (or fre-
cluadrant." To answer the question, Recall that each " grarllle" (the set of quencies) can be considered continu-
we need to divide the area of the first three numb ers i, j, and k) describes ous rather than discrete. Equation (4)
quadrant by AS-,r, : I (in other an individual standing wave. There- says that the number of equilibrium
words, we can skip the division). fore, we have found the total num- electromagnetic waves filling the
Thus ber of modes-that is, the number of volume 13 of a "stove" is
standing waves with wavelengths
less than /-generated inside a cube o:::
* =xnzIl+ = n[1,1r.
(r/ with edge length 7. N(r)=
3),"=*u'=ru(u)
3c"
rs)
However, sounds are not the only
The tt-t' sign reminds us that it's not things that canbe musical. hr a sense, Every photon of frequency v car-
easy to cover a round expanse of electromagnetic waves and visible ries an energy hv (ft is Planck's con-
floor with square tiles. light can also be "musical." This stant). The equilibrium electromag-
Let's move into three-dimen- "musicality" is called color, and any netic radiation is sometimes called
sional space now and consider a spa- color is characterized by its own fre- the "photon gas." It's similar to con-
tiai figure lsay, a cube with edge quency v and wavelength )" = clv, ventional gas in that the photons
where c is speed of light. In this case, travel in all directions like mol-
a laser producing a monochromatic ecules. However, unlike molecules,
wave could be considered analogous the photons do not collide with one
Ai to an organ pipe or piano string. If the another-they only "strike" the
distance between the two parallel walls of the vessel (our "stove"). In
ffi^,
'r* ok mirrors of a laser is 7, the laser gener- addition, the speed of aI1 tfie photons
ates a wave whose wavelength is de- is the same (it equals the speed of
AV-t. scribed by equation (1). light), so physicists say they are dis-
Is it difficult to construct a cube tributed in frecluency (while the
filled with electromagnetic waves? molecules in a gas are distributed in
Not at all. We only need to pump speed). So what is the mean energy
everything out of a cube of volume of the photons?
13-air, water vapor/ carbon dioxide, First, let's consider a molecular
and so on. Will the cube be empty? gas. The number density of its mo1-
Paradoxically, no. It will be filled ecules is n and the mass of each
Figure 5 with so-called "equilibrium radia- molecule is m. It's known that the

12 JUI.Y/AUOIJST 2OOO
 mean kinetic energy of a molecule The denominator of this expression
of gas at a temperature 7 is propor- is n(-) - n(0) : 0 - ro : -no (here we
tional to ksT, where k, is Boltz- took into account that n(-) = 0-
mann's constant: that is, at an infinite height the den-
sity is zero). The numerator can be
mv2
- - kBT
2
(6) obtained by means of the identity

Thus the energy density of this gas

d(yn): ydn + ndy,
mtn
is
1

nmv-
from which we get

jro"til =j oO4-i,ou
ffio
2-
-nksT=P, Figure 6
where P is pressure.
000
* mBY
the largest portion of the energy is
In the Earth's atmosphere the carried by photons with frequencies
density of the gas varies with alti-
tude according to Boltzmann's for-
= r, i_o- ",ir* o(, ffi)( **)
0
near the characteristic frequency v-
- kBTlh. Thus the total energy of the
mula: kBT. equilibrium radiation in the volume
=o-no m8
_mgy of the stove can be found from the
n=noe kT Fina11y,
formula

This formula gives the characteris-

tic altitude at which the density of mgY = 11u7'
u -.rr(v"
) nu. -)n :
air decreases a-fold compared to that Equations (5) and (7) show that Correspondingly, the energy density
on the Earth's surface: the mean values of both the kinetic (energy per unit volume) is described
and potential energies of the gas by the formula
,T K,T
--e
RT molecules are of the order of ku7.
mg Mg This result is also true for any other u= ut,-.4 ' l9l
8.31.300 case where thermodynamic equilib- lr-
= ___________;_m = g.B I<m. rium is present for a large number of
"W(ksT)
29.L0-" .g.g For the sake of this formula-
chaotically moving molecules.
At this altitude the potential energy What can we say about the equi- u: aTa
of a moleculeis mgHu: kuT. It's cu- librium radiation in our "stove,,? covered a lot of ground.
rious that this value is also equal to According to equation (1), the long- -1sg/ys
That's because we derived it rather
the mean potential energy of mol- est wavelength is )'" : 21, so.for a than simply cite the Stefan-Boltz-
ecules in an isothermal atmosphere: stove with a side length 1- 1 m we mann 1aw. Note that these formulas
*sy k*T have ),*", - 2 m. This lies in the ra- describe only the dimensionality
= (or H, =i). l7l
dio range. The shortest wavelength and qualitative dependence of the
According to the mathematical can be anything down to that of X- values sought. The correct fre-
definition of a mean value, rays. Of course/ nobody uses a stove quency distribution of photons is
as an X-ray generator or radio sta- given by Planck's formula, which
: tion. However, for a given tempera- contains the dimensionless factor
msyN - *s)ydn(y), ture a certain value v- will be most I f (shvlkrr - 1). However, here we,re
0
characteristic ("will be found most oniy trying to estimate the physical
where Nis the total number of mo1- frequently among the photons," values (which is a very important
ecules in the column of air above a "will be most likely," "wlll be the step in exploring any problem). The
unit area: mean"-choose your own terminol- above coefficient is dimensionless,
ogy). We may expect that the coffe- so it will not change the dimension
sponding energy of equilibrium pho- of our formula foiu. \ /e're on the
x = lang1. ton radiation is also of the order of right track.
0
ku7-that is, We can take another step and cal-
This relationship can be rewritten as culate the energy radiated per unit
hv. - krT, (8)
time from a unit sur{ace area oI a
while the frequency distribution of body in thermal equilibrium (that is,
) ranlv) the radiation energy is described by at a constant temperature). As noted
mgy = mg+-. a bell curve (figure 6). This curve earlier, such a body radiates and ab-
shows that there is only a small sorbs the same amount of energy.
lo"u) amount of energy at very long and When we multiply the energy den-
0 very short frequencies. In contrast, sity u by its rate of propagation (the

0UAllIUliil/ttAIUIt l3
speed of light c), we get the density temperature. Indeed, the energy ra- from which we get
of the energy flow uc (its dimension- diated from the entire surface of the
:
ality is |/m3(m/sec) Il,m2. sec)). Sun per unit time is A_ 4Fc
The photons propagate in every di- 20(oTq"
Qs = 4s .anRr2 = oTlnDzr.
rection, ar.J tle of the photons 4.1.3.108
travel to the su{ace of the body (be- The disk of the Earth, whose areats ,r2
cause it's one of six possible direc- rR.2, receives only a small fraction z. to-4 5.67. to-8 . 60004
.

tions: forward, backward, up, down, of this eflergy I which is equal to

right, and left). Thus the density of. lxR]l I ftnL2). All this "intercepted" -10s m2.
the energy flow is energy is radiated into space from
the entire surface of the Earth anRl. This is almost ten hectares, to speak
1 k,a Equating the solar energy striking
61:-ug--n-Ta. (10) the Earth to the energy radiated by
in agricultural terms.
o c-h" Knowing the temperature of the
the Earth, we get Sun, we can now make the relation-
This is another form of the Stefan- ship (8) more precise. Since v. : cf )".,
Boltzmann law. -n2 we can rewrite it as
If we look a little deeper into the
or{nD!19 = arRlor!.
4fif hc
problem, we might guess that the N,l - , .
l<B
correct coefficient is Il4, rrot ll6. Note that we don't need the precise
But this is a fine point that doesn't value of the Stefan-Boltzmann con- We see that the product of the tem-
concern us here. Our aim was to stant, because it cancels out. Thus perature and the characteristic radta-
obtain not only the Stefan-Boltz- we have tion wavelength is some constant
mann law in the form u = uTa - q = composed of fundamental physical
o7a (where o is the Stefan-Boltz- 16 2 constants. Since the temperature of
mann constant), but also the rather ' 'i \D,lL)'
,S = rF+ )
'Ll-
' .% '
the Sun is about 6,000 K and the char-
important nontrivial relationship acteristic wavelength of visible light
between the proportionality factors Plugging 7e = 300 K and 0, = 0.$ = is about 0.5 pm, this constant is about
and fundamental physical con- 10-2 rad into this formula, we get 0.5 ' 104 m. 5000 K:3 . 10+ m'K.
stants: The relationship between the tem-
Tr= 2OTr: 6000 K' perature and the characteristic radra-
t.4 kf; Now that we know the tempera- tion wavelength is known as the
o- l*. ' ",
ture at the Sun's surface, we can cal- Wien displacement 1aw, which is
c3hJ czh3'
culate the area of a solar "sail" that one of the universal physical laws.
The precise value of the Stefan- can generate a force of one new.ton From this law it follows that at
Boltzmann constant is o = 5.67 . 10-8 to propel a spacecraft We'l1 assume room temperature all bodies emit
Il,:r,2 .s . K4). Note that these com- that the spacecraft and the Earth are electromagnetic radiation predo-
binations of fundamental constants traveling at the same distance I minantly at the wavelength )". :
could be obtained (as has happened from the Sun. The surface of the sail 3 . 10-3/300 m : 10 pm-that is, in
many a time in the pages of Quan- is covered by an ideally reflecting the infrared range. Therefore, this ra-
tuml by dimensional analysis, pro- layer. Every photon that hits the sail diation cannot be seen in the dark
vided the set of related values is perpendicular to its surface is re- (by the human eye). However, if
known (here they areh,k* and c).In flected back and thereby changes its there were such a thing as an "invis-
this article we took a step further own momentum by hvlc - (-hvlc) : ible man," an inJrared camera would
and showed how to obtain the for- 2hvlc. Since the energy falling on a detect this "warm" object quite eas-
mulas by playing with basic laws sail of area A per unit time is ily against the background room-
rather than basic constants. temperature radiation.
Clearly the dependence of q onT We can draw one more conclu-
is very steepr if we double the tem- Qr =oftno?-4-.
4nI]' sion from our reasonin'g. A very
perature, the density of radiation small stove or a speck of dust heated
energy increases by afactor of 16! (we showed earlier how to derive to a certain temperature under con-
Now that we've obtained such a such a formula), we multiply this ditions of thermal equilibrium
powerful law, it's tempting to use it value by 2 and divide by c to obtain should radiate a number of wave-
right away. For example, we can cal- the change in momentum of all pho- lengths limited by condition (1): the
culate the temperature at the Sun's tons striking the sail (which is the longest wavelength will be about
surface knowing only its angular propulsive force F): the size of the speck. Therefore, the
diameter 0s = Ds/I (D, is the Sun's spectrum of radiated frequencies will
diameter, I the distance between 2Q, 2 n nA
Earth and the Sun) and its mean
F=____-v =;oTjei7, CONTINUED ON PAGE 20

1 4 JU[Y/AUGU$T 2ooo
 HOW DO YOU
FIGURE?

ChallBltUs$
spheres that are tangent in pairs are layer of the plate with the largesr
also tangent to line m and plane L. charge is carefully disconnected,
tlllAIll The radius of the largest sphere is l. moved away parallel to the other
Find the minimum possible radius plates, and positioned as the third
M296 of the smallest sphere. layer on the outer side of the plate
Exacting equation. Let a3 - with charge Q. A very narrow gap is
cr -I left between this third layer and the
= 0. Find the exact value of the ex-
plate, which prevents any electric
pression Plrysics contact between them. What work
i/30l-4CI +a\[rc2+3,r+2. must be expended in this transfor-
P296 mation? (A11the actions are done at
Pendulumpuzzle. a distance so as not to influence the
M297 A wire arc of length I distribution of charges on the
Copper conundrum. Two pieces and radius R is sus- plates.) (A. Zilberman)
of metal have masses of I kg and 2 pended on two light 1(
kg, respectively. They are alloys of inextensible strings P299
copper with some other metals. The of the same length R. Spot-light. A plano-convex lens
two pieces were melted down and Find the period of made of glass has a refractive index
reformed into two new pieces. One small oscillations for L n:1.5 and a diameter D :5 cm. The
of the new pieces had a mass of 0.5 such a pendulum if the strings and radius of the convex spherical sur-
kg and was 40"/" copper, the other the arc always remain in the same face R : 5 cm. A broad parallel beam
had a mass of 2.5 kg and was 88% plane. (M. Yermilov) of light hits the flat side of the lens
copper. What was the percentage of along its optic axis. Calculate the
copper in each originai piece? P297 size of the light spot formed on a
Extr ateft estfi al ozone. According screen set behind the lens perpen-
M298 to some estimates, the mass of dicular to the incident beam. The
Linear thinking. Let the perpen- ozone (Or) in the Venusian atmo- screen is positioned so as to obtain
dicular to side AD of the parallelo- sphere is cx: 10-5 percent of the en- the smallest light spot for a narrow
gram ABCD passing through vertex tire atmospheric mass. What would beam (restricted by a diaphragm) di-
B intersect line CD at point M, and the thickness of the ozone layer be rected along the optic axis.
the perpendicular to side CD passing if it were collected at the planet,s (A. Ztlbermanl
through vertex B intersect line AD surface and had a temperature and
at point N. Prove that the perpen- pressure equal to that at the surface P300
dicular dropped from B onto diago- of Venus? The acceleration due to Con-fusion! The nuclei of deute-
nal AC passes through the midpoint gravity on Venus is g = 8.2 mf s2, and rium D(]H) and tritium T(!H) can
of segment MN. the temperature at its surface is ? = fuse according to the reaction D +
800 K. (A. Sheronov) T -+ fHe + lon to produce a neutron
M299 and alpha-particle (!He). In addition,
Circulu logic. A circle lies en- P29B each pair of interacting nuclei re-
tirely inside a given angle. Construct Two-plate special. A parallel leases energy E = 17.6 MeV. What
another circle, tangent to the first plate capacitor of capacitance C is energies are carried awayby the neu-
and to the sides of the given angle. composed of two large conducting tron and the aipha particle? The ki-
How many such circles are there? plates, each of which is a double netic energy of the nuclei before
layer made of electrically connected nuclear fusion is negligible.
M300 thin sheets of foil. The plates are (Y. Samarsky)
Think small Let a line mbe per- charged with charges Q and 2Q of ANSWERS, HINTS & SOLUTIONS
pendicular to a plane I. Three the same polarity. The outer foil ON PAGE 51

OUAlllTUltil/IIO[I DO YOU fIGURI? 15

The BorsulFUlam lheol'eln

Horsing around with continuous functions on a circle

by M Krein and A. Nudelman

HE WEATHER IS CAPRI- Theorem of the zero of a function. is continuous on its domain l-2, -Il
cious. The parameters that de- If a t'unction f is continuous on the w ll, Zl, is negative on l-2, -1], and
scribe it (for example, pressure, intervalla, bl and takes values with positive on [1, 2]. However, it doesn't
temperature, and humidity) opposite signs at its endpoints, then have a value of zero at any point.
vary continuously over time and thare exists a point xo between a In this arttcle, we'Il examine
from place to place. The isothermal and b such that f(xo) : O. some properties of pairs of continu-
and isobaric curves on weather maps We won't prove this theorem-it ous functions defined on a sphere.
take whimsical (and, alas, often un- may seem obvious geometrically, But first we'll look at a simpler case:
predictable) shapes. Yet no matter but strange to say, the proof is far an unexpected property of continu-
how convoluted the weather map from elementary. ous functions de{ined on a circle.
looks, the following theorem is true. For the "zero theorem" to be true,
Weather theorem. At any mo- both the continuity of the function The mse ula rhcle:o circu$ hu'$E
ment theru exists a pafu of diametri- and the connectivittz of the segment
pErlorllls
cally opposed points on the Earth are essential. The reader is invited to
("antipodes") where both the tem- explore why we need continuity, By Suppose that a circus horse begins
perature and the pressure are iden- "connectivity" we mean that there running smoothly around a ring
tical. are no " gaps" in the segment. For ex- from a point A, and stops smoothly
Although we've couched this ample, the function at the same point, after making a fuli
proposition in meteorological terms/ circle. It turns out that no matter
it is actually a property of continu- f(x)= -,)(o- *)+zx how the speed of the horse varies,
ous functions defined on a sphere there exists a pair of diametrically
rather than properties of the atmo- opposite points where the horse's
sphere. The theorem lies within the speed is the same.
realm of topology, a division of It goes without saying that the
mathematics that deals, among horse isn't the cause of this-in fact,
other things, with functions or col- we'll turn the horse into a point later
lections of functions that are con- on. Rather, it's a property of con- o
tinuous for certain sets. tinuous functions. .C
a
Some properties of such functions We can determine the position of tr
are determined by the structure of the horse H on the circle by the p
C
the set on which they are defined. magnitude of the angle 0 l0<e< }Tc) o
()
J
For example, for numerical func- that the radius OH forms with the
radius OA drawn from the starting -o
tions that are studied in high school
the following theorem holds: Figure 1 point (figure 1). The corresponding

1 0 JUIY/AUGUsI zooo
 lt'., ,, ,, ,:

j
il
a i

i
:
"ffi _i

# e
fi :

fls
I

1'

il

:l
l
.:
. -c,

-:

i
ll:r'.. rlt ,

OUAilIU[l/IEATURI 1t
speed will be denoted by v(0). We the sphere S. We assume they are
assume that the function Y is con- continuous. Indeed, the values of
tinuous on [0, 2n] (the horse stops these functions cannot vary too
and starts running smoothly), and much when the location o{ point P
according to the statement of the changes slightly.
problem v(0) = 0, vlLnl = 0. The point A precise definition of the conti-
H' drametrically opposed to H is de- nuity of a function defined on an ar-
termined by the angle 0 + n (where bitrary point set X (located on a line,
0 < e S ru). We need to prove that . on a plane, or in space) can be formu-
there exists 0o e [0, n] such that v(00 lated as follows. The function g is
+ n) : v(00). called continuous at the point Po e
Consider the function u(0) = v(0 + X If lor arry e > 0 there exists 6 > 0
Figure 2
n) vl1). We want to find 0o e [0, n]
- such that, for any point P e Xwhose
such that u(0s) : 0. The function u is distance from Po is less than 6, the tion that is continuous on an inter-
continuous on the interval [0, n] (as inecluality lq(P) - q(P6)l < e holds. val to have a zero oL this interval.
the difference of two continuous The function is continuous on the We now formulate a theorem that
functions), and its values at the end- set X if it is continuous at every gives a criterion for the existence of
points of this interval have opposite point of this set. a zero that is common to two func-
signs (if vlnl + Ol or are both equal to We will denote by P' theendpoint tions that are continuous on a circle.
zero llf v{m) = 0). Indeed, ul}l:v(nl - of the diameter of the sphere whose Let every point Q of a set K ({or
v(0) : v(n) and u(n) : vlLxl - v(nl :' other point is P. our purposes, this will usuallY be a
v(n).In the case when v(n) = 0, we Leaving aside the "meteorologi- circle) in a plane be assigned a vec-
can set 0o = 0, otherwise, the exist- cal" meaning of the functions f and tor a(Q) in the same plane. In this
ence of 0o follows from the theorem g, we'll formulate the theorem in its case/ we say that atwo-dimensional
of the zero of a function. general form. vector on K (figure 2).
fieldis defined
We can see that the nonnegativitY Borsuk-Ulamr theorem. If the A vector in a plane is defined by two
of the function v(0) is not needed to functions f and g are defined on a coordinates. Therefore, the defini-
prove the theorem. (That is, the sphere S and arc continuous, then tion of a two-dimensional vector
horse may sometimes run in the there exist diametrically opposed field is equivalent to the definition
opposite direction around the ring.) points P, and PI on the spherc such of two functions: a(Ql: (x(Q), y(Q)).
From the conditions v{0) : 0 and that f(P[) = f(Pol and g(P[l = s(Po). The vector field is said to be con-
vl2nl :0, only the equality v(0) : Let's introduce two functions F(P) tinuous at the point Qo if both func-
vlZn) is essential (the horse may pass = flP'l - flP) and clP) = slP'l - slPl. tions x(Q) and y(Q) are continuous
the initial point A with a nonzero Both these functions ate continuous at this point. The vector field is con-
speed. What matters is that it return on S and are antisymmetic'. F(P'I = tinuous on a given set Lf it is con-
to this point with the same speed). - F(Pl and G(P'I = - G(PI.For in- tinuous at every point of this set.
It's clear that v may be considered stance, F(P'I = fl!'l'l - f(P'l = f(Pl - The vector field is said to be degen-
a function of the point H rather than f(P'l = -F(P). For the points Po e S erate on the set R if a(Qo) = 0 for a
a function of the number 0. There- such that flP;l = flP,l andg(Pj) = s(Po) certain point Qo e K. We want to
fore, our theorem can be formulated and only for such points, I(Po) = 0 obtain conditions that are sufficient
as follows. and G(Po) = 0. for a continuous vector field in a
The "horse" theorem. If a func- Therefore, the Borsuk-Ulam circle to be degenerate.
tion is defined on a circle and is con- theorem would follow from the {ol- First, we give a vector interPreta-
tinuous, there axist two diametri' lowing theorem. tion of the zero theorem. A vector
cally opposite points on the cfucle Common zero theorem. If the defined on a line is given by a single
whera this function takes on equal t'unctions F and G are continuous coordinate. Therefore, a function
values. and antisymmeftic on a spherc S, defined on an interval may be inter-
then therc exists a point P o at which preted as a one-dimensional vector
Mathematical lol,mulations ol the both these functions are eclual to field (consisting of vectors that are
zero:F(P6l: G(Po) :0. oriented along the given line). Thus,
tlll8alh8r lhe0l'8ln the theorem of the zero of a function
Let's rephrase the weather theo- I[eonem olile nnos ultlocmr lield$ may be formulated as follows.
rem in mathematical terms. At a Zero theorem gives the condi- Theorem of the zero of a one-di'
given moment in time, each point P tions that are sufficient for a func- mensional vector field.If a continu-
of the Earth's surface S can be char- ou s one - dim ension al v e ctor f i el d i s
1 Karol Borsuk (born in 1915) is a defined on an interval and the cor-
acterizedby two numbers-the pres- Polish mathematician; Stanislaw
sure /(P) and the temperature g(P). Ulam (1909-1984) was an American rcspondingvectors at the endpoints
Thus two functions are defined on mathematician. of this segment have opposite di'

18 JUI.Y/AUGllST 2OOO
 eral revolutions counterclockwise
and several revolutions clockwise as
Q moves along an arc of Co, these
revolutions "cancel out.// We sug-
gest that you try to find v(C) for the
fields depicted in figure 4 (the field
of velocities of a rotating circle and
the field corresponding to paralle1
transiation).
Let a(Q) (where Q e K) be a vec- Figure 5
tor field satisfying the conditions of
the theorem. Now let's assume the
opposite-that is, that a(Q) + 0 for all
Q e K. Then, in particular, a(01 + 0.
Figure 3 Due to the continuity of the field,
the direction of all rrectors a(Q) is
rcctions, then this vector field is de- close to the direction of a(0) in the vi-
generate (figure 3a). cinity of 0. Therefore, for a small eo
This theorem can be extended for > 0 we have v(C.) : 0, since the direc-
the case of a two-dimensional vector tion of the vectors on C.n is "almost
field as foilows. the same" as the direction of a(0) (see
Theorem of the zero of a two-di- figure 5), and the vector a(Q) makes
Figure 6
mensional vector tield.If a continu- no full revolutions as Q moves along
ous two-dimensional vector field is C.o. We now gradually increase e to continuous or constant. Thus, in our
defined on a circle and the corre- its extreme value e : 1. Since a(Q) + case, v(C.) : coltst, from which we
sponding vectors at any diametri- 0, the number v(C.) is defined for ali get v(C) : v(C*) : 0.
cally opposed points on its circum- values of e, 0 < e < 1. That is, the Now consider a point A on the
ference have opposite directions, function v(C.) (considered a function circumference C and its antipodal
then this vector field is degenerate of e) is defined for all e on (0, 11. This point A' (figure 6). Since the direc-
(figure 3b). function is continuous (since our tions of the vectors alA) and a(A'l are
Proofs of the "one-dimensional field is continuous) and takes integer opposite to one another, the vector
zero theorem" and the "two-dimen- values only. However/ the integer- a(Q) makes an odd number of half-
sional zero theorem" would take us valued function can change only by revolutions as Q moves korr.Ato A'
well beyond the high school curricu- steps (each step is greater than or along the arc AmA'. As Q moves
lum. But while the one-dimensional equal to 1). Therefore, it is either dis- further frorn A'to A along the arc
theorem is easy to grasp visualiy, A'nA, the vector a(Q) makes the
such is not the case for two-dimen- same number of half-revolutions in
sional vector {ields. So we'1lprovide the same direction. Therefore, v(C)
a "plausible reason" for you to ac- is an odd number, which contradicts
cept the validity of this theorem.

(F),,
the equality v(C) = 0 proved earlier.
Let O be the center of a circle K This completes our reasoning.
andr its radius. We'1I denote the cir-
cumference of K by C, and for any Pl'ool d I[e theol'em ol tfle
number a, we'lllet C, denote the cir-
cumference of the circle centered at c0lnlll0[ I8r0
O with radius a. Let a be a continu- Let the functions F and G be con-
ous vector field in K. Suppose that tinuous and antisymrqetric on the
alQl + 0 for ail Q e C,. Then the sphere

2/./'
S:
vector a(Q) will change its direction
F(A',) : - FIA), G(A',l : - G(4)

z//z
somehow as the point Q moves
along Co. Let's denote by v(Cr) the Construct a plane passing through
number of revolutions that the vec-
tor a(Q) applied to a point l performs
in the counterclockwise direction as
point Q passes along C, counter-
'ar./, the center O of S. The cross-section
is a circle K, and we denote its cir-
cumference by C. Let's introduce a
rectangular coordinate system in the
clockwise. The integer v(C,) can be secant plane whose origin is the cen-
positive, negative, or zero. To be pre- ter of the sphere. Assign the vector
cise, we note that if alQ makes sev- Figure 4 a(Q) with coordinates FlPl and G(P)

g
OUAIITU'iil/IIATURE I
 whole sphere such that the coordi-
nates are continuous functions of
the corresponding point on the We recommend that readers who
sphere! Naturally, different points wish to gain a deeper understanding
must have different coorfinates. of the basic concepts of topology
It follows from the Borsuk-Ulam read the excellent book by W. G.
theorem that this is impossible.In- Chinn and N. E. Steenrod, First Con-
deed, if continuous coorfinates x(P) cepts of Topology: The Geometry of
andy(Pl were given on the sphere, a Mappings of Segments, Curves,
pair of antipodal points Po, Pd e S Circles, and Disks (New York: Ran-
would exist for which xlPo) = xlP[) dom House, 1965li. O
and y(Ps) = ylP;).
Figure 7
to every point Q e K, where P is the $ome Ueltgl'fllizrliom CONTINUED FROM PAGE 9
point on the upper hemisphere that If we thoroughly analyze the angle of the rhombus. Then the layer
prolects onto Q (figure 7). We sug- proof of the "horse" theorem, we'11 is translated. Therefore, if the angle
gest that you verify that this vector see that it isn't so important that the of the rhombus is incommensurable
field is continuous on K. By virtue of points H and H'be diametrically op- with n (that is, if a cannot be written
the antisymmetric property of the posed. The reasoning remains valid as (mln)n for integers m andn), then
functions F and G, we find, for all if we take an arbitrary point O, in- no two biprisms from different layers
Q e C, that side the circle instead of O and in- can be parallel. On the other hand,
terpret H andH'as the opposite end- any translation that maps a layer
a(Q) = -a(Q').
points of a chord passing through onto itself cannot map any other
By the zero theorem for two-dimen- Or. This also applies to the zero Iayer onto itself. Thus, if the angle a
sional fields, there exists a point Qo theorem for the two-dimensional is incommensurable with zr, there are
e K such that a(Qo) = 0. Therefore, field and to the Borsuk-Ulam theo- no translations that map the decom-
FlPol = 0 and G{Po) : 0, where Po is rem (figure 8). It's only important position described onto itself.
the point on the sphere that proleits that the property (P'l' : Pbevaiid for However, if we ask the same ques-
onto the point Qo. Thus, the theo- the new definition of the point P'. tion about a plane figure, the answer
rem is proved. Any continuous mapping P -s P' is not known. It is possible that there
that possesses this property is called are no such apefiodic tlles on the
A goograilical mnsequence aninvolution The Soviet mathema- Euclidean plane. An analogue of the
There is a consequence of the tician A. Fet used ingenious and aperiodic tile on the Lobachevskian
Borsuk-Ulam theorem that is rather powerful topological methods to plane has already been found. It
distressing for geographers. The loca- prove that the Borsuk-Ulam theo- would be wonderful if a reader of
tion of a point on Earth is given by rem (even its n-dimensional version) Quantum discovered an aperiodic
geographic coordinates: latitude 0 remains true for anarbitrary involu- tile on the Euclidean plane. O
and longitude <p. These may be con- tion P -) P' on the sphere. Here is
sidered functions of a point on the the precise formulation of this theo-
terrestrial sphere. In this coordinate rem for three-dimensional space. CONTINUED FROM PAGE 14
system/ the poles have a peculiar Let an arbitrary involution P -+ P' be "cut off" on the low-frequency
property: the latitude of the poles is be given on a sphere S. For any pair side, as shown qualitatively by the
90" (N or S), and they can be assigned of continuous functions f(Pl and glPl dashed line in figure 6-thatis, they
an arbitrary longitude. So if we go to defined on S, there exists a point Pn are shifted toward the "violet" por-
the North Pole aiong a meridian and e S such that f(Ps'l = f(Po) and g(ps') tion of the spectrum. If the specks of
then continue moving along another = s{Po) dust could be heated to the tempera-
meridian upon reaching the pole, our ture of the Sun's surface, the smaller
motion will be continuous and the they are, the bluer they'd look (in the
latitude will vary continuously, but visible range of the speetrum).
the longitude undergoes a disconti- All these considerations come
nuity. If we assign the plus sign to into play in many fields of science
East longitudes and the minus sign and technology-for example, in
to West longitudes, then the longi- studies of the energy balance in plan-
tude undergoes a discontinuity when etary atmospheres, metallurgical fur-
it crosses the meridian that is antipo- naces/ rocket jets, and so on. So we
dal to the Greenwich meridian. The see that the waves generated by a pi-
question arises: rs it poss ible to intro- ano string have reverberated far into
duce a cootdinate system on the Figure B diverse areas of research. o
20 JUI.Y/AUOl,lSI 2OOO
 rtZtrItfl I

I EII)TI TI'i !

rli t[iHil.l(* 1*'{r I a !

Move with Science:

Energy, Force, & Motion
Bring rea[ transportation situations jnto the ctassroom
through anatogous hands-on activities and background
reading sections. Focus on the physics and bio[ogy of
transportation safety and safety devices. Move with Science
uses methods of transportatjon that are most famitiar to
high schooI students to connect the basic concepts of
physics and human bioLogy to the concrete sights, sounds,
and physical sensations that students experience nearly
everyday. (Grades 9-1.2, 1.998, i60 pp.)
#?8744X $21.95 Member Price $19.76

Evidence of
Energy: An Methods of
Introduction Motion:
to Mechanics, An Introduction
Book Two to Mechanics,
Do your students Book one
understand why and how objects move? Topics Could Isaac Newton reatty have betieved that a
such as projectite motion, work, energy, thrown object woutd continue at a constant
machines, torque, and center of gravity are vetocity in a straight l.ine? This manuaI js
introduced with the first-time student of designed to help teachers introduce the daunt-
mechanics in mind. The informat, hands-on ing subject of Newtonain mechanics to
actjvities use a variety of instructjonal students. The teacher-created activities pre-
techniques to make the subject of mechanics sented here use readity avaiLable materiats to
accessible-and fun-for teachers and com bat students' misconceptions.
students. (Grades 6-1.0, 1.990,200 pp.) (Grades 6-10, 1998, 163 pp.)
#PB080X $17.95 Member Price $16.16 fiPB039X $21.95 Member Price $19.76
P70002

NiEIiEffi l-^rlTmF{ Ff,ir{ItE TFIlTfiEI iraT[arilra]-ilnm

 AlUehl'aic and ll'anscendenlal numher$

Thought-provoking for thousands of years

by N. Feldman

ATURAL NUMBERS, INTEGERS, RATIONAL

AlUehl,aic ltttlnhsls
numbers, real numbers, and complex num-
Every rational number a =
bers-this expanding chain,
alb (a e Z, b e Z) is the root
NcZcQcRcC, of a polynomial with integer
has been familiar to mathematicians coefficients-for example, of
for cluite some time. the polynomial bx - a. Any ir-
Perhaps you've read the articles rational number of the form 4la
in Quantum describing how hard la e Zl is the root of a polynomial
it was for negative (N c Zl and of this kind as well-for example, of the
complex (R c C) numbers to be polynomial xn - a. We now want to consider only the
accepted as links in the chain. numbers of this sort-that is, roots of polynomials
This article will have something to with integer coefficients.
say about another part of the chain-the By definition, arealnumber is called algebraic
inclusion Q c R. if it is the root of a polynomial with inte-
You've certainly heard about irrational * ger coefficients that is not ec1ual to an
numbers (that is, numbers that cannot be -o"l',;t
,t-t a
l"\,
* o
identically zero polynomiai.l We de-
represented as a fraction mfn, whete m e Z
a
note the set of all algebraic numbers by
and n e N). These numbers had already been A. As we've already seen, Q c A c R. To
discovered in anticluity. The fact that the diagonal get a better feel for the conccpt of algebraic
of a scluare is incommensurable with its side (translated numbers, prove the following propositions.
into the language of algebra, this means that "uE is an 1. If s e A (o+0), then l/u e A.
irrational number)was one of the most exciting (and dis- 2.If uis the root of a polynomial with rational coef-
turbing) scientific discoveries of the time. Nowadays, a ficients, then cr e A.
proof of this fact is given in elementary textbooks. If you 3. If cx e A and a e Q,then aue A and a +'ue A.
have any feel for the mathematical way of thinking, the It canbe proved that if ote A and F. A, then u+ B e
elegance and deceptive simplicity of this proof cannot A, o-F. A, cr. F e A, ando/pe A (in thelatter case, B
I
leave you cold. cannot equal zero). In other words, ailthmetic opeta'
o Besides being classified as rational and irrational tions do not take us out of the set of algebtaic numbers.
a
(6 numbers, real numbers can also be classified as algebraic

(such as 213, Jr,5 - "tV, or 4l4) and transcendental 1 The set of polynomials with integer coefficients is
'a denoted by Zfxl.In this article, we consider only
(U numbers
(such as rEt e, ot lg 2). This classification isn't
polynomials that are distinct from the zero polynomial
as well known, but it's very important nonetheless. In
_o this article we'll examine these two classes of numbers,
without specifying this fact every time. For readers who are
familiar with complex numbers, we note that it is possible
their properties, and their (ongoing) history. to introduce and siudy complex algebraic numbers as well.

22 JUI,Y/AUOUSI 2OOO
 (The proof of this faci is more complex than the solution
of the first three problems, so I won't give it here.) Thus, Liouuilht fieorem
the set A with the operations + and x, as well as the set At first glance, the formulation of this theorem is
Q with the same operations, form a number field-that unrelated to the existence of "nonalgebraic,,numbers.
is, a numbdr set with all four arithmetic operations de- Liouville's theorem lLB44l. If u is an algebraic num-
fined on it (except for &vision by O, ofcourse), and all the ber of degree n> 2, thentherc exists a number c > O such
operations act as they usually do in arithmetic. that,foranypeZand4€ N
The question naturally arises: are there any real num-
bers that are not algebraic? To give an answer/ we need IlC[-'l>-
nl c
a notion that is explained in the next section.
I qlqn
Ihe degree ol an aluelnaic numhel' This theorem in effect says that an irrational alge-
If o, is a root of the polyno- braic number 0( cannot be approximated by rational,frac-
mial Plxl, then it is also a tions "very well." Therefore, if we find an irrational
root of the polynomial number that can be approxim ated "very well', by ratio-
P(xlQft), where Q(x) is nal numbers, it is not algebraic.
an arbitr ary polynomial. Proof. Let u be an algebraic number of degree n > 2.
Therefore, every algebraic Then there exists a polynomial
number u is the root of an infi-
nite set of polynomials from Z[x]. P(x) = arf + ar_rxn-l + ... + atx + aox", (ar+Ol
Clearly we can find polynomials of minimum with integer coefficients such that Plo,l: 0. Denote
degree among them. If this minimum degree is n, we by H the greatest of the absolute values among la,l,
say that cx is an algebraic number of degree a and write larl, ...,larl. We'l1 show that the number
deg cr = n. We can see that deg s = 1 if and only if cr e e.
It's also clear that the degree of an irrational number of
the form nG (a. Zl is Z-that is, deg ,G = Z. "=-----1-
nzn(r* l"l)"-'
To go any further, we'Il need the following simple,
yet important, theorem. possesses the desired property. Notice that c < 1. Let,s
The remainder theorem (l77gl. The remainder when take an arbrtrary p e Z and q e N. Then
the polynomial P(xl is divided by , - y is p(y).
Proof. Let us divide P(xlby x - y. The remainder is a
constant/ which we denote by c: P(!\= anp' + ar-rpn-rq +...+ atpqn-t + aoq' _ a
lrt ) qn qo'
P(x) = lx - y) PoEl + c,
where we have denoted the numerator of the fraction,
where Po(x) is a polynomial. Plugging 76 = yinto this for- which is an integerby a.
mula, we find that c = P(yl. By the lemma,
On the basis of this theorem, we can easily prove the
following 1emma. p(L)*0.
Lemma. If an algebraic number a of degree n > 2 is \q)
a root of the polynomial Plxl e Zlx] of degree n, then p(xl Therefore, a * 0. Since a e Z, we have lal > 1. Therefore,
has no rational roots.
Proof. Assume, on the corltraryt that p(albl: 0, where
a e Z and b e N. By the remainder theorem, the remain-
der when P(x) is divided by , - alb ts O. Therefore, p(x) l,(#)=a
is divisible by x - alb: Since P(a) : 0, we obtain

P(')=("-f)nel'
,,#..-ffi
*ffi where Pn(x) obviously has rational coefficients. If
frffi
,ffi
M is a coirmo.r multiple of the denominators of
the coefficients of po(x), then f r(xl: tW pok) e Zfxl.
*W Since P(u) : 0 and s * alb (the degree of o is gteater
than 1), we have lro(cx) :0. Therefore, pr(cr) :0. How- If
ever, the degree of the polynomial pr(o)-is n - | < n :
deg cr. So we have arrived at a contradiction.
lo-el-r,
The decisive step in the search for numbers that are lql
not algebraic was the following theorem. then we have

OUAlllTllfil/IIAIllRt 23
 lrlc
la_:_L >l>_>-
ApruximatinU alUnhl'aic ltttlnher$ tltil[ mlioltal ltulnhols
We say that the number cx allows approximations of
lqlqnqn ordet m tf, for a certain constant y, there exist infinitely
and the assertion of the theorem holds. On the other many rational fractions p I c1 that satisfy the inecluality
hand, if
o.lo-zl.-1.
l*-al.r I ql ct'' lz)

lql Liouville's theorem shows that algebraicnumberc of

then degree n do not allow approximations of an order
greatu than n.Indeed, if u allows approximations of or-
der m, then it follows from (1) and l2l that, {or an infi-
l4l . tot* r. nite secluence of natural numbers q, the inequality
lql
Obviously, Icrl < lsl +1. Then, forany 1 <k <n (k e N), c -y
we obtain - '
,1rt ,ttt

'l
which can be transformed
-;l o"' -,[;)- into
l"-
(;)-l= l" l**-'. t-,
1c
-;lol'"1 *')o-'
'l" ' l" -;l<r"r* 1)" cl"' " y

from which we get

" holds. However, this is im-
possrble (or m > n and sui{i-
ciently large q.

;.1" -11-6ul+r)"-rr=1, -il:,,d l,

-i,,; txample ol a tnanscendenlal

Thus the theorem is proved. llulllhE['

It's possible to reformulate Now we have an instrument for
Liouville's theorem so as to constructing real numbers that are
eliminate the condition deg not algebraic (such numbers are called
/'dg
a> 2. f tanscendentall.For this purpose, it's suf-
Theorem. If a is an al- .frf-l ficient to construct a number that admits
#.,; g
gebraic number of degree du" d; approximations of an arbitrarily high order.
n, then there exists a #g We define such a number as an infinite deci-
number co , 0 such that, mal sequence 0: O.ararar..., where
foranypeZandqeNfor
which a + plcI, It, iIt=mt (m=1,2,...),
6rG
a-
' [0, rIt+ml.
(3)

t* c*, I
o-41--L (1)
| . 2. 3 . ... . lm -
I ,tl q" @ ,i' (Here m! denotes the product
ru, which is called the "m factorial.")
Ll
. ' r.di
In particulart (tt: az = ae = az+: atzo = atzo: ... =
Proof. The case deg o > 2 has 1 and a, = a+= a5: a7= ...= azz: (lzs= ...= arrtr: (tr2,
aheady been considered above. = ... = 0. Then, for any fl > l,
Let deg cx = l-that is, cx,: alb (a e Z,
b e N). Then, the number c' : llb possesses
the desired property. Indeed, if p I c1 + a f b, then a: O. arara3 ... a1r, ,,, + 0.0 ... Olt,rr,arr,. *,. = !Q,,, * Frrr,
lpb - clal + 0, so lpb - clal > 1. Therefore,
where
l" - sl =lg - zl -lPb - qal,f- = !
-bq
l- qllb ql bc1 q 2lI yot are familiar with the concept o{ countability,
Setting co equal to the minimum of the numbers c and you can easily prove that the set o{ algebraic numbers is
countable. I{ you also know that the set of all real numbers
c', we obtain the desired inequality. is uncountable, you can immediately conclude that
Liouville's theorem can also be proved by examining transcendental numbers do exist. However, this reasoning
the difference P(o) - p(pld and using Lagrange's mean does not provide a single concrete example of a
value theorem. Try to find this proof! transcendental number.

24 JU[Y/AUGU$T 2ooo
 p,r= arara^... alnt-t)!r ent= 1Oim-1Jl/ 0,,r:0.0 ...\a*,a,r, *y...,
a,l{t,o}-{kro}=
0 < Fr, = \}-n/' a,r,arrr r... = l0{rrl . l, ... < 2. l}--t - +[kr"] = l{r, - rr;" - ([t,"] _ tftr"l)l
+ ]\m lt,u - [t,o] k2u
\q-)
Thus If we now set q = k, - k, and p : [k, o] - [krcr], we obtain
the desired inequality bydividing the above inequality by
lt
o.lu- bl.:-, m : -'t,2,-' ..., q and taking into account the fact that 0 < k, < k, < m.
I q-l (r* Corollary. Any irrational number a can-be approxi-
mated to an order of 2.
which means that o allows approximations of arly or- Proof. For any m e N, there exist p e Zand q e N
der whatever. Therefore, it cannot be algebraic. such that e3 m, and inequality ft) ho1ds. Since q < m
Exercises and o is irrational, from (4) we obtain
4. Prove that c is transcendental if in equation (3)

lrfort=m- o. io-zl. +
*'
,t
-lo lort*m"'
-
)
(m =1,2,...),
I qlq' 16)

A
It follows from (4) that the quantity lu - p I ql
Li
5. Find several more transcendental numbers us- * becomes arbitrarily small as m increases.
ing Liouville's theorem. & Since this quantity cannot be zero, the
& fraction p f q comes ever closer to o( as m
[ificfihtb fieonem increases. Thus (6) holds for an infinite
In 1955 the English mathematician Klaus Roth &e B ; number of rational numbers p/4.
proved that no furational algebraic number can be ap- &*,' *
proximated to any order greater than 2. tam0us lrall$c8lldellla!
At the same time, every irrational number can be ap- llll1ll[8r$
proximated to an order of 2. This fact was proved by
Although Liouville's
the German mathematician Peter Dirichlet using a and Roth's theorems aI-
principle that now bears his name. This principle is
low us to construct infi-
simple yet fruitful: if n items are disftibuted among nitely many transcenden-
n - | boxes, then at least one box contains Z or mote
tal numbers, they have
7tems.
in di-
been of no use so far
Exercise 6. Construct several transcendental rectly proving the transcen-
numbers using the Roth theorem.
dence of such well-known
Dirichlet's theorem (l924l. For any real number '' numbers as fi/ e, InZ,lg2, and
u and any natural numbet m, therc exist p e Z and so on. These numbers have been
qe N suchthatq<mand attracting attention for centuries.
The number zc is especially famous. The
lolI mathematicians of ancient Greece posed the problem
l"-;l'n*' @)
of squaring the circle: given any circle, construct a
squarc of equal area with a straightedge and compass.
Proof. The interval [0, 1) is a union of m intervals This problem is reduced to constructing a segment of
length n, given a segment of unit length. For 2,000 years

[0,*), l**), ,lr#,+),1+) (s)

all attempts to solve this notorious problem have failed.
Eventually such a solution was shown to be impossible,
and to establish thatfactit,s sufficient to prove the tran-
Consider the numbers {ko} (k : O, I, ..., ml (where {x} scendence of n (in fact, it's sufficient to prove that n is
denotes the fraction al part of x. We recall that, by defi- not an algebraic number of a certain type). .
nition, {r} = , - [x], where [x] is the integer part of x- The irrationality of the numbers e and n was proved
that is, the greatest integer that does not ex- by |. Lambert in 1756.In 1873 C. Hermit proved the
ceedx). Each of these numbers belongs transcendence oI e. The method he developed for this
to one of the intervals (5). We have purpose continues to play an important role in number
m+ | numbers and m intervals. theory. In 1882 F. Lindemann improved on Hermit,s
Therefore, by Dirichlet's prin- method and proved that n is transcendental. He also
ciple, at least one of the inter- proved that the number ao is transcendental for cx, e A
vals (5) contains two or more (u + 0). This fact implies that natural logarithms of all
numbers. Let these numbers be algebraic numberc distinct from I are transcendental
{k,o}and {kro} (k, , kr). Then (try to prove this).

OUANIUIi{/IIAIUBI 25
 In 1748 Euler suggesied that if a, b e Q, and logob things, he proved the transcendence of eN. A
is irrational, then it is also ftanscendental. Certainly, j year Later the Soviet mathematician R. O.
it is clear that rational-for example,
logob can be Kuzmin showed that Gelfond's method with
logo8 : 3l2.This conjecture was not proved in the certain improvements could be used to
18th or 19th centuries. prove the transcendence of the numbers o,F

In 1900, at the International Congress of Mathema- when cr is an algebraic number different

ticians in Paris, David Hilbert formulated twenty- from 0 or 1 and g: nE , where d is a natu-
three problems that he thought would ral that is not a perfect square. In parlicu-
stimulate the development of mathemat- 1ar, he proved the transcendence of 212.
ics. The seventh problem was as fol- A complete solution of the Hilbert's seventh prob-
lows: if o and B are algebraic numbers, lem was given by A. O. Gelfond in 1934 by means of a
cr is not 0 or 1, and p is irrational, then new method, which was called Gelfond's second
s0 is transcendental. In particular, method.
Hilbert suggested that someone prove Gelfond's theorem. Let u,0 . A; u is not 0 or I, and
that 212 and en are transcendental (the $ is irrational. Then aF is ftanscendental.
second number can be reduced to the form s0, Exercise 8. Prove that if the numbers o, F, and p are
where u, B e A; however, this requires some knowledge such that the expression logoo/logoB is defined and u, p
of functions of a complex argument). e A, then the number logoo/logoB'is transcendental or
Exercise 7. Prove that Hilbert's proposition implies rational.
Euler's hypothesis. Gelfond's second method makes it possible to prove
The first partial solution of Hilbert's seventh prob- many other theorems. An improvement of this method
lem was obtained in 1929 by a postgraduate student at by A. Baker in1966led to significant advances in num-
Moscow University, A. O. Geifond. Among other ber theory. Work in this area is far from finished. O

iL;qt ii: E r e: -l1 H .&n'K,q'g:.Fg g::;1,{,&-"$ E {,-: ih f-,*6 {# {: E *i'3'-H

r!::';
iai': !-jii: i. r1'-j. 1:'11 i;l i',1.u., .!'E:i: liilr,'j,'-.

The History of Mathematics from introductory notes provide evaluation and context to
Antiquity to the Present: A help users decide where to go next.
Selective Annotated B i bliography, The CD-ROM's content is in PDF format with book-
edited by foseph W. Dauben marks, enabling browsers to link from the table of
Revised Edition on CD-ROM contents to sections of interest, and there are links
from the author and subject index items to entries in
Albert C. Lewis, Purdue University, West
the text. Adobe"Acrobat@ Reader 4.0 with Search
Lafayette, /N, Editor, in cooperation with

ffi'
makes it possible for users to search text throughout
the ICHM
the CD using PDF: the search term is highlighted, and
ri"Sr$xliil';**a Thirty-eight experts have critically anno- you can move from hit to hit across files. This edition
tated a selection of the literature within is unique in that it also offers a listing of World Wide
their respective specialiies in the hisiory Web resources and links within the text to selected
of mathematics. The 4,800 entries Web sites. This publication on CD-ROM makes it
(twice as many entries as the original easier for both browsers and researchers to locate the
edition of 1985) include full biblio- cited works that have been selected and annotated for
graphic information and cover the this bibliography,
history of mathematics beginning with Adobe" and Acrobal@ are trademarks ol Adobe Systems
the ancient civrlizations. lt also focuses on the various lncorporated.
branches of mathematics and their applications to 2000; CD-ROM; ISBN 0-8218-0844-3; List $49; All AMS
many other fields of human endeavor. Annotations and members $39; Order code HMAPQ070

continental U. S., please include $6.50

All prices subject to change. Charges for delivery are 53.00 per order. For optional air delivery outside oi the conl
pet ilem. Prepayment requlred. Order lrom: American Mathematical Society, P 0. Box 5904, Boslon, Boslon. l\,4A 02206-5904. USA. For credit card orders,
l,4A 0220{
{ax 1-401-455-4046 or call toil {ree 1-800-32'1-4AMS (4267) in the U. S. and Canada. 1'401'455-4000 worldwide. Or place your order through the
A[.4S bookstore at www.ams.org/bookstore. Residenis of Canada, please include 7% GST.

Circle No. 1 on Reader Service Card

26 JUI.Y/[UGUST 2OOO
 HAPPEN INGS

Bulletilt hoard
Cylerleasel' winnens Maxim Bachmutsky (Kfar-Saba, Israel) Quantum button. Everyone who
The following are the names of Oleg Ivrii (Toronto, Ontario) submitted a correct answer (up to
first ten people to submit a conect Mikhail Agladze (Ithaca, New York) the time the answer is posted on the
answer to this month's Cyber- Theo Koupelis (Wausau, Wisconsin) web) is entered into a drawing for a
Teaser-Wrong-Way Mercury. We're Chris Ridgers (Cambridge, England) copy of Quantum Quandaries, a
sure the problem generated heated Howard Brown (Idaho Falls, Idaho) collection of 100 Quantum brain-
debate among our contestants, but Clarissa Lee (Selangor, Malaysia) teasers. Our thanks to everyone who
cooler heads prevailed and the solu- Igor Astapov (Kingston, Ontario) submitted an answer-right or
tion was found. Our congratulations to the win- wrong. You will find our next
Dimitrios Vardis (Ithaca, Greece) ners, who will receive a copy of this CyberTeaser at:
Eu fin Teoh (|ohor Bahru, Malaysia) issue of Quantum and the coveted http://www.nsta.org/quantum. O

Mdth endticdl Olyrpiad Challansas

TituAndreesct,Antericon Mathemotics Competitions, Ltniv,ersitv^ of Nebraska, Li.nc:oln, NE
Rizvan Gelca, University o/ Michigun, Ann Arbor, Ml
This is a comprehensive collection of problerns rvrittcn by two experienced and well-known
mathematics educators and coaches of the U.S. lnternational Mathematical Olympiad Team.
Hundreds of beautiful, challenging, and instructive problems from clecades of national and
international competitior-rs are prescnted, encouraging readers to move away from routure exer-
cises and mentorized algorithms toward creative solutions and non-standard problem-solving
techniques.
The work is divided into problerns clustered in selt-contained sections wlth solutions provid-
ed separately. Along with backglound material, each section includes representatir,e example s.
beautiful diagrarns, and lists of unconventional problems. Additionally, historical insi-shts ancl
asides are presented to stimulate further inquiry. The ernphasis throughout is on stimulatin.'c
readers to flnd insenious and elegant solr,rtions to problems with rrultiple approaches.
Aimed at tttoti\ated high school and beginning college students anrl instructcrrs. thi-r uork can
be used as ate\t tor arh anccd problen-r-solving coul'ses, for self-studr. or as a resLrurce ibr t.-ach-
ers and students trarning tbr n-tathenratical competitions and fbr teacher prolcs-rronal i,ierelop-
ment. senrinals. and ir orksl.rr-rps.
Fron-r the forenorcl b1' \'Iark Saul: Ilrt' lrooA rr'r,cn ci rrtgttltet Ol.., ttrTtiutl pruhlem; y.itlt t:t tont-
mon theme, .so lhat irtsights betonte rti'lttticlLtt'.:. tt i(k.\ b(,tt)tltt ntethoLl.s, ctntl nteThotls build to
master.r... Much is demcntled o;f'tlte reuier.h rrirr ol tliot.r trtttl putience, but the rnve-stnent i.:. 2000 i 280 PP. 85 ILLr.iS.
SOFTCOVER r' ISBN 0-E 176-.+ I 55-6 I .$29.95
greltl,v" repaid."
I L\RDC'OVl-.1{ I lSBN 0-8I 76-4190-,1 i $59.95

Io Order: Call: 1-800-777-4643. Fax: (201) 348-4505. E-mail: [email protected]: www.birkhauser.com

Textbook evaluation copies available upon request, call ext. 669. Prices are valid in North America only and subject to change without notice.
Birkhluser Boston
-
International Publisher for the Mathematical Sciences z/00 promorion #y1097

Circle No. 2 on Reader Service Card

0 U A I'l T U I'l / ll A P P t ltl I il I S 27

 KALEIDOSI

0eomelricsu
i
HEN YOU STUDY GEOM- Let's divide an inscribed polygon
etry in school, you often have into triangles by drawing all the di- I

to prove theorems that are agonals through one of its vertices

pretty obvious-f or example, (figure 3). Then inscribe a circle in
the fact that when two parallel lines each of the triangles obtained. It
are intersected by a third line, the tums out that the sum of the radii of
alternate interior angles are equal. these circles is a constant, and is in-
However, the fact that three alti- dependent of the choice of vertex o{
Figure 2 the polygon.
tudes in a triangle meet at a point is
not so obvious. In fact, mathemati- We owe the next geometric sur- Later itwas proved that the same
cians in ancient Greece didn't know prise to Archimedes. While studying sum of the radii is obtained for any
this fact, even though they were ex- lunes formed by circles, he discov- decomposition of the inscribed poly-
cellent geometers and knew how to ered that two circles inscribed in gon into triangles (figure 4).
prove that three bisectors/ as well as "curvilinear triangles" (figure 2) are
three medians, of a triangle meet at equal. The figure obtained from the
a point. largest semicircle by removing the
When a mathematicalfact is un- other two semicircles reminds me of
expected, it gives that fact a certain a medieval battle-ax. Archimedes,
charm. This adds to the beauty of who lived much earlier, thought this
mathematics, along with unexpec- figure looked iike the cobbler's knife
ted and elegantly brief proofs. called arbelos; so this theorem is
Inthis atticle, we'll acquaintyou known in mathematics as the arbe-
with some surprising geometrical los theorem.
facts. It's interesting that many surpris-
ing geometrical facts can be seen on
the walls of |apanese temples. |apa-
nese mathematicians discovered
them several centuries ago. In 1800 Figure 4
an inscription was made on the wall
of alapanese temple which made You have certainly dealt with
the following observation. quadrilaterais that are inscribed in a
circle or circumscribed about a
circle. However, many interesting
properties of these quadrilaterals are
less well known. One such property
Figure 1
was discovered by Claudlius Pto-
Let's construct two circles and lemy, who lived in the second cen-
draw tangents from the center of tury. He is known as an outstanding
each circle to the other circle (figure astronomer, but he also contributed
1). Now connect the points where to the development of mathematics.
these tangents intersect the circles. Ptolemy discovered that the sum of
The quadrilateral obtained turns out the products of the lengths of the
to be a rectanglel We don't know opposite sides of an inscribed quad-
who first discovered this unexpected rilateral (figure 5) is equal to the
fact. Try to prove it-it's not very product of the lengths of its diago-
difficult. Figure 3 nals. Ptolemy used particular cases

2 I JtttY/AtlGUsI 2ooo
)OSCOPE

!surprN8$
i
we'11 mention Napoleon Bonaparte,
I who was a serious student of geom-
etry and even read a papet at the
Paris Academy of Science. The theo-
rem ascribed to Napoleon is as fo1-
lows. Let's construct equilateral tri-
angles on the sides of an arbitrary
triangle ABC (ftgure 7) and mark
their.centers O1, C)2, and O.. It turns
out that triangle OrO20Bii equilat-
eral.

03
Figure 5
Figure 9
sides as triangle OrO20B, and it,s
of this theorem, now known as not hard to determine that its angles
Ptolemy's theorem, in his astro- are equal to 50'. Thus the upper tri-
nomical calculations. angle in figure 9 is equilateral; there-
Another interesting theorem in- fore, triangl e O rO rO ris equilaterai
volving a circumscribed quadrilat- as we1l.
eral belongs to Isaac Newton. He I'd like to round out this collec-
noticed that the center of the circle tlon of surprising mathematical
inscribed in a quadrilaterul lies on facts with an elegant miniature by
the line connecting the midpoints of the Moscow mathematician V. V.
its diagonals (figure 5). Proizvolov. Consider a strip formed
Figure 7
by two parailel lines. Let's superim-
The proof of this fact is simple
and elegant. Let's connect the points
O 1, O2, and O, with the nearest ver-
tices of triangle ABC. Then rotate
two of the three triangles obtained
about points O, and O, as shown in
figure B. The triangle iomposed of
these three triangles has the same

Figure 10
pose on this strip a square whose
sides are equal to the width of the
strip. Then connect "crosswise,,
the points where the sides of the
Figure 6 square intersect the boundaries of
the strip (figure 10). The angle
Continuing the list of prominent formed by these lines is 45o. Sur-
people who discovered unexpected prising, isn't it? CI
properties of geometrical figures, Figure B Savin
-A.
0llAlllrlJ]til/t(il.il[0sc0Pr 2g
 PHYSICS
CONTEST 5
*#
E1

ffi
#
E

Batlerio$andhuffis
by Larry D. Kirkpatrick and Arthur Eisenkraft

r
!i

HE RULES OF BASEBALL ARE The habit of analysis

the same for everyone-from
the smallest Litt1e Leaguer to has a tendency to wear
the biggest Major Leaguer. away the feelings.
However, we expect the exPertise of
the player to increase with age. The Stuart Mill
laws of physics are the same for ev- -John
eryone. We expect that the problems of a logical framework for under-
adults can tackle are more difficult standing electric circuits and was an
than the ones we give children. early example of the kind of inquiry
That's usualiy true-but not always. supported by the National Science
Given a flashlight b attery t a flash- Foundation.
light bulb, and a single piece of wire, After unsuccessfully trying to
hold them together to make the bulb light the bulb using arrangements FtEue 2
light. We have seen adults take more such as that shown in figure 1, most
than an hour to light the bulbl And students discover that they must A11 for-rr rial-s have one thing in
yet, this is the first activity in a les- conlmon-a c ontinuous conducti.ng
son on circuit electricity for fifth path allorvs charges to flow from one
graders. Experience has shown us end of thc battery through the light rt
that fifth graders are much more bulb to the other end of the battery. s
successful at this task than adults. Thrs path is known as a complete i
ctrcuit. By examining a broken 6
Experience has also shown us that 7
studying Ohm's law does not guar- flashlight bulb you can sec that €
antee that one can successfully ana- there is a continuous conducting
path from one metal part through
*
lyze circaits containing batteries s
and bulbs. Elementary education the light bulb to the second metal _1

majors who have studied Batterias part as shown in figure 3. {The com-
and Bulbs in physical science Figure 1
courses at college have often re-
ported that their friends and spouses use two parts of each of the obiects:
in electrical engineering did not
have the conceptual understanding
the two ends of the wire, the two
ends of the battery, and the two
ill I €
to help them with their homework. metal parts of the bulb. The two \
Batteries and Bulbs was devel- parts of the bulb are the metal tip
oped and written by the Elementary and the metal around the base. ilJ ( -Y
C
l
Science Study project in the mid Whenever all of these six parts are ( c0
a
60s. Gerry Wheeler, currently the connected in pairs-no matter how ( (0
E Si
Executive Director of the National you do it-the bulb lights. One such o
F E
Science Teachers Association/ wrote way is shown in figure 2. Can you _o
the final version of this popular unit find the other three wavs of doing
in 1968. It stresses the development this? Figure 3

3ll JUtY/AuotlsI 2ooo

.:lJ,rR"G.U.E T}: Ut,#X n
OUAlllTUll,l/PlIYSICS
plete circuit is preserved when bulbs Let's now look at what happens bulb does not pass through the other
are screwed into sockets as shown in if we use a single battery to light two bulb. You can check this by discon-
the rest of the figures.) identical bulbs. We start by connect- necting either bulb and noticing that
Combining the concept of a com- ing the bulbs as shown in figure 5, an the other bulb is not affected. This
plete circuit with the law of conser- arrangement known as seiles. Be- means that the current through the
vation of charge leads to the conclu- cause there is onlY a single Path battery must be twice that in the
sion that electricity flows from one through the two bulbs, whatever standard circuit and the total resis-
end of the battery and back into the charge flows through one of the tance of the combination must be
other. All of the charge that leaves bulbs must flow through the other one-half the resistance of a single
' bulb. If we use identical bulbs, we bulb. You can verify this experimen-
one end returns to the other end.
Charge does not get lost along the notice that the two bulbs have the tally by letting the battery run
way. same brightness. We also notice that down. It does so in approximately
these bulbs are dimmer than the one-half the time. In general, adding
standard brightness. If we leave the a path in parallel always reduces the
bulbs lit, we discover that the bat- resistance of the combined paths.
tery lasts longer than the battery in
the standard circuit. From this we
infer that there is less current in the
circuit and, therefore, the resistance
of the two bulbs in series is greater
than the resistance of a single bulb.
(We assume that you are aheady fa-
miliar with the concept of resis-
tance. If not, we would spend time
developing this concept more care-
Figure 4 fully.) From these observations we
If we light one bulb with one bat- infer that the brightness of a bulb is
tery as shown in figure 4, we find a rough measure of the current pass-
that the battery runs down in some- ing through the bulb. We will as-
thing less than a day. (The actual sume this from now on.
time depends on the type of battery
and the type of bulb.)However, if we
connect the wire directly between Figure 7
the two ends of the battery, the bat-
tery runs down in less than an hour Let's use these ideas to atalyze
and the wire usually gets too hot to the circuit in figure 7 containing
touch. We infer that the current is three identical bulbs. Which of the
larger through the wire than through three bulbs is brighter and why?
the bulb. We say that the bulb pro- How do the brightnesses of the other
vides more rcsistance to the flow of two bulbs compare to each other?
charge than the wire. Let's denote Notice that the entire current from
the battery must pass through bulb
the brightness of the single bulb con-
nected to a single b attery as the stan- A. Therefore, it must be the bright-
est. At junction f, the current must
dard brightness.
split. Because each path following
the junction contains a single bulb,
the two paths are equivalent and the
Figure 6 current must split equally. Conser-
vation of charge tells us that the
We can also connect two identi- currents through bulbs B and C are
cal bulbs to the battery so that each each one half of the current through
bulb is on its own path from one end bulb A. Therefore, bulbs B and C are
of the battery to the other, an ar- equally bright but dimmer than bulb
rangement known as p atalLel. (Note A.
that the paths may share some of the We can check our understanding
same wires/ as seen in figure 6.) In of the model by answering the fol-
this case, each bulb has the standard lowing cluestions about this circuit.
brightness. The current through one (1) What happens to the brightness
Figure 5

Jl]tY/AUOUST
of the bulbs when bilb A is removed B. The second part of our contest
from its socket? (2)What happens to problem is a modification of one of
the brightness of the bulbs when the questions on the exam given to
bulb C is removed from its socket? select the members of the 2000 US
(3) What heppens to the brightness Physics Team. Which of the identi-
of the bulbs when a wire is con- cai bulbs in the circuit in figure 10
nected across the two terminals of are the brightest? Which are the
socket A? (4lrWhat happens to the dimmest? What happens to the
brightness of the bulbs when a wire brightness o{ the bulbs for each of
is connected across the two termi- the following? (1) Bulb A is re-
nals of socket C? Be sure to write moved from its socket. (2) Bulb E is
down your answers to these ques- removed from its socket. (3) Bulbs
tions before you read on. A and E are both removed from
Now that you've committed your their sockets. (4) Bulbs A and D are
answers to writing, we are ready to both removed from their sockets.
look at the answers to these ques- (5) A wire is connected across the
tions. terminals of socket A. (61 A wire is
(1) Bulbs B and C will go out as Figure 8 connected across the terminals of
the single path to the battery has bulb E. (7) Wires are connected
been broken and there is no current. across the terminals of sockets C
(21 After the removal of bulb C, and E. (B) Wires are connected
bulbs A and B arewired in series and across the terminals of sockets A
are equally bright. Removing bulb C and D.
removes aparallel path to the right Please send your solutions to
of the junction and therefore in- Quantum, 1840 Wilson Boulevard,
creases the resistance of this part of Arlington YA2220l-3000, within a
the circuit. This, in turn, reduces the month of receipt of this issue. The
current from the battery. Therefore, best solutions will be noted in this
bulb A becomes dimmer. Two com- space.
peting effects determine the bright-
ness of bulb B. There is less current I[nnelFouhle
from the battery but it all passes The |anuarylFebruary 2000 con-
through bulb B. Qualitative argu- test problem concerned gravity and
ments do not tel1 us the answer, but the construction of gravity tunnels.
observation tells us that bulb B gets Art Hovey of Amity Regional Hs in
brighter. Connecticut provided a solution to
(3) Connecting a wire across the Figure 9 all parts, and a number of his stu-
terminals of socket A provides a dents (Brian Chin, Alex Rikun, |osh
very low-resistance path around Leven, and Victoria Buffa) were able
bulb A, so bulb A goes out. This also to present solutions to parts A and
reduces the resistance in the circuit, B.
so there is more current from the Part A asked for the force a hoi-
battery. Therefore, bulbs B and C lowed-out lead sphere exerts on a
brighten. small sphere of mass m that lies at
(4) Connecting a wire across the a distance d from the center of the
terminals of socket C provides a lead sphere on the straight line con-
very low-resistance path around necting the centers of the spheres
both bulb B and around C. There- and the hollow.
fore, they both go out. Because this There are three equivalent ways
also reduces the resistance of the of looking at the solution to this
circuit, bulb A brightens. problem. The first is to fill in the
A. For the first part of our contest missing mass of the hollow in the
problem, examine the circuits sphere and add an equivalent mass
shown in figures 8 and 9. In each on the opposite side of the small
case, which buibs are the brightest sphere. The difference of the two
and which bulbs are the dimmest? forces is the desired force. A second
Repeat the questions asked above approach is to calculate the {orce of
for each of these circuits. Figure 1O the sphere as if it were solid and sub-

0llmrTU[t/PllY$tC$ s0tlrrrsr 33
tract the force due to the mass imag- x/R
ined to fili the spherical hollow.
0.5
The third approach is to calculate
the force o{ the sphere as if it were 0.4
solid and add'a second force due to 0.3
a "negative mass" filling the hollow.
0.2
(The positive mass and the negative
mass add together to produce the 0.1
hollow.) Let's use the second ap- .
0
proach.
Figure 13
- GMm D- GM'm
\= Oz
,,r-16_R1Zp,
This force is proportional to the segments that takes less time.
where M' = llSM because the mass that lies inside the sphere of Consider the path from A along d
spherical hollow has ll2 the radius radius r. and then to B as shown in figure 12.
of the sphere. Therefore the force on Since every cord requires 42 min-
the small sphere is:
/.^ utes for the trip, path wwill require
GM'm GP-w"m
-
/, = ------srnU = ., sinO I 12 thattime, or 21 minutes. P ath d
cMm(- I -r) f" -! r' + x will also require 21 minutes,
\-bz= d-^ lI-; 8(1 ^4 2 showing that path d requires less
\ -Rl'jdr)' G?iftr"mx (4nGom\ .lzv than 21 minutes.
: ------2--- I Iv

Part B asked for an analysis of a rz =

r\ ---]- 3 ) =
We can find the minimum path
tunnel drilled along a chord of the by finding the path d + xthat maxi-
Earth connecting points A and B as mizes x. First, we do some trigo-
shown in figure 11. At the position When the displacement is to the nometry.
shown, there is a component of the right, the force is to the left, so the w sin(Q+0)=-R
d+x
gravitational force along the tunnel. correct form of the equation is sinQ=7, ,

F: -kx.
Once again/ we see that the path coso =
through the tunnel is simpie har-
#
monic motion with the same pe- We now write down an expres-
riod (84 minutes) for a tunnel along sion for x.
an Earth diameter and also equiva-
lent to the period of an orbiting x = Rsin(O +ol- d,
satellite.
The tunnel would not be particu- x=Rsin(O+e)--w
larly feasible due to the difficulties cos0
of drilling through the Earth and 4gtg
the presence of friction, heat, and = Rsin(Q + o)- .
cos0
air resistance. If the Earth's molten
Figure 11 core doesn't present enough diffi- At this point we could take the
culties, we will also have to worry derivative of this equation and set it
about the walls melting and col- equal to zero,but a simple solution
lapsing. does not emerge. Alternatively, we
Part C asked if the straight tun- can solve it numericaiiy using a
nel provides for the fastest journey spreadsheet and finding.0 for any
from A to B? We found that the pe- grven Q. As a specific example, let's
riod for any chord is 84 minutes, or choose O = 20".
a one- way travel time of 42 min- From the graph in figure 13, we
utes. HoweYer, a chord is not the obtain a maximum value {or 0 of 45"
fastest path from A to B.It is best and a corresponding distance x :
to travel a curved path that passes 0.43R. The resulting time savings
nearer the center of the Earth. Find- can be determined by analyzing the
ing that curved path requires the equations for the simple harmonic
use of the calculus of variations. oscillation... but that's another prob-
Figure 12 Let's find a path with two straight lem.

3 4 JUIY/[tlotlsT zooo
 AT THE
BLACKBOARD I

$hal'inU a poiltl
by l. Sharygin

ROM TIME TO TIME YOU

may encounter problems where
you need to prove that three or
more lines meet in a point. For
example:
Problem. Three isosceles tri-
angles are constructed on the sides
of triangle ABC as shown in figure
1. Prove that the perpendiculars

Figure 1

dropped from the points A, B, and C Similarly, if we have to establish numberc. Then the locus of points
onto the lines BrC1, CrAr, and ArB, that three or more points belong to a M such that
meet in a point. straight line, we cattry to prove that
Here's a method that is useful in all the given points satisfy a condi- krlArM)2 +kr(,+rtwlz =k (.)
solving such problems: prove that tion and then prove that all the points is as follows.'
.
two of the given lines intersect in a of a line and only such points satisfy (a) a circle, a single point, or the
point that satisfies a certain condi- this condition (this line of reasoning empty set if k, + kr* 0;
tion, and then prove that aII points can be used for circles as well). (b) a perpendicular to the seg-
of the third line and only these Now let's see how we can find the ment ArArif k, + kr: O (and k, + 01,
points satisfy this condition. The locus of points that helps us solve A generalization of proposition 1
following well-known theorems can this type of problem. for several points holds. o
be proved by this method: three bi- Proposition 2. Let Ar, Ar, ..., Anbe a
o
sectors of the internal angles of any Fot,mulaling lle U'orusitions
fixed points in a plane, and lei k, (f,
triangle meet in a point, and three Proposition l. Let A, and Arbe kr, ..., k, (a11 k,* 0) be real numberi. o
perpendtcular bisectors of any tri- two fixed (differcnt) points in a Then the locus of points M such that a
CD
angle meet in a point. plane, and let k, kr, and krbe real the sum l
o

OUA|llTUll,IiAT TllI BI.A8I{BOARI I 35

 then
klArMlz + k2(AzM)z +,.. + knlAnM)z
is a constant is as follows: (rtto)'
r-t,(a,o)'-t (e D)'
=c.
=
(a) a cfucle,,a single point, or the t- 1.
^l T^2
emqt7 setif \+kr+,,, +kn*0;
(b) a line or tha enttue plane if k, The right-hand side is indepen-
+kr+...+kr=6. dent of M, so it is constant. Thus,
AID (MDlz is constant. Therefore, if C
Using assertion lb, we can prove
the following useful condition. > 0, then point M lies on the circle
Figure 2
Proposition 3, Let perpendiculars of the radius "uE centere d at D. If
be dropped from points A, Br, and C = O, then M coincides with D;
C, onto the sides BC, AC, and AB, and if C . 0, there are no points M
rcspectively, of triangle ABC. In or satisfying the conditions of the
der for these perpendiculars to meet problem.
in a point, it is necessary and suffi- The converse assertion-that is,
cient that the following ecluation DAT42 every point M of the set obtained
holds: satisfies the equation krlArMl2 +
Figure 3 kz(AzM)z = k-can be easily verified.
lA,B)z - lBC,f + lC,Alz -lAB)' It's sufficient to substitute the ex-
+ (B,cl'- (CA)': 0. (1)
Now let's prove the propositions pression for lMDl2 in ecluation (2).
This proposition implies another. we've formulated. We've considered the case k, , 0,
Proposition 4. Let p ery endicularc k, , 0. The case k, . 0, kr.0 can be
dropped fuom the vertices A, Br, Pnool ol fie pnopositions reduced to the previous one by re-
and C, of tiangle ArBrC, onto sides Proposition 1a. Let k, > 0 and k, > 0. versing the signs ol k1, k2, and k. In
BC, AC, and AB of triangle ABC Consider the point D that divides the case k, , 0, k, . 0 (or k, < 0, kz,
meet in a point. Then the perpen- the segment ArAzin the rutto krl. kr. 0), our reasoning can fo1low that of
diculars dropped from points A, B, Then ktlApl : kz4zD). Let the first case. However, in this case,
and C onto lines BrCr, ArCr, and ZMDA. = q (see {igure 2}. Choose we must take D outside segment
ArB, also meet in a point, any point M in the plane, and draw ArA, (see figure 3-try to perform all
Try to prove all these proposi- MA' MD, MAr.We can then aPPIY the computations). Equation (2) re'
tions. In the next section we'lI see the law of cosines to triangles MDA. mains true for all cases-we'll make
how we can use proposition 3 or and MDAr: use of this fact later.
proposition 4 to solve the problem Proposition 1b. The relation
formulated at the beginning of the lArM)z:lAtDl'+(MD)z
-ZMD'DA, cos tP, kllArM)z -ktlAzMl2 = k
article; then we'Il prove the propo-
sitions themselves. lArMlz:(AzDlz+lMDlz is ecluivalent to the relation
+ ZMD. DA, cos tP.
Soltlilion ol flte prohlem k
According to proposition 3, it's Multiply the first equation by k1, lAtM)z - lArM), = k,
sufficient to verify that the second by k2, and add them to
obtain Choose any point M on the piane,
IAB)r-lBrCl,+(CA1l2 and let D be the projection of M onto
- lAtBl, + lBCl)2 - (CA)z :0. k lA rM)z + krlArMlz =k t(AP)z the line ArAr.Then, we have by
+ krlArDlz + (k, + k2)lMDY. l2l
(See figure 1. Note that the equation Pythagorean theorem (see figures 4
in the statement of proposition 3 has Therefore, i{ M equation satisfies (. ), and 5)
here been multiplied by -1.) Indeed,
this ecluality holds since
ABr: BrC, CAr= AtB, BCr: CrA.
We can also use proposition 4. In
this case it's sufficient to note that
the perpendiculars dropped from
points 41,81, and C, onto the sides
of triangle ABC pass through the
midpoints of the sides of ABC and, At
therefore, meet in a point that is the
center of the circle circumscribed
about triangle ABC. Figure 4 Figure 5

38 JUI.Y/AUGllST 2OOI
 (ArMlz=(A,Dir+(MD)z, Proposition 3. Necessity. Let P be escribed circles all meet in a point.
(ArMlz=(A2Dl2+(MDlz. the point of intersection of the per- 5. Let the distances from a point
pendiculars dropped from points A, M to the vertices A, B, and C oL a
Therefore,
. triangle ABCbe a, b, and c/ respec-
spectively. The following relations tively. Prove that for any d * 0, the
(ArM)z - lAzMlz = lA P)2 - (AIDP : + follow from proposition lb:
'R1 distances to the vertices A, B, and C
(ArB), from any point of the plane (taken in
Thus, the problem is reduced to - lCA,)z = (PB)z - (CPlz, the same order) can never be
finding the points D on the line (B
rClz - @a sz = (PClz - (APl2,
ArArthat satisfy this equation. It is lc rAl, - IBC tlz = (PAlz - (BPlz. x7i, ^E
*d, ^Fe.
clear that such point is unique, and
a Adding up these equations, we 6.Let an equilateral triangle ABC
it can be easily found. Thus, the see that condition (1) is satisfied. and an arbitrary point D be given:
point Mmust lie on the perpendicu- Sufficiency. Let condition (1) be Let A1, Br, and C, be the centers of
lar to ArA, erected at point D. The satisfied and let P be the point of the circies inscribed in triangles
details of this discussion are left for intersection of the perpendiculars BCD, ACD, andABD, respectively.
the reader. dropped from points A, andB, onto Prove that the perpendiculars
The converse proposition is also BC andAC, respectively. It follows droppedfrom A, B, andC onto BrC,
trte: for any point on the perpen- from proposition lb that ArCr, and ArB, respectively, meet
dicular to ArA, the differcnce of the in a point.
squiles of the distances to A, and lArB)2 - (CAtlz + lB rClz - (AB 1)2 7,Let Ay, 42, A, and Aobe arbi-
A, is constant. The proof is left to : lPBlz - (APl2.
trary points in a plane. Prove that
the reader. Condition (1) implies that the there exist four numbers X 11 x 2t x gr
Thus, proposition 1 is proved. left-hand side of this equation equals andxo(not a1l of them equal toZero)
Proposition 2. We conduct the lB$12 - (C,Al2.That is, IBC)2 - such th-at xrlArMl2 + xz(A2M)z +
proof by induction. For n = 2, this (CtAlz = lPBl2 - (Apl2, which means
proposition has already been proved. that the point P iies on the perpen-
\(AzM)z + xolAoMl2 is constint for
any point M of this plane.
(For n = 2, the locus of points coin- dicular dropped from C, onto AB, 8. Let a triangle ABC be given.
cides with the entire plane if k, + k, which was to be proved. Consider all pairs of points M, and
= 0 and A, coincides with Ar.Thei, Proposition 4. The validity of this M, such that AMr: BMr: CM, =
for all-points of the plane, (ArM), - proposition follows from the fact AMr: BMr: CMr. Prove that all lines
(A2Ml2 = o.l that condition (1) is symmetric with MrMrmeet in a point.
Now we assume that proposition respect to A and A' B and Br, and C 9. A circle is tangent to side AB of
2 holds {or a certainn and prove that and Cr. triangle ABC and to the extensions
it is true for (n + 1). Notice that I{ n of sides AC and CB at points M and
> 2 and all k1, k2, ..., kr, k,
*, are dis- Exercises. N, respectively. Another circle is
tinct from zero, thenthere exist two 1. Use Proposition 3 to prove that tangent to side AC and to the exten-
of them such that their sum is not all three altitudes in the triangle sions of sides AB and BC at points P
zero. Let them be k, and kr. Con- meet in a point. and K, respectively. Prove that the
sider the point D constructed in the 2. Three pairwise intersecting intersection point of lines MN and
proof of proposition 1a and apply circles are given. Prove that al1 com- PK lies on the altitude of triangle
formula (2). Then the equation mon chords of any two of these ABC drawn from vertex A.
kL(ArM)z + kr(ArM)z + kr(ArMlz + ... circles meet in a point. 10. Two segments, AB and CD,
3. Prove that if the perpendiculars are given. Find the locus of points M
+ kn* r(An, rMlz = k
dropped from the points 41, A2, ..., such that the sum S*r,ro+ So"r, is
can be written as A, onto the lines BtBz, B2Bz, ..., BrB, constant.
meet in a point, then 11. Use the previous problem to
lk, + krl(DM)z + kr(ArM)z + ...
+ k, : prove that the midpoints of the di-
* rA, * Nlz k - klArDl' - k 16 rD)r. (nrer)' * (ersr)' + (a, ar)2 agonals of any circumscribed quad-
On the right-hand side of this gr)' +.. + (a,A,)' - rilateral and the center of the circle
equation, we have a constant/ the -(n 1e,sr\' = g. inscribed in it lie on a straight line
number o{ points on the left-hand 4. An escribed circle of a triangle (Newton's problem) (see "Kaleido-
side is reduced by one, and the sum is a circle that is tangent to one side scope").
of the coefficients remains the of the triangle and to the extensions 12. Prove that the locus of
same. By the induction hypothesis, of the other two sides (so that its points-such that the ratio of the
proposition 2 holds for the last center lies outside the triangle). distances from these points to r\Mo
equation. Therefore, it holds for (n Prove that the three perpendiculars fixed points of the plane is a con-
+ 1) points. Thus, proposition 2 is to the sides of a triangle at the stant different from l-is a circle
proved. points of tangency of one of its (called a circle of Apollonius). O

0lJArlrTllltil/Ar iltr BtAct(00ARI I ll

 AT THE
BLACKBOARD II

The litlle house olt lhe lundl'a

by A Tokarev

HE GREAT 16TH-CENTURY areas are built on piles instead of a cial physical parameter, the coeffi-
Italian architect Andrea Palla- conventional solid foundation. But cient of thermal conductivity. Natu-
dio (1518-i580) thought that this isn't enough. rally this coefficient is different for
any building worthy of public Are there any other ways to re- different substances. Thermal con-
approval must satisfy three recluire- duce the flow of heat to the ground? ductivity is highest for metals (which
ments. These are usefulness (and In this particular case, heat is trans- correspondingly have the highest co-
comfort), beauty, and durability. ferred by thermal conduction only; efficients of thermal conductivity),
Leaving aside the principles of use- therefore, we should make the piles whiie it's lower in licluids and much
fulness and beauty, 1et's talk about out of material with the lowest pos- lower in gases. For the purposes of
durability-in other words, the reli- sible thermal conductivity. What our analysis it isn't important how
ability and safety of buildings. material could that be? the coefficient of thermal conductiv-
Just about everybody knows that The thermal conductivity of vari- ity is determined and in what PhYsi
the construction of any building be- ous substances is described by a spe- cal units it's measured. We need only
gins with laying the foundation. A compare the capacity of various sub-
good, solid foundation is a token of stances to transfer heat. So we take a
relcrtive thermal
further success. But just what is a sttbstance reference book and construct a grid
conductivity
(tabte 1) that shows the coe{ficients
"strong foundation"?
Builders face many challenging gasoline 0.2
of thermal conductivity relative to
tasks, and one of them is laying that of water.
the foundation. This is especially cottoll 0.07 Take a good look at table 1.
dif f icuit in permaf rost areas. Among solid substances/ cotton
Many buildings in these areas are water 1 and cork have the lowest values of
subject to cracking due to settling thermal conductivity, while wood
of the foundation in soil that has at 0.04 and brick are next. Now, a pile
melted. must be durable and strong, so the
felt 0.1
Can we prevent thawing of the substances with the lowest values
ground under buildings erected in wood 0.2-0.6 won't make good piles. But what if
permafrost, or at least minimize it? we take some steel pipe and stuff
Let's try to solve this problem Tron t22 it with cotton/ felt, or sqme other
using the simplest physical laws and porous substance that contains a
ru1es. We begin by analyzrng the kerosene 0.2 lot of air? I think we've iound a
conditions of the problem. Why does solution!
brick 1.1
the ground under a building start to Piles shouldbemade of a durable $
thaw? Clearly because the founda- rce 3.7 soiid material and filled with a po-
tion transfers heat to it. So we need rcus substance. e=
to focus on the foundation. cork 0.07 Analyzing our result, we arrive at 9
The first thing to do is decrease this conclusion: due to its low ther- ft
the area of contact between the alcohol 0.3 mal conductivity, a pile of this de- \
ground and the foundation. This is sign will actually decrease the flow 3
why some buildings in permafrost Table 1 of-heat from the surrounding air to &

38 JUIY/AUGUsT 2ooo
.j

OUAIITUT AT T11I BI.ACI(BOABO II 3g

the ground (that is, downwardl dur- thermal conductivity but also by
lleat cap\city,
ing the warm season. However, a convection. How would such a Pile sLlbStlnce
kl/(kg K)
properly constructed pile can do work?
even more. It would be nice if in In winter, the upper layer of fluid
gasoline t.4
winter, when the temPeratures are will be cooled. The cold fluid has
well below freezing, the piles could higher density than the warm fluid, water 4.2
cool the ground (that is, they could so it will sink. Warmer, less dense
transmit heat upward). This would fluid layers will rise and release heat a1r 1

add strength to the soil and decrease ' to the surrounding air. Then this
portion of the fluid will be replaced glycerin 2.4
its thawing during the following
summer. by cold fluid from the bottom, and 2.t
kerosene
Let's try to formulate more Pre- so on. As a result, the lower part of
cisely the details of the physical pro- the pile and the adiacent soil will be nTotor oil ),.9
cesses in a pile and the surrounding cooled to the temperature of the sur-
soil in summer and winter. During rounding air. Notice that we don't Table 2
the warm season/ the upper Part of need to construct any special refrig-
a pile is heated due to contact with erators and waste energy to cool the ways greater than the mass of a gas
the warm air. Gradually the lower foundation-everything is done "at in the same volume, we prefer to use
part of the pile, buried in the ground, the expense" of the naturally occur- a liquid as our filler for the pile. Now
is also warmed. The less the lower ring cold winter air. let's compare the heat capacities of
part of the pile (and the soil sur- In summer, the upper layers of various iiquids {we'11 use a reference
rounding it) is heated, the better. In the fluid in the pile will be heated by book again). We can see from table
winter the air cools the upper part of the air. But being less dense, they'll 2 thatnot only density but also heat
the pile. Gradually the lower part of stay in the upper part of the piIe. As capacity is higher in liquids com-
the pile and the adjacent soil are also a result, no convection occurs in pared to gases. Thus the piles should
cooled. The colder the ground gets, suffIrner; the heat will be transferred be filled with a lic1uid.
the better. downward only because of thermai What liquid is the best for this
Therefore, a pile should have the (molecular) conductivity, which is purpose? Although water is cheap
following properties: very small in fiuids. and readily available, it won't do: in
(a) if the temperature of the upper Since this type of pile conducts winter rt fueezes. Both giycerin and
part of the pile is higher than the heat poorly in summer, the temPera- motor oil thicken at low tempera-
temperature of the lower part, the ture of the surrounding soil will in- tures, which makes convection inef-
pile should conduct heat very crease only insignificantly. ficient. This leaves gasoline and
weakly; Thus we've arrived at another kerosene from our list. The freezing
(b) if the upper part of the pile is important conclusion: the pile, con- point of both these liquids is less
colder than the lower part, the pile structed of a durable material, than -50'C, so either could with-
should conduct heat efficiently. should be filled with a fluid (gas or stand the low Alaskan and Siberian
In other words, downward heat liquid). temperatures. Of the two we should
transfer should be small, but the One question remains: With probably choose kerosene, because
upward heat transfer should be what exactly should we fill the pile? it's cheaper and has a higher heat ca-
large. The pile should be a "heat In winter, the properties of the fluid pacity than gasoline.
semiconductor." aten'tcrucial, since there's plenty of At long last we can formulate the
It's known that heat exchange in time to cool the ground to the am- "ttnal answer" to the problem:
solids is due entirely to thermal con- bient temperature. In summer, how- To decrease thawing of the
ductivity, which doesn't depend on ever, it's very important that the gr oun d b ene ath buildings in p erm a -
direction. So an ideal pile cannot be ground be heated as little as pos- frost areas, they mustbe erected on
completely solid. Our previous sible. Therefore, we should use a piles; the piles should be made of a
model of a metal pipe filled with a fluid whose temperature rises the hollow durable material and filled
porous material is of no use either, least when heated. From this well- with kerosene.
because its porous interior will con- known formuia- We should note that this method
duct heat weakly not only in sum- of strengthening frozen foundations
q = cmltr- trl (decreasing thawing of the ground
mer but also in winter/ when it's
necessary to cool the ground. find that the change in tem- under buildings) isn't just idle specu-
What if we fill a strong and du-
-we
perature (tr-t1l depends not only on lation. It was calculated theoreti-
rable pipe (or some other empty the amount of heat transferred but cally and tested by the constmction
metal shell) with a fluid-that is, a also on the heat capacity c of the industry in permafrost areas. As ex-
licluid or a gas? In this case, the heat fluid and its mass m. pected,
-duced
construction costs were re-
is transferred not only by molecular Since the mass of a licluid is al- significantiy. O

JUI-Y/AUGl,$T
 AT THE
BLACKBOARD III

The eniUmalic lnaultelic lorce

by E. Romishevsky

I T IS KNOWN FROM the advance of a right-

I .*p.r,.r.e that, In gen- hand screw whose head
! .rrt, the iorce acting on lies in the plane of the
I a point electric charge q vectors v and B, and
placed in electric and mag- which is turned through
netic fields depends on the the smaller angle from
position of the charge and the vector v to the vector
on its velocity. Usually, B (figure 1). The magnetic
this force is resolved into force E- is normal to both
two components: the elec- of the vectors v and B.
tric force E,= eE, which is The total electromag-
independent of the motion netic force F = F. * F_ act-
of the charge, and the mag- ing on a particle with a
netic force F_, which de- charge q is called the
pends on the velocity of "Lotentz force." By mea-
the charge. In this article suring the Lorentz force
we discuss the nature of acting on a test charge of
this magnetic force and its known sign (positive or
interplay with the electric negative), one can deter-
force. mine the magnitudes and
At every point in space directions of the vectors E
the magnetic force is per- and B.
pendicular to the velocity Note that the mag-
of the electric charge. The netic force does not affect
magnetic force is also per- an electric charge at rest.
pendicular to a special direction, This property of the magnetic force Another important feature of the
which is defined at every point as can be described in another way us- magnetic force is its direction: it is
well. The magnitude of the mag- ing the concept of the magnetic always normal to the velocity, so
netic force is proportional to that field. The direction of the magnetic that it performs no work while act-
component of the velocity of the field coincides with the special di- ing on a charge. Therefore, in a con-
charge which is normal to the "spe- rection in space. stant magnetic field thq kinetic en-
cial direction" mentioned above. The magnitude and direction of ergy of a charged particle does not
the magnetic force are determined change, whatever motion this par-
by the formula ticle undergoes.
As an example, consider the mo-
F^: qvB sin o, (, tion of two particles with opposite
where v and B are the magnitudes of charges +q and-q that have different
o
the velocity and the magnetic field, masses Mr = 2m and M, = m. Ini- C_
while the unit vector ( (according to tially, the velocities of these par- o
a
the right-hand rule) serves to indi- ticles have the same value vo, whose o
G)
cate the direction of the magnetic direction is perpendicular to the D
o.
Figure 1 force. This direction coincides with boundary of a homogeneous mag- CD

0llArurl,lIi4/Ar ilrr StAct(80Afl0 ilt 4l

 mogeneous magnetic field is capable
of separating, in time and space, par-
ticles that move in the same beam
but that have different masses and
charges. This property is used in
mass spectrometers/ which can
separate isotopes (atoms of the same
charge but different masses). the magnetic field decreases with
. Moving charges (that is, electric distance as lf P, just like the magni-
currents) generate magnetic fields. tude of the electrostatic field gener-
Numerous experiments with mag- ated by a point charge. The analogy
netic forces yielded a simple law between the electric and magnetic
that gives the magnetic field B gen- fields is not universal: the magnetic
erated by a point charge q moving field has no "sources" and "sinks,"
with constant velocity v that is so that the magnetic lines are always
much less than the speed of light c. closed. Such a physical vector field
This law can be written as has specific features and is referred
Figure 2 to as a vortical or solenoidal field.
B= I =Q"!.nA'-t
1.
netic field B (figure 2; the vector B is Now let us consider another ex-
normal to the plane of the page and
4xtnc' t' ample. Suppose that two fairly mas-
directed away from the reader). where u is the angle between the sive point particles 1 and 2 with
When the positively charged particle velocity v of the charge and the ra- equal charge q move parallel to each
enters the magnetic field, the mag- fius vector t drawn from the charge other with the same nonrelativistic
netic force F^ = evsB, which is ini- to the observation point; ( is a unit velocity v (figure 5). Each particle is
tiaily directed upward. The negative vector obtained by applying the affected by a repulsive electric (Cou-
particle experiences the same mag- right-hand rule to the vectors v and lomb)force.Q : qE andanattractle
netic force, but this force is initially r (figure 3). The constant ll,eoczl is magnetic force -fu : qvB (the veloc-
directed downward. Each particle usually denoted by po and is called ity of one particle is normal to the
describes a semicircle, after which it the magnetic permeability of free magnetic field generated by the
leaves the region of the magnetic space. other particle). Let us compare these
field. The radius of each circle can be By multiplying both sides of this two components of the total electro-
found from Newton's second law: formula by the number of electrons magnetic (Lorentzl force that act,
AN: nAJS in a segment of wire of say/ on particleZ:
qvoB: - M"3
R t
length Al, electron density n, and
F^2
cross-sectior.al area S carrying an - QvBzt .

whence - A
i, qEzt '
R:M'u. where B^ and Er, are the magnetic
qB and electric fields generated by
charge 1 at the position of charge 2.
The angular velocity of the particle Inserting the corresponding expres-
and its period are

61 :vo - QB
RM
Figure 3
2tr ZnM
electric eurrent I: c1nvS, we obtain
oqB the famous Biot-Savart law for the
Clearly, the positive pafiicle lM, contribution AB to the magnetic
= 2m) describes a semicircle whose field generated by an electric curent
radius is twice that of the negative element IAJ:
Particle lMz = ml, which moves in
the opposite direction. The heavier on=Pl*rye
4n r:
positive particle will return to the
no-field region in a half-period In this case the lines of magnetic
which is twice as large as the core- field are concentric circles drawn
sponding interval of time for the around the trajectory of the moving
lighter negative particle. Thus, a ho- charges (figure 4). The magnitude of Figure 5

42 JU[YiAttGttsT zooo
sions in this ratio, we get erence frame is replaced by another. Second, there are situations where
At the same time, the total Lorentz a "negligible" magnetic force is really
F^2:F"2 = ,1r, J-tq -! " force, just like any other force, does a single unbalanced force in a physi
4neoc'r' 4ntst' not depend on the choice o{ nonrela- cal system. This is the case for elec-
c c' tivistic inertial frame. Therefore, in trons moving in a conducting wire.
=V-:C-.
the reference frame in which the Here, there are no net electric forces
This ratio shows that at nonrela- magnetic component of the Lorentz as a result of the almost ideal balance
tivistic speeds the magnetic {orce force disappears, the electric compo- of the negative and positive charges
produced by moving charges .is nent of this force must change to in a conductor. Recall what a huge
much weaker than the electric force compensate for such a loss. In other number of charged particles partici-
acting on the charges. In other words, dividing the total Lorentz pate in generating an electric curent
words, under these conditions the force into electric and magnetic com- in metals-about 1023 elementary
magnetic force is a minor contribu- ponents is meaningless without charges in one cubic centimeter!
tion to the total electromagnetic specifying a reference frame. This enormous number produces a
force. The last example raises the ques- very large magnetic force-for ex-
What will happen if we choose an tion of whether it is reasonable to ample, in electric motors.
inertial reference frame that moves study and take into account such Third, sometimes the electric
with the same velocity v as our par- relatively sma1l magnetic forces. Of charges move under the action of
ticles? In this moving reference course/ it is worthwhile, and here various combinations of electric and
frame the particles are at rest/ so are the reasons why. magnetic fields generated by differ-
that their magnetic fields and mag- First, the ratio obtained is valid ent sources. In general, the relation-
netic forces disappear! also at relativistic speeds v - c. In ships between electric and magnetic
Well, this paradox could have been this case the magnetic forces are forces can be quite versatile, includ-
expected: the magnetic component comparable with the electric ones. ing the case when the magnetic
of the Lorentz force depends on the For example, they play a major role force dominates over the electric
velocity of a charged particle, and in a rapidly moving beam of charged one. Therefore, magnetic knowledge
this velocity changes when one ref- particles. is power in itself, isn't it? O

Number
From Ahmes to Cantor
I'lidhat Gazal€
We might take.numbers and counting for granted, but we
shouldn't. ln his companion volume lo Cnomon: From Phoroohs
to Froctols, Midhat Cazal6. introduces us to some of the most
fascinating and ingenious characters in mathematical history.
As he deftly blends together history, mathematics, and even
some computer science, we discover what "number" truly
means. Number is for all those who enioy mathematical recre-
ations and puzzles.
Cloth $29.95

June 8,2OO4-Venus in Transit

Eli Maor
ln 2004, Venus will cross the sun's face for the first time since
1882. Here Eli Maor tells the intriguing tale of the five.Venus
transits observed by humans and the fantastic efforts made to
record them. This is the story of heroes and cowards, of
reputations earned and squandered, told against a backdrop of
phenomenal geopolitical and scientific change. With a beguiling
storytelling style, Maor guides readers to the upcoming Venus
transits in 2004 and 2012 and provides a wealth of scientific lore
and joy.
Cloth $22.95

Princeton University Prgss Ar FrNE BooKSToRES oR ,ALL800-777-4726 . ww,v.pup.pRrNCEroN.EDU

Circ1e No. 3 on Reader Service Card

OUA[IIU]llliAI TIII BI.A8I(BOARO III 43

 IN THE LAB

Can you carry ttualer in a sieue?

by A. Dozorov

O YOU KNOW THE STORY have the minimum number of mol- You can use this trick to pilot the
of the little ant who was in a ecules at its surface. As a result, needle all around the saucer. Ma-
big hurry to get home? Many the surface layer of a liquid is neuver it to the very edge of the sau-
creatures heiped him. For ex- "stretched" much like cellophane. cer and watch what the needle does.
ample, a water strider carried him Let's call this surface a "film" (in Keep in mind that the soap spreads
across a river. Maybe you've seen quotation marks). across the water surface verY
this insect. The water strider stands We see that not only water strid- quickly, so don't {orget to change
calmly on the surface of the water, ers but even denser bodies (such as the water in the saucer from time to
which sags slightly under its weight. a metal needle) can stay on the sur- time. You can replace the needle
Why doesn't this insect sink? And face and not sink. They don't swim with a match and repeat the experi-
can water really " sag" ? or float in the usual sense of the ment. (A needle that keeps sinking
It turns out that the surface layer words-they're held up by the sur- is an unnecessary complication in
has a number of unusual properties. face tension of the liquid. However, your experiments.)
We can investigate them in some if we test thicker and thicker Take two matches andplace them
simple experiments. needles, we'll eventually find one carefully onto the water surface par-
l. The watu surface can suppofi whose weight is greater than the al1e1 to each other. What happens?
vafious obiects. supporting force of the surface ten- The matches are drawn toward each
Pour some water into a saucer. sion. Of course, this needle (and other (figure 1). Pull them apart and
Take a needle and place it carefully heavier ones) will sink to the bot- touch the water on both sides of this
on the surface-it doesn't sink. If tom. It's interesting that the length pair with the tip of a third match,
this experiment failed, don't give up. of the needle has virtuaily no effect which you've rubbed with soap, as
Rub the needle with your fingers (or on its ability to "float." before. What do you see now?
oil it slightly, or rub it with a 2. The tension of a suf ace 'fiLm" Using the same principle, you can
candle). Repeat the experiment and depends on the liquid used. make a number of simple toys and
look closely at the surface. Now do Place a needle on the water sur- entertain kids with them. Slit the
you see that the surface is bent? It face. Take a wooden matchstick and end of a match and skewer a piece of
looks as if the needle is lying on a cut off the head. Rub the end of the paper with it (figure 2). Soak the pa-
film. match with soap and touch the wa- per in the soapy soup that often col-
A rather good comparison is that ter about 1 cm from the side of the lects in soap dishes (or make some
surface layer of a liquid is similar to needle. The needle will immedi- "sorfp" yourself, if there isn't any).
a stretched piece of cellophane (al- ately "jump away" from the match. Now lay this match "ship" on the
though the specific properties of the Why? Weil, you created a soap solu- water-it starts "sailing." Did you
surface layer are quite diiferent from tion near the needle when you
those of stretched cellophane). Let's dipped the matchstick in the water.
try to guess why. The molecules of the soap solution
A molecule within a liquid is don't attract the needle as strongly
surrounded by other molecules, as the pure water molecules on the
which pull it eclually in every direc- opposite side. Since unbalanced
tion. By contrast/ molecules of the forces act on the needle, it moves in
surface layer have no molecules the direction of the greater force. In
above them, so they are attracted other words, the surface tension of
only by the molecules below them. pure water is greater than that of a
It looks as if the liquid "tries" to soap solution. Figure 1 Figure 2

44 JUIY/AUGUsT 2ooo
notice the direction in which it some water in it and And now the moment of truth:
sailed? Now slit a matchstick at plug the bottom with how would you answer the clues-
both ends and your finger (figure 5). tion posed in the title of this ar-
slip pieces of pa- You'1l see that the ticle? Can you carry water in a
per into the slots water surface in the sieve? WeII, let's take a sieve and
as shown in fig- tube is curved. This spread butter on it, or even better,
ure 3. Soak these curvature-ca1led a rub it with a candle. Pour some wa-
pieces of paper in "miniscus"-is ex- ter into it-the water doesn't run
the soap soup. plained by the fact outl It's supported by a surface
This match will that water molecules ttfilmtt that forms because the wa-
Figure 3 rotate on the sur- are atttacted more ter doesn't wet the edges of the
face like a propeller. Figure 4 shows strongly to the walls Figure 5 sieve's tiny openings. If you donlt
a " gurt" cut out of thick paper. To of the tube than to each other. In have a sieve, you can do this experi-
"fire" the gun, touch the water sur- this case we say that the liquid wets ment with a can with a small hole
face at point A with the end of a the surface of a container. punched in the bottom.

ffi
match that you've rubbed with soap. Now let's do one more experi- As we saw/ a liquid that doesn't
ment. Pour tea from a cup, but leave wet the surface doesn't spread out
some tea and a few tea leaves at the but collects itself into a drop. In
bottom. Carefully touch the surface this case, the smaller the drop, the
of the liquid with a teaspoon or a nearer its shape approximates a
match and watch how the surface sphere. Why? Due to the strong
quickly " crawls" upward, drawing mutual attraction between mol-
the tea leaves along with it. ecules in such a liquid, the drop as-
Figure 4 4. Not all liquids "cling" to the sumes the shape with the least sur-
waLls, and it doesn't happen in ev- face area. As a rule, this is a sphere,
Try testing other substances in- ery pipa. which is easily (and often) demon-
stead of soap. Here's a nostaigic There are sases when a liquid in strated in the weightlessness
scene from summer camp: a bunch a capillary doesn't rise-not only aboard an orbiting spacecraft. If an
of kids are gathered around a little that, but the miniscus is curved in astronaut releases water from a
puddle. They've taken wood chips the other direction (it's convex). container (one can't "pour it" as
and rubbed their ends with resin Why is that? It's because this par- one does on Earth), it immediately
from fir trees. Now they're holding ticular liquid doesn't wet the surface assumes a spherical shape.
races-thc chips slide swiftly along of the walls-the mutual attraction The brief weightlessness of mo1-
intricate paths on the water surface. between molecules in the liquid is ten drops of metal as they fa1l from
3. Sttrface tension can raise the stronger than the attraction between a high tower has been used for gen-
liquid rather high. these molecules and the wal1s of the erations to produce pellets. The
Take a glass tube with a very nar- tube. This is how mercury behaves drops become spherical as they fall
row internal diameter {much less in a capillary (figure 5). and stay that way long enough to so-
than 1 mm)-a so-called " capillary lidify as spheres.
tube" (or just "capillary" ). Lower You can do a similar experiment
one end into a 1ar of water and at home. Tilt a burning candle and
watch the water rise in the capillary pore the melted wax into a basin
to a height greater than the water filled with cold water-you'11 get
level in the jar. The thinner the cap- small wax pe1lets. Hold the candle
illary, the higher the water is iifted as close to the watet as you can/ so
in it. If you've ever had blood drawn that the wax solidifies right at its
from your finger at the doctor's of- surface.
iice, -vou've seen how the nurse col- 5. Sometimes surface tension is
lecrs it in a capillary tube. This Figure 6 so strongyou can literally "feel" it.
"capillarr- action" can be observed Take two identical plates of glass.
e \ cr\-\\-hcrL, You can see it in tea Collect some water in a pipette. Clean them carefully and put one on
rising rn tl-ie trnl' holes in a sugar Carefully release one drop onto top of the other. You can easily sepa-
cube, in ke rosenc rrsing in the wick clean glass and another onto a piece rate them again. Now wet one plate
oi an oil larlp tn \\-ater absorbed of buttered bread. The first drop with water and put one atop the
from soil l',r- tr. r' , r: or r arious spreads out on the glass,
while the other again. Try to pull the plates
plants, and so on. other one maintains its round shape. apart (without sliding one over the
A mOre mOJc.t c\:,r'lln(nt Can So we conclude that water wets other). It's not so easy, is it? That's
bc made wirh a rh..^-r :-r.'. Pur glass but doesn't wet butter. surface tension at work. O

OUA[IIUltJl/IN IllI I.A[ 45

 LOOKING BACK

[l'nsl Ahhe altd "Cal'l IEi$$"

by A. Vasilyev

demands of science and technology He defended his doctoral thesis in

led to the invention of devices for 1851, and in 1853 he became an as-
observing various objects, which re- sistant professor at the University of
AHl#t';#i,,tr#mii';;
work of Abbe and Carl Zeiss (1816-
sulted in the rapid development o{
applied optics and optical engineer-
|ena. Abbe lived in |ena for 35 years
and brought worldwide fame to the
1888), a brilliant engineer and inno- ing. The production process for sci- town. As a professor at the univer-
vator in the commercial production entific instrumentation was radi- sity, Abbe focused all his attention
of optical instruments, standards in cally improved. Small workshops on optics: the theory of optical in-
optics were raised to a level that has were replaced by scientific and struments/ analytical and math-
remained essentially unchanged to manufacturing conglomerates such ematical optics, and the technology
the present day. as "CarLZeiss" (its modern name). of optical experimentation.
Theperiodfrom the middle of the Abbe's work in this firm contributed The period of Abbe's life from
19th century to the beginning of the to its prosperity and made it possible 1865 to 18BB was closely tied to the
20th was a time of revolutionary to produce outstanding optical in- work of the legendary German opti-
discoveries in various fields of natu- struments. ca1 engineer Carl Zeiss.In 1846 Zeiss
ral science, which enriched human- Ernst Abbe was born in Eisenach founded a workshop in |ena that ini-
kind with sophisticated new tools in 1840, but he finished primary and tially produced magnifying glasses
and methods of investigation. The secondary school in |ena, where he and primitive microscopes. Very
entered the loca1 university. He
later transferred to Gottingen. At
that time Wilhelm Edward Weber
{i804-1891), George Friedrich Ri-
emann 11826-1855), and other re-
nowned mathematicians worked in $l :.'!

Gottingen, and personal contact

with them helped Abbe develop his ;i
.,!
considerable mathematical gifts.

.,: .:

4$ JUrY/AUGUsI 2ooo
 soon/ howeuer, Zeiss microscopes
received high praise and were widely
used by scientists and engineers due
to the exceptional workmanship of +
the lenses. Beginning in 1858, Zeiss =_->
produced sophisticated microscopes, .+
and later he added other optical in-
struments to his list of products.
Zeiss continually strove to "base the ------------>

practical design of microscopes en- ---------.-.>

tirely on scienti{ic theory," so he in- +

_____>
vited a number of outstanding spe-
cialists in applied optics to work with
him, Ernst Abbe being the brightest
star among them.
By that time, optical toois had
o
been manufactured in Europe for
FigUfe 1 . gbbe's microscope diffraction theory: p,p, is the plane of the
three centuries. However, produc- o!l9ct, FF is the focal-plane of obiective OO, p2pz ii the image planei. qo is the
tion was based mainiy on intuition diffraction angle, and Ao is the diffraction maiiita in the foiaf plane.
and traditional workmanship. One
of the inventors of a two-lens micro- of the incident light. This phenom- tions. But Abbe's biggest achieve-
scope with a biconvex objective and enon is known as dispersion. ment was discovering the limits
a biconcave eyepiece was the great From the 17th through the 19th imposed on designers of optical sys-
Italian scientist Galileo Galilei centuries, investigators tried not tems by the wave nature of iight.
11564-1642J. The father of the mod- only to improve the quality of the Abbe explained how a lens forms
ern microscope is Cornelius Drebbel images formed by microscopes, but the image. First an interference pat-
11572-I 634), whose microscope con- also to construct a microscope with tem is formed in the plane perpen-
sisted of a biconvex objective and a the greatest possible magnification. dicular to the axis of the iens. This
plano-convex eyepiece. A funda- It's known that the magnification of is a system of alternating maxima
mental improvement was made by a microscope increases as the focal and minima of illumination, which
Robert Hooke (1635-1703), who in length of its objective decreases, so plays the role of a diffraction grating.
1653 inserted the third "collecting" opticians started to work with short- The light flux passes from the lens
lens between the objective and the focal-length objectives. In addition, through this grating and interacts
eyepiece. The next step was taken in the resolving power of a microscope with it. Only then does an image ap-
1716by Hertel, who added atotat- depends on its aperture-that is, the pear a short distance from the plane
ing stage with a mirror beneath it to angle between the outermost beams of the grating, which can be seen on
reflect light. This 1ed to better illu- from the object to the edge of the a piece of frosted glass or photo-
mination of the object and a clearer objective. An aperture of almost graph. This is how an image is
image. This is essentially the micro- 180o was achieved by the middle of formed with one lens. In a micro-
scope design that is used today. the 19th century. However, the scope/ however, according to Abbe's
Any further improvement in im- short-focai-length and wide-aperture theory, the image is obtained in two
age quality would have to be made objectives suffered from even greater stages/ shown schematically in fig-
by eliminating defects in the opti- aberrations. ure 1.
ca1 system-above all, spherical Attempts were made to improve In the first stage, the light illumi-
and chromatic aberrations. Because the performance of microscopes and nating an object PrP, falls on the
of spherical aberration, paraxral calculate their magnification on the microscope's lens after being scat-
rays (rays traveling near the optic basis of geometrical optics. It turned tered and diffracted by tlre details of
axis) pass through different parts of out that geometry could not com- the obiect, so that the structure of
a lens and cross the optic axis at dif- pletely explain the process of image the light beam is determined by the
ferent points, causing the image of formation in microscopes. This fail- object. After passing through the
a point source to look like a ure directed Abbe's attention to objective of the microscope, the
nonhomogeneously illuminated physical optics. light beam produces a diffraction
disk. Chromatic aberation causes Abbe published his studies on mi- pattern in the focal plane FF, which
a ray of white light to split into a croscope design in 1873. In these is a system of illumination maxima
number of rays of different colors, papers he considered the role played whose angular sizes depend on the
which cross the optic axis at differ- by the objective and the eyepiece in structural details of the object. The
ent points because the focal length image formation. For the first time directions to these maxima are de-
of a lens depends on the wavelength in optics, he classified the aberra- termined by the condition nd sin Q

OUAlllTUllll/IOOI(IlllG sACI( 4t
= ki",where n is the refractive index the order of d. The larger sin u is with a substance whose refractive
of the medium, d the characteristic relative to ),old, the more high-order index is greater than that of air. In
size of the object's details, q the dlf- spectra will contribute to the image 1878 Abbe and Stephenson made a
fraction anglp, k = 0, l, 2, ... the formation, and greater detail will microscope in which cedar oil was
number of a particular maximum, appear in the image. placed between the object and objec-
and ), the waveiength. Usualiy an object is illuminated tive. Their efforts met with success:
In the second stage, the illumina- not only by light beams passing this instrument improved the resolv-
tion maxima are considered point along the optic axis, but also by ing power by one-third.
sources emitting eoherent beams of beams atlarger angles, and this im- Of particular interest are Abbe's
light. These beams mutually inter- proves the resolving power. If the il- ideas about improving the resolving
fere behind the focal plane of the luminating beam makes an angle cx power of microscopes by decreasing
objective and produce an image of with the microscope's axis and dif- the wavelength of the light used to
the obiect in the PrP, plane. Abbe fracts at an angle o,s, the condition form the image-speciIically, the pos-
called the pattern in the focal plane for the maxima takes the form sibility of using ultraviolet light. This
of the objective the primary image sin o,o - sin ot : kLld. idea was realizedin one of the micro-
and the pattern in the linked plane In order for the first spectrum to scopes made by Abbe's colleagues in
the secondary image. enter the obiective completely, the the Carl Zeiss firm not long before
To obtain the correct image of an following requirements must be the death of the great inventor. Later
object, the secondary image must be met: cr = -1i, ao = u, k: 1. Also, such microscopes helped in studying
formed as a result of the interaction 2 sin u >)"ol(ndl or d>i.olQn sin u). the structure of DNA and RNA, the
of the beams emitted by all the Abbe called the value A = n sin u the large information-bearing molecules
maxima of the primary image. Of "numerical apertute." According to in living organisms.
particular importance are the first- Abbe's theory, the numerical aper- Abbe aiso devoted a great deal of
order maxima situated at small ture determines several important attention to correcting aberrations in
angles to the principal axis and pro- properties of a microscope-for ex- optical systems. Since the various
duced by the largest and usually ample, the brightness of the image zones of a simple lens produce an
most important details of the object and the degree of similarity be- image of a plane element with differ-
being examined. The maxima corre- tween the object and its image. The ent magnifications, the images of a
sponding to large angles are pro- larger the numerical aperture of a point source formed by various zones
duced by smaller details of the ob- microscope, the smaller the details coincide only at the optic axis of the
iect. Minute details of the object in the object that it can resolve. system/ while the sharpness of the
(smaller than the wavelength of Abbe's theory says that it is impos- image degrades sharply outside this
light) cannot be seen at all, because sible to see objects in a microscope axis. Abbe showed that allthe zones
the waves diffracting off such small that are smaller than half the wave- of an optical system magnify an ob-
details do not reach the screen/ even length of the light illuminating ject to the same degree as long as the
through an objective with the larg- them. Abbe confirmed the validity "sine condition" is met. This re-
est possible aperture. This sets a of his theory by experiment (in quires that for ali rays emerging from
limit on the resolving power of a which the objects examined were a point on the axis of the optical sys-
microscope : d 2)" = 7"ol n, where ),0 is the absorbing $atings), and in 1BB7 tem and then collecting after rehac-
the wavelength of hght in vacuum. he formulated a strict mathematical tion at the point of an image, the ra-
Usually there are no obstacles for theory of the microscope. tio between the sines of the angles of
light inside a microscope, so the In order to improve the resolving the respective rays with the optic
number of diffraction maxima pass- power of microscopes, Abbe tried to axis must be constant:
ing through the objective is limited increase the numerical aperture. sin ur/sin ur: Knrlnl,
only by its mount. The smaller the There were three ways to do this:
object or its detail, the larger the dif- increase the angular aperture, in- where n, arrd nrare the refractive in-
fraction angles it produces (half this crease the refractive index of the dices of the media on the object and
angle is called the aperture u) and medium, or decrease the wave- image sides, and K is the ihagnifica-
the larger the opening of the objec- length of the light beam. Even at the tion of the optical system.
tive must be. beginning of his optical research Two points that have no spherical
If the aperture is less than the dif- Abbe realized that microscopes had aberration, and for which Abbe's sine
fraction angle Q, corresponding to reached their limit in angular aper- condition is valid, are called apla-
the first-order spectra (that is, if sin ture and that this was a dead end. natic. Abbe showed that only one
u < sin qL = ?"oldl, onlY raYs from the The second approach looked pair of aplanatic points exists on the
central maximum will pass through more promising: Abbe proposed in- axis of an optical system. He also
the objective into the microscope, creasing the medium's refractive found a simple method for determin-
and we won't see an image corre- index n by filling the open space ing the degree to which the sine con-
sponding to details whose size is o{ between the obiect and objective dition is satisfied. He drew the pat-

4 I JIrY/[ttGll$I 2ooo
 to the idea that an optical system ber of new optical instruments/ as
must include special diaphragms well as organizing scientific research
that limit passage of light rays. He aimed at producing new types of op-
showed that to form an image, an tical glass. The Carl Zeiss firm de-
optical system needs only those rays signed and produced prismatic bin-
that pass through the device to the oculars, new types of photographic
image without adelay, whereas rays lenses, refractometers (devices that
Figure 2. lbbe's pattern for testing that pass through only a part of the measure the refractive index of a
tlte Abbe sine conditictt. optical system (held back, perhaps, substance), and various optical de-
by the lens mount) are not only use- vices to measure angular and linear
tern shown in figure 2, which is less, they're harmful. values. A11 these instruments raised
viewed with the optical system be- Abbe performed a great service by the standards of the optics industry
irrg tested. If rhc sinc rcquirement is developing and constructing a num- to a higher level. o
met, it rs irossible to find a location
ior the Lratrcrn such that the ob-
server sees it as a recrangular grid.
Abbe tested lr-tan)' mlcroscope objec-
tives made b1- trLal-and-error by the
old masters and tound that the sine
condition was r.ahd ior all the good
objectives. Today, Abbe 's sine con-
dition is always taken inr,'account
in the design of any optrcal system.
In his struggle with chrrturatic
aberration, Abbe spared no eiiort in
persuzrding the glass workshops to
produce new kinds of optical glass
rvith certain properties. To compare
the properties of various t.vpes oi
optical glass, Abbe proposecl the fo1-
lowing method: selcct a number itf
reference wavelengths in thc visible
range of the spectrum and use the A High-Performance,
concept of rclative dispersion {de-
fined as the ratio of the refractive in- Hand-Held Microsope for Under $30?
dices corresponding to the chosen
wavelengths). In applied optics the
(ltt About Time!)
We borrowed a brilliant idea from 1/th-rertr', - -',,.,.: - ,'
value y is known as the Abbe num-
Antonyvan Leeuwenhoek,perferted itwith l1s.-::" -', :."'. .- : -
ber.In 1873 Abbe managed to make
the first objectivc in history that was
achromatic {or three colors. The co-
(E m*reffiKffi€srtr(opg'"
rodicol new personal nicrosropethot outperforms troditionoi nioostopes rcsting 20 times nuch!
incidence of the foci for rays of three A os
--::,-:'-sio-reopttca
different wavelengths was achieved
Weusedadvancedcomputer ardedoLrta: t:r t" .-:::":' ,-..' -:.
system e\/er manufactured,then mounted ii in durability,and ease of use
by using various types of glass with The P5-1 50 Pocket5cope rs a superb 21 st c y price!
different Abbe numbers. Abbe called The PS 150 PocketScope delivers exiraordin;ry por,,'r: and value:
such objectlves "apochromatic.// In . 5uperior 1 50X Iens shows deiari qcrri"raiif reluiring 300X.
1886 Abbe managed to clesign and . Bright image uses arnblet ii.ahr. \i etennr oc,,ter required
prodllce an apochromirtic objectrr.e
rn rihrch i.,-.th srherrcal and chro-
rrratlc aberratt,,ls r\ ;:e',-irru..llv
ually indestructible
clrllrrnar.J ]r '... ... ., :t-:--: '.' : .- Visit our web site
for nore informotion:
OUtcl' ici'lSq: ir-rte -:r:-l- .--:--. www.Po*etScooe.com

while the inner lens cL,ns-.re.d ,-.:

Erploing llew l$orlds From the fulm of Your Hond"
three lenses glued together each oi
which had a diiferent Abbe numl-,er PocketScope.com LLC, 1 246-A Old Alpharetta Road, Alpharetta, GA 30005
To I -f ree 87 7 .7 1 8.63 57 . 7 7 0.7 7 2.63 57 . F ax 7 7 0.663.47 26
I
In designing optical systems Abbe
www.PocketS(ope.com
always started from a theoretrcai (Leeuwenhoek would be prcud!) Potent Pending
analysis. It was theory that led him
Circle No. 4 on Reader Service Card

0l,lilIUr/r.0 0rilr{ 0 8Ac r( 4$

 xcr0$$$GlEIlc0 by David R. Ma(in
gl, ll
.2 I 2 .J 1 5 7 8 9 t0 t2 13

r
c5
14 l5 16

11 18 19

20 2t 22 23 21

25 26 7

28 29 30 31 32

34 35 36 31 38

39 40 4l 12

43 44 45 16

47 18 +9

50 51 52 -53 54

55 56 57 58 59 50

51 62 63 54 55

66 6l 58

AGRO$$ 7l
69 10
1 Pronto: abbr. 54 Civet rel:rtive
5 Quartz deposit in a 56 Gibbs encrgy
furnace 45 Small amount A{zelius i1750-1837) 40 Samuel's teacher .) / GrcE-k ch(csu
9 "March King" 46 Pelvic bones 4 Orbital extremum 41 Biological duct 58 1977 Cher.n.
14 Point of minimum 47 5 Alpha {ollowers 44 Aged Nobelist
disturbance indicator)
-paper
(pH 6 Ol{actory sensation 46 CH3(CH2)rCH, and Prigogine -
15 Adams or McClurg 49 Mitosis star 7 Silicon monoxide CH"CH(CH"),, e.g. 59 Esker
16 Type of circle or 50 Chemical connec- B Element 2 48 Strong nuclear 60 A logic circurt
tube t10n 9 Trigonometric Particles?
(r2 Friend: Fr.
17 Shebat {ollower 53 Wave guide mode: {unction 49 [t's mostly nitrogen (r4 Variablc resistor
18 Hammer or wrench abbr 10 Yoko 50 Large raft (s1ang)

19 Like Neon 54 Narrowcoastalinlet 11 Release

- 51 Reproductiveorgan SALUTION IN THE
20 Tlpe of flow 55 Actress Gardner 12 Physician Hans 52 Zenith's opposite /VFXI /SSUE

22 Eisenhower's 55 Ferrous sulfide Hugo Bruno

SOLUTIAN TO THE MAY/JUNE PUZZLE
nickname 57 Half-spin particle 11907-1982) -
24 Author Rand 61 Plane detector 13 Competition site
F L B S E D M E
25 State of -matter 53 Unlock 21 Sci. org.
I, x o N
M I N R E B E E
26 Status _ 55 Famous lion 23 Biochemist
R I N D R E T E F (_) L D
27 Width times length 55 Sootydirt -
Bloch (1885-1944)
D R Y L C L N E S
67 Musical sound
C)
28 Cooks 25 This :oit'agazir'e
L
31 1946 chemistry 68 Baseballer Nolan 27 Appendage D E C I M S T Y R

Nobelist James 28 Nail C) T C H R o N{ I U T4

{ 1887-1e55) - 59 Eagle nest 29 Phosphatase unit K o N S S C () o T R

33 _ Descartes 70 Volt or coulomb 30 Sack or weed C E R S S K R N M E N

34 Southern constella- preceder precedcr N B I C I N E M B Y

tion 71 Redstone 31 S I L I C T E P R o
35 Type o{ engine
39 Abscisic acid: abbr.
40 Nickel-chromium
Dotllltll
I sphincter wing
:i -Lanka
ffffil':?,i,"- U N
M
I
o
N
N
D
S

F
D

o
E

R
G

S
R E

C
E

S
S

I
L E D I N E R T C o K E
alloy -{certain muscle) 36 Bump
N E E U N T I L S L G
42 Mine output 2 Sodium carbonate 37 Buffalo's lake
R E S M E S S C I R E
43 Electric doublet 3 Swedish botanist 38 Fissure
-
5ll JUIY/AUGUsI zooo
 ANSWERS,
H INTS &
SOLUTIONS

transformation, the line AC goes to

M29B the perpendicular line PN. Thus BN
tlllAIh For definiteness/ consider the is perpendicular to AC.Bttt BTLAC"
case when angle ABC is obtuse (see by construction, so BT ll PN, and BT
M296 figure I ). (The case when this angle passes through the midpoint of side
It follor,r,s from the grr-en equation is acute can be treated similarly.) Let PM rn triangle PMN. A theorem of
that u = s,r - 1. Thereiore, 3o.1 - 4a" = P be the point symmetric to M about elementary geometry says that a
3sr -3u- o.t - 1 = ,1- s'1. Thus, the point B. We first prove that triangles line parallel to one side of a triangle
cube root rs equal ro ,1 - al. Let's ABC andBPN are similar. Note that and passing through the midpoint of
transforrn the erpre ssron under the ZABC = ZPBN, since they are ob- a second side must pass through the
second root. \A'c lt,rr . 'ul - )a - 2 = tained by adfing a right angle to the midpoint of the third side as well. It
u2 + (u2+ u) + 21o. - 1 = sr - cr(u+ 1) angle CBN. Triangles CBM and follows that BT passes through the
+ Zlu + 1) = crr + st - rc{' = 1ul + cr)2 = ABN are similar right triangles. midpoint of MN.
ol(u + 1)2 = ss. The second term in Therefore,
the sum is cx3 = 1 - cr. Thus, the given M299
expression equals L BP BM BC Suppose that we have already
BN BN AB. constructed the desired circle C (see
M297 figure 2a). We draw a third circle,
Let us ask horr nuch oi each old Thus triangles BPN and ABC are centered at C, with a radius eclual to
piece could har-e gone rnto the mak- similar. the sum of the radii of the other two
ing of the larger nerr' piece. If all of Triangle BPN can be obtained circles. This new circle will then
the old 2 kg piece r, e re used, then from triangle ABC by rctating ABC pass through the center of the given
the larger lte\\ lt(Ls rluSt Contain by an angie of 90" followed by a di- circle, and is inscribed in an angle
0.5 kg of the 1 kg prece. Ii all of the lation centered at B. Under this (A, in the figure) with its parallel to
o1d 1 kg piece rrere used, then 1.5 kg
P
of the o1d 2-kg pre ce rras used. In ei-
ther case, the nerr 1.5 kg piece con-
tains at least 0.5 kg oi each oldpiece.
The reader is rnr-rted now to show
that at least one oi the original
pieces conrains no more than 40%
copper.
ArrnuJ rr ith rhcsc trto proposi-
tions, rre can no\\- ask: u-hat rs the
traritnutn potsrbi. pcrcenrage oi ,'lA
copper in the 2.5 kg piecei We ob-
tain it if we melt 0.5 kg containing
40ok copper with 2 kg of 100% cop-
per. That is, the maximum possible
value is

Oi.O-1 r)
).5

This is exactly the percentage of

copper given in the conditron of the
problem for the 2.5-kg piece. Thus
b
the original pieces were 40% and
1009/" copper, respectively. Figure 1 Figure 2

0 lJA llrTll lil/illr8 l1l r B s, 1lilllTs & s0tlJTr0llJs 51

the given angie at a distance equal to tangency.) Note that triangle O tO zC minimai sphere is r, then our lemma
the radius of the original circle. is a right triangle, COr: AB, OrO2 shows that MR = 2'rE , and we know
If we can construct this third : x + y, and that CO, : lx - yl.Then, that OM = l, KO = r. But OM: MK
circle, we can then shrink its radius by the Pythagorean theorem, + KO, so | = r + 2^l r .Fromthis equa-
by a known amount to find a circle tion, we find that r = 3 - 2",12.
which solves our original problem.
co.= @d-cd Now let us turn to the given prob-
So we have reduced our problem to
)l lem by re-introducing the middle
(r*y)'-("-y) =2rxy. sphere. It's clear that the radius of
that of constructing a circle (the new
one) tangent to a given angle (ang1e. (Note that the result for this the smallest sphere cannot be any
A1l, and passing through a given lemma is realiy a property of tangent less than the value for r found above.
point inside the angle (the center of circles, not tangent spheres.) We will show that we can in fact
the original circle). Before proceeding to the stated introduce a middle sphere such that
This problem can be solved by problem, we solve a simplified ver- the radius of the smaller sphere is
similarity, as shown in figure 2b. sion. Suppose we remove the mid- exactly r.
(1) First, we inscribe an arbitrary sized sphere (in the original problem If such a middle sphere exists,
circle ro centered at Q in the given statement). What then is the mini- then it must touch plane I at some
angle. mum possible radius for a sphere point P. We will find a point P and a
(2) We find the points where the tangent to plane I, line m, and the radius R for this middle sphere,
line A'O intersects the circle ro; call unit sphere touching plane L and which makes it touch the other two
them I and N. line m? spheres ifi our auxiliary problem.
(3) Draw the lines parallel to IQ In figure 3b, we again see a cross Figure 3c shows the situation on
and NQ through point O and find perpendicular to plane I through the plane I. Points M, K, and O are as in
the intersections of this line with centers of the two given spheres (and figure 3b, and point P is the point
the bisector A'Q of angle A' . These through line m). It is clear from this where our new sphere is tangent to
points are the centers of the desired figure that the center of the mini- plane I. If there is such a point P,
circles. mum sphere lies on the plane of this then it is not hard to show that OP
The original problem has, in gen- cross-section. If the radius of this = R, and our lemma tells us that MP
eral, four solutions. We have shown : zJR, and KP : 2"[k. As before,
how to get the two new circles that OM = L and OK: r.
are tangent externally to the given We apply the law of cosines, 1et-
circle. The construction of the ting cos IPOK: i.. In trtarrgle POK,
circles that are tangent internally is we have 4Rr : R2 + t2 - Z)"Rr.In tri-
left to the reader, as is the analysis angle POM, we have 4R : R2 + 1 -
of the special cases which result 2)"R. Multiplying the second equa-
when the given circle is itself tan- tion by r, and subtracting it from the
gent to one or both sides of the given first equation, we find that R : ^,lr .
angle. Now we can construct our
middle sphere. It is not hard to see
M300 that if we assign the line segments
We first obtain a simple and use- lengths according to the computa-
ful formula. tions above that MO < MP + OP.
Lemma: Suppose two spheres, Thus we can construct triangle
with radii x andy, are tangent exter- MOP to find point P, arrd the radius
nally, and each is also tangent to R wil1be the correct value so that a
some plane, at points A and B, re- sphere tangent to plane L at P with
spectively. Then, radius R will be tangent to the other
two spheres. Since R < I and greater
=2tW
AB .
thanr :3 - 2"{, , this lastValue for z
Proof: Let the centers of the is the smallest possible.
spheres be O, and Or. Figure 3a
shows a cross section of the situa- M
tion, taken through the plane deter- Physics
minedby the 01, 02, A, andB. (The
reader can prove that these four
points are in fact coplanar, and is P296
invited to recall that the line con- Let's displace the arc through a
necting the centers of two tangent C very sma11 angle Q. The restoring
circles passes through their point of Figure 3 torque of the force of gravity relative

52 JUI.Y/AllGlJSI 2OOO
to the pivot point is'determined by Plugging the formulas for V and P After the outer foil of the plate
the "surplus" of mass m on one side into this equation, we obtain the with charge 2Q is disconnected, the
and the " deLicit" on the other: thickness of the ozonelayer: charge of the outer plate remains on
it, and we carry this charge onto the
2mgRsirru = rf{oO]rRsin o,
-
h=
*"-
o"RT
other side of the capacltor. Now the
- \L )- gM
=1.7'10-3 m.
charges of the plates of the modified
where 2o, is the angle between the capacitor become 2.5Q and 0.5Q.
strings and M is the mass of wire arc. P298 The field between the plates
The moment of inertia of the sys- For the given charges, the electric changes direction (which is not es-
tem relative to the pivot point can field outside the capacitor is not zero sential for the energy calculation)
readily be calculated since all the (in contrast to the case when the and increases two-fold. Therefore,
parts having mass are located the total charge on the plates is zero). the energy of the field located be-
same distance R from the reference Any rearrangement of the plates tween the plates increases by a fac)
point: modifies only the internal field in tor of four and becomes
the capacitor; it doesn't disturb the
I = MR2. oz
external field. The outer foils of both W1 --
The next step is to write down plates (figure 4) collect equal charges
'2C .

Newton's second law for rotational of the same polarity, each oi which The outer field doesn't change, so
motion: is half the net charge of the capaci- our work was expended on increas-
tor (in a " coruectly" charged capaci- ing the internal field between the
MR)q,,= _2MRzEsinu ,, tor this half-sum is zero). In our case plates. Thus the work necessary for
the half-sum is 1.5Q. the charge transfer is
from which we obtain the period of Therefore, the inner foils of the
oscillation: piates cary charges -0.5Q and er er rez
L +0.5Q. The internal field is deter- )C 8C 8C
T =2n mined only by these residual
k'*" " charges, because in this region the P299
where the angle cx is fields generated by the external According to the statement of the
L charges cancel. The energy of the problem, the lens is placed in a way
A.=-. field located between the plates can that simplifies our calculations: the
2R
be calculated as usual: parallel beam hits the ilat srde of the
When the arc is small, we can set lens perpendicular to its suriace and
sin cr = o" and obtaln the usual ex- t^/ @lz)' e2
,2C8C
-:----:---!-_
doesn't re{ract. Therefore, ll e should
pression for the period of a math- consider reiraction onlr' at the
-
cmzrtical pcndulum as expccted. spherical boundart' benteen thc
This problem has an elcgant solu- 1.5Q 1.sQ 1.5Q -o.sQ glass and air. To begrn, rre iind the
tion that is based only on energy thickness d oi the lens aiong the
conservation ancl doesn't need the optrc aris irr he re it's thickestl:
value of the moment of inertia.
F

lil^-
Hrnt: compare the maximum values
of thc potential and l<inetic energies
+
F 1r/ -rn-d1 '

and recal1 a similar relationship for

harmonic oscillations. iritrl rrhrch \re get
P297
Ler ::: L-re rlass oi the Venu-
tl-re
S,:l1l .i:Ill,'S'a l.l'.' \l = J: I rlOlC the
fu d:0.67 cn.
Tl-rc thickness of the lens is impor-
ranr Lrecause we'll measure the dis-
Figure 4 rances frorn various points on the
':- K :-.- -:-, --:::: r.::-: sr-rriace of the lens. Thus a narrow
'..ri--- \ \ tr'-r.. -^. I l-: -_-.:I ldiaphragm-restricted) pencil of light
cupic. a volunrc \'- ----" :- .r. : p:.-- parallel to the optic axis is {ocused
sure P and temperature f. Accerr.Lins at a distzrnce
to the statement of the problem, P:
ntgf 1nr2, where r is the radius oi the R =locm.
r= n-l
planet. On the other hand, the equa-
tion ot \tatc for ozone is Now let's consider the ray far-
thest {rom the optic axis (figure 5).
PV =air
M
RT.
The angle of incidence for this ray

0lJArlrrlJlll/AlrslilrrR$, ilrlllT$ & s0tlJTt0ils 53

measured relative to the radius
drawn to the point of refraction at
the spherical surface is u:30o, since
sin u : (Dl2)l.R = 0.5. The angle of
refraction can be obtained from
Snell's law: sin p = n sin a:0.75,
from which we get B :48.6'. Simple
calculations yield the point on the
principal axis crossed by the ray af-
ter refraction. It's located at a dis- Figure 6 Figure B
tance I : (D 12) cotan (p - u) from the
flat side of the lens. Taking into ac-
count the thickness of the lens, we 8299
find that the outermost rays of the See the figure B.

beam intersect 3.2 cm from the

screen, and so the diameter of the 8300
light spot is about 2.2 cm. When the thermometer was
It would be interesting to investi- taken out of the molten tin, the
gate the question: Are there rays Figure 7 temperature of the glass bulb
that produce a spot with a larger di- dropped sharply. The bulb shrank
ameter than that produced by the point, in which case the lines of in- and squeezed the mercury, causing
outermost rays considered here? tersection of any two of them go it to rise in the glass column. To
through this point, or (2) they do not quantify this effect, compare the
P300 intersect in a point, in which case coefficients of thermal expansion
Since momentum and energy of the lines of intersection of some pair for glass and mercury.
the system before nuclear fusion are of them are parallel. But the'lines of
zero, the newly formed particles fly intersection' of the 'planes' of the
off in opposite directions with nu- three quadrilaterals shown do not
merically identical momenta (this intersect in a point, and no two are
follows from the 1aw of conservation parallel (figure 6).
of momentum): Challenge: What could the
teacher's pentahedron have looked Cml'ections
Pn=t\m,Er=Po=$4"h. line?
Vol. 10, No. 4
Energy conservation requires 8297 p. 7, col. 2: Third line after the
E=En+Eo.. This situation is possible (see {ig- second display equation:
ure 7).
Solving this system of ecluations si- for "xr"read "\r" .
multaneously, we get 8298
Suppose there were more chem- Vol. 10, No.5
nfrrr t
E, =-md+mn = 3.5 MeV, ists than alchemists. Since the to- p. 36, col. 2: The formulas should
tal number of participants was 100 be:
(an even number), the number of
.)
nffior _c chemists was not less than 51; then I b\ 4ac-b)
r:.
" = mo+mn = 14.1 MeV.
there were not more than 49 alche- x=alx+ +_
za) 4a
mists. Thus there was at least one
chemist among those who an- +ac-u2\
swered the cluestion, and she must
(
_u
have said that chemists were more lro' 4a )
BnainteasEl'$ numerous. Or suppose there were
more alchemists than chemists.
8296 Then there was at least one alche- Have a question about your
The drawing is incorrect. The mist among those who answered ibscription? Contact Springer-
New York:
quadrilaterals shown would not lie the question, and he must have said
1 .800.$PHINGER (777 .4643)
in a plane. that chemists were more numer- 201.348.4033
To see this, note that we can say ous. Therefore, 50 chemists and 50 Fax 212.533.3503
the following about any three alchemists attended the confer- [email protected]
planes: either (1) they intersect in a ence.

5 4 JtlIY/AUGllsr 2ooo
 INFORMATICS

Ihul'E$
by Don Piele

ACH SPRING, THE USA COMPUTING OLYM- Farmer fohn's list is nicely ordered, and chore K (K > 1)
piad staff is faced with the iob of creating a chal- can have only chores l, ..., K - 1 on its dependency list.
lenging set of informatics problems for our national Write a program that reads both a list of chores from I
competition. These problems are similar to the pro- through Nwith associated times and a list of chore pre-
gramming challenges presented at the International recluisites. Calculate the shortest time it will take to
Olympiad in Informatics. They focus on tasks that can complete all N chores. Of course, chores that don't de-
be solved with an efficient algorithm that will quickly pend on each other can be performed simultaneously in
dispose of the 10 test data sets. An optimal program parallel. In fact, a large number of chores could be tak-
must find a solution for each data set within a few sec- ing place simultaneously.
onds to receive the maximum score. Less efficient pro- INPUT FORMAT:
grams receive partial credit, depending on how many Line l: One integer/ A{, the number of chores (3 < N
test cases it can solve within the time 1imit. The scor- < 10,000)
ing is done with an automated grading system. Line 2, ..., N * 1 : N lines, each with several integers:
Creating the problems is anything but automated. The chore number (1, ..., N, supplied in order in the
The entire staff participates by submitting new problem input file).
ideas to the head coach-Rob Kolstad. A list of approxi- The length of the chore in minutes (1 < length < 100).
mately ten possible problems are considered, and five A list of no more than 100 prerequisite chores, if any
problems are selected for the competition. Then the fun are needed.
begins: every problem gets "cowified." SAMPLE INPUT (file CHORES.IN): 7
Cowification is the process of transforming an ordi- 15
nary looking programming task into a barnyard chore. 2tt
It is also a license to work in some of our " dairy state" 332
humor. Presented in this column is one of the easiest 46t
tasks from our recent US Open Competition. 5t24
6824
The Rl'ollem 74356
Farmer |ohn's famlly pitches in with the chores dur- OUTPUT FORMAT:A single line with a single integer
ing milking time. They round up the cows, put them in that is the least amount of time required to perform all
the stalls, wash the cows' udders, and perform many the chores. SAMPLE OUTPUT (file CHORES.OUT): 23.
other tasks. Organizing the chores and completing them The sample INPUT file is represented graphically in
as quickly as possible is always desirable, because it figure 1. The chores are numbered 1 to 7 and highlighted
leaves more time for hang gliding with the cows. Of 13

q \>@€#
course/ some chores cannot be started until others have
been compieted. For instance, it is impossible to wash
a sow's udder until a cow is in the stall, and you ,,ffi--.--@------
wouldn't want to attach the milking machine until you u/ \ \ l\
have washed the cow's udder. Farmer |ohn has created
list of N chores that must be completed. Each chore re-
quires an integer number of minutes to complete, and
\ o ..Ys --'
there might be other chores to be completed before this
chore can be done (i.e., prerequisites). At least one chore
has no prerequisite: the very first one, numbered 1. Figure 1 ffi1ffi-
:

0 UA llIU ll/l/l il F0 B lil AT I I S 55

in green. The time required for the chore appears above We now have the minimum length of time it would
the green dot. Lines between chores denote a depen- take to complete each chore and all dependent chores
dency. Lower numbered chores must be completed be- for all n chores. The solution to the problem is simply
fore the highe.r numbered chores can be done. to take the maximum of a1l these values, which in this
It is clear from figure 1 and7,
that chores I,4,5, case is 23.
which follow a line of dependency, will take 5 + 6+B
+ 4 or 23 minutes to complete. This is the worst case Pseudo mde
and hence the shortest time that it will take to do all Here is the pseudo code which encapsulated the al-
chores. gorithm described above.
1) Initialize
Ihe solution Chore[k] : Time to complete chore k.
What we are looking for is an algorithm that guaran- DependentChores[k] = List of chores that must pre-
tees a solution without summing up all chore paths. To cede chore k.
sum up all chore paths would be, in the worst case, of 2) Recursive algorithm for computing MinTime[k] :
order (n - 1)l for n chores if chore i depended on nearly Minimum time needed to complete chore k and all de-
all previous chores for i :1 to n. A more efficient solu- pendent chores.
tion uses recursion and works as follows. If we have only Fork=1ton
one chore to complete, the solution is known: the time MinTime[k] : Chore[k] + MaxfMinTimefl], 7 e
it takes to complete chore one. Now assume we know the DependentChores[k]l
minimum time it takes to complete chore k - I and all 3) Solution: Max[MinTime[k], for k = 1 to nl
dependent chores. The shortest time that chore k can be
completed is eclual to that maximum of all shortest times Mafienatha nle
each dependent chore can be completed plus the time it Now for the actual code in Mathematica which mir-
takes to complete chore k. These times are indicated rors the pseudo code:
below in blue as they are generated for each chore from
(* INPUT file *)
k : 1 to 7. Each frame in figure 2 shows one step in the
process of finding the minimum time it would take to
CHORES = ({1, 5}, t2, L, tl, {3, 3, 21,
complete both a chore and all its dependent chores. 14,5,1), {5, L,2,41, t6,8,2,41,
t7, 4, 3, 5, 5));
I 3
(* n = nurnber of chores *)
2 F.-s-- n = LengthlCHORESL
5/ \ \ l--.- 4 5 ///
(* pick off chore times *)
1

\o \ >@;€ I
Chorelk_l := CHORES[[k, 211
sezH---'- (* pick off dependent chores *)
DeBend.entChores [k-] := CHORES[ [k,
6 rffi ffi* Range[3, LenghtICHORESt tk] I I I I I
2 F--s-- $=*-- (* comlrute MinTime to complete each
'#,
/ \\b-- 4 ffi.l \\)---
ffi.
4
chore and all dePendent choreE *)
\# \Xs)sr€
,.i1:
1 ,-W-- 7
MinTime[1] = chore[1];
\o :,t
$&--- --'
e--s- AllMinTimes = TablelMinTimelkl =
Chorelkl + MaxlMinTime /G
6, & DependentChoreslkll, {k, 2, n}l
?:
-6
F<ffi--._ {6, 9 , Ll, 1,2, 19 , 23}
J
/// \ \ e)--.- 4
1
\.Ysiffi7ffi
\ws'Fa&- -.''
tvlax [A1U{inTimes]
ZJ

6t r@ lhlll' ilrll
2i F.-s.- Certain sequences of chores follow apath of depen-
dency and take the full time needed to complete all
i' \>%>e chores. These sequences are called critical paths. There
\gt2#--- can be more than one critical path. Your chore is to
modify the code ever so slightly to find a critical path.
In our example, {1, 4, 6,7} is a critical path because chore
Figure 2 7 depends on chore 5 which depends on chore 4 which

5$ JULY/AUIUSI 2OOO
 Reid Barton, Arlington, MA; |ohn Danaher, Spring-
2 field, VA; Vladimir Novakovski, Springfield, VA; Percy
Liang, Phoenix, AZ;Yuran Lu, Presque Isle, ME; |acob
,,/ 4 Burnim, Silver Spring, MD; Steven Sivek, Burke, VA;
1 7
fack Lrndamood, Dallas, TX; George Lee, San Mateo,
CA; Gary Sivek, Burke, VA; Richard Eager, Falls
Church, VA, TomWidland, Aibuquerclue, NM; Gregory
\o Price, Fal1s Church, VA; Kevin Caffuey, Oakton, VA;
4 Thuc Vu, Anaheim, CA.
The USACO team will have the opportunity to repre-
Figure 3 sent the United States in the Central European Olympiad
depends on chore 1, and the total time to complete all in Informatics, August 24-3L, in Romania lhttp:ll
chores is 23: the minimum time needed to complete a1l ceoi.ubbcluj.ro) and the International Olympiad in
chores. Your output should be a list of chores. Figure 3 In{ormatics in Beijing ChinA (http://www.ioi2000.org.cn),
shows graphically a critical path. September 23-30,2000.
The complete listing for all participants in the 2000
2000 U$ 0pen US Open can be found on the USACO website at
The 2000 US Open, held in Aprii of this year, at- www.usaco.or8.
tracted 378 entries from 34 countries, including 223
from the United States. For US students this is the fi- Finally
nal competition of the year leading up to the selection Waiting for two months to see a solution is not nec-
of the fifteen finalists. The finalists will spend an all- Ail solutions to the prob-
essary, thanks to the Internet.
expenses-paid eight days at the University of Wiscon- lems presented in this column are available at the
sin-Parkside in fune competing for one of four spots on Informatics website:
the USA Computing Olympiad team. The finalists this http://www.uwp. edu/academic/mathematics/usaco/
yeaT aTe'. informatics/. O

W'orldco LLC
IVY LEAGUE, GRADS FROM TOP
lllalG$ a [8l'l8cl uilll
SCHOOLS
Wantedfor ltosition of Use t1-rerespclnse card in this issue to order
Equity Trader Qtttutttutt ior r,our chi1d, grandchild, rliece,
ncphcu-, mother, father, friend. . . Or call 1 800
WORLDCOLLC,COM
SPRNGER 1777-4643). Give them six colorfu1,
challengrng, entertaining issues of Quantum-
\florlclco LLC, an establishecl "se1f clearning" tracling
.t \-e.u's \\.orth of reading pleasurel
firm, seeks bright determined individuals rvith the
drive ar-rd cor-nmitment to beconre successful trade.s.
Factor x into the
Frr rliose uitl-r 3,8+ GPfs. e chance to rrade u'ith OUANTUM equation,
' :1r , ::l'r::1 l rrh tto cap.ital cor-rtribrtti6n oi r ortr where x is any potential.
OUANTUM reader you kriow!

lnder ulldmfisel,s
American Mathematical Society 26
Fasrc-.ttl:.-:-r:- :-."- - - Birkhluser Boston 27
ATTN: ll,tgrl,t 1,,.:. : Princeton University Press 43
mktczot'owskat tt ln.. -. .
PocketScope 49
\l-or1c1co LLC Covcr 3
 *-tEg@*TEGilE*&=*#**re* **"+;z=;Zt+-s*+r*--&:#?*#E
Think thermo- How do boats work?
dynamics is beyond Why do they float?
your students' grasp? Exptore principtes of
Engage student interest buoyancy, hut[ design,
in heat transfer and scaLe modeling, and
insutation with this seaworthiness. In this
votume. A chaLlenging, voLume, students
hands-on opportunity for investigate the physics
students to compare the of boat performance and
functjon and design of work with systems and
many types of handwear modeling. Through
and to design and test a research, design, testing,
gLove to their own and evaluation of a
specifi cations. Students modeL boat, students
Learn the basic principLes of product design white experience the practica[ application of mass, speed,
expLoring princip[es of physics and technotogy and acceteration while apptying the math and science
necessary to construct necessary to bujld a
and test an insutated scaLe model of a boat.
gtove. #PB152X1 #PB7s2X2
Science by Design Series ,

Ihisseries offers a hands-on approach for students to

successfutty develop and carry out product design. Each
book includes teaeher-tested units that introduce
students to the desiqn process and sharpen students'
abititjes to investigate, build, test, and evaluate famjliar
'-.-;-. i 1;,.':-=. :.-..;-.i"' .:-;=.= products. Al[ four volumes are keyed to the National t:=.a+-,1::-;'.:-"=a.=:===-.4=-:--

: ::= i i.:
;.
=.-.::
-:1:1::: Science Educatian Standards, the Benchnarks far Science : 1=='-::=':==:::'::=
=
How==
can physics Literacy; and the Internatianal Technalagy Education How much can you
hetp your garden Standards. Produced in cooperat'ion with IERC. {Grades launch? How far
grow? 9-12, 2000) witt it go?
Engage your students into
Each votume in series $19.9s Cataputt physics
in a probLem-solving and technology with
chaLtenge to design and
Set of alt four volumes $63.95 SAVE 2o%!
the heavy weaponry of
buitd a physicaI svstem the MiddLe Ages. This
that provides an optimal environment for plant votume integrates history, physics, mathematics,
growth. This volume helps you cultivate student and technology in its
interest in optics, energy challenge to students to
transfer, and photosynthe- design and buitd a
sis. In addition to learning working cataputt system.
and apptying concepts in Students investigate
thermodynamics, tight e[asticity, projecti[e
absorption, and p[ant [aunching, and [earn
bioLogy, students must about frequency distri-
make a range of decisions bution while working
as they encounter cost through the process of
constrai nts, construction product design.
alternatives, and environ- #P8152X4
mental changes while
buiLding a greenhouse
mode[. #P8152X3