Arch. Hist. Exact Sci. (2011) 65:6797 DOI 10.

1007/s00407-010-0071-0

The mathematics in the structures of Stonehenge

Received: 11 February 2010 / Published online: 4 November 2010 Springer-Verlag 2010

Abstract The development of ancient civilizations and their achievements in sciences such as mathematics and astronomy are well researched for script-using civilizations. On the basis of oral tradition and mnemonic artifacts illiterate ancient civilizations were able to attain an adequate level of knowledge. The Neolithic and Bronze Age earthworks and circles are such mnemonic artifacts. Explanatory models are given for the shape of the stone formations and the ditch of Stonehenge reecting the circular and specic non-circular shapes of these structures. The basic mathematical concepts are Pythagorean triangles, thus adopting the construction procedures of the Neolithic circular ditches of Central Europe in the fth Millennium bc and later earthworks and stone circles in Britain and Brittany. This knowledge was extended with new elliptical concepts. Approximations for the values of and the square root of 2 are encoded in the henge. All constructions were performed using a standardized Babylonian metrology that shows a remarkable consistency and comprehensible development over some 14 centuries. 1 Introduction In ancient script-using civilizations, the achievements in mathematics and astronomy have been investigated to an adequate degree (e.g., Neugebauer 1969, 1975; van der Waerden 1954, 1973, 1983). In illiterate ancient European civilizations, there is

Communicated by Menso Folkerts. Electronic supplementary material The online version of this article (doi:10.1007/s00407-010-0071-0) contains supplementary material, which is available to authorized users. A. Kainzinger (B ) Elektrastr. 46a, 81925 Munich, Germany e-mail: [email protected]

123

68

A. Kainzinger

some evidence of basic astronomical knowledge due to prehistoric monuments and earthworks (Thom and Thom 1980), but there are almost no secured ndings about the level of mathematical knowledge. Because there are many identical solutions for mathematical problems in the ancient civilizations of Mesopotamia, Egypt, India, China and Greece, van der Waerden assumed a common origin for the basic mathematical concepts, which he placed in Neolithic Europe (van der Waerden 1983). Two examples should highlight the basis of this supposition: (a) in all the abovementioned civilizations, the Pythagorean Theorem was used, and a common set of Pythagorean Triangles was known and/or calculated in the same way; (b) in India and in Greece, comparable ritual geometric constructions were performed (generally for the shapes of altars). This assumption was supported by earlier articles of Seidenberg (Seidenberg 1961, 1978, 1981). A gure summarizing these ndings of van der Waerden is included as Electronic Supplementary Material (ESM Fig. 1). In illiterate civilizations, knowledge is generally transferred by oral tradition to subsequent generations. Mathews proposed an ancient core of a Neolithic oral tradition of mathematics (Mathews 1985). In addition, the knowledge can be documented by the depiction of illustrating pictures or symbols on mnemonic objects (e.g., Vansina 1985; Haarmann 1991) or by the construction of appropriate devices and monuments. The majority of the Neolithic and Bronze Age earthworks and stone/wood rings are object of this mnemonic procedure. The encoding of the mathematical concepts in these monuments was performed by the appropriate construction of ditches and rings which are composed by circles and circular arcs. With their centers and endpoints, these elements determine the geometrical models. This procedure is distinct from the documentation of geometric problem solutions nowadays. With the present method, linear geometric models are depicted directly by their line segments, while in the ancient monuments, the linear geometric models are represented exclusively by the respective vertices. The principal objective of the investigation of ancient earthwork and stone/wood rings now is to determine the original construction concept. So much effort in the construction of these outstanding monuments could not have been expended without a comprehensible and consistent plan. A determination procedure comprising six steps was established to work out highly probable concepts of the ancient plans (see Sect. 8, Fig. 13). Two steps have to be explained at this point to familiarize the reader with the often-used terms backward construction and forward construction. A backward construction is the deduction of a potential construction concept from the available data of the plan of the ancient monuments present remains. There might be several attempts and different results in this step for the search of the original concept. The six-step determination procedure aims to reduce these options to preferably one plausible plan. The forward construction is the exact construction based on the selected mathematical concept and the established metrology. This article is concerned exclusively with the construction concepts of the Stonehenge structures. Some efforts have been there to uncover the principles of the Stonehenge constructions, e.g., the early attempt by the use of computers in the early 1960s (Hawkins 1965). However, this statistical approach and other methods produced no satisfactory results. A new, sound approach was published by Johnson

123

The mathematics in the structures of Stonehenge

69

(2008), although the author cannot agree with the proposed mathematical explanation models (Johnson 2008). 2 Z and Y holes We begin with the last phases of Stonehenge. Although the construction of the Z and Y holes document in some way a decline of the mathematical knowledge of the Stonehenge society, the construction concepts of these ring-like structures reect the beginning in an especially descriptive manner. The Y and Z holes obviously do not form an exact circle. We encounter this characteristic in nearly all early rondels (circular ditches) and wood and stone circles in Central Europe and Britain. In the case of the exact circles in Stonehenge (Aubrey holes, Sarsen circle), we will show a clear reasoning for these constructions. The basis of the construction of these ringlike structures is always a mathematical concept. In the majority the construction starts with a Pythagorean triangle. The knowledge of Pythagorean triangles traces back to the fth millennium bc (rondels/Kreisgrben in Central Europe). 2.1 Z holes The construction of the Z holes is an excellent example of this procedure. To exemplify the a.m. determination procedure, we will depict the results of the backward construction for the Z holes (Fig. 1a) and also the forward construction as well (Fig. 2). (For the remaining structures, a combination of the sampling points and the resulting forward construction is displayed, with the exception of the Bluestone horseshoe.) At rst, an isosceles triangle (Z 31 Z 32 Z 33 ) composed of two mirrored Pythagorean triangles of the shape (3,4,5) is constructed by sharing the leg 4. Then, another isosceles triangle, (Z 31 Z 33 Z 34 ), also composed of two mirrored Pythagorean triangles of the shape (3,4,5) is attached, but in this case sharing the leg 3. The outline of the overall geometry is another single Pythagorean triangle (Z 32 Z 33 Z 34 ) of the shape (3,4,5) (Fig. 1c). From a mathematical point of view, the basis for a Pythagorean triangle is a Pythagorean triple of three integers a, b, and c fullling the Pythagorean theorem a 2 + b2 = c2 . If these integers have no common factor, then they are called primitive Pythagorean triples.1 In the depiction of the construction concepts by Pythagorean triangles, the integers i of the corresponding primitive Pythagorean triples are denoted by squared brackets: [i]. The position of the vertices Z31Z34 (which are the centers of the respective circular arcs) are calculated by a nonlinear regression analysis (see Sect. 8); the sampling points for the regression analysis are depicted in Fig. 1a by a small cross symbol. The construction was performed on the basis of a consistent metrology (see also Sect. 7). The concrete length of the legs of the respective Pythagorean triangles are 6, 8, and 10 cubits; the resulting overall shape is a Pythagorean triangle with legs of 12, 16, and 20 cubits. The position of vertex Z31 is near the generic center of the Stonehenge structures.
1 For example, (6,8,10) is a Pythagorean triple; (3,4,5) is the corresponding primitive Pythagorean triple.

123

70

A. Kainzinger

Z34

Z31 Z33 Z32

10 m

20 m

C
Z34

z1

z3
z2

Z31

[3] 6c

Z35

z1

[5 10 ] c

ro

d
[3] 6c

Z36

[3] 6c [3] 12 c

Z33

Z32

Fig. 1 Z holes. a The points Z31 Z34 are the center points of the respective circular arcs provided by circular regression analysis. b The arbitrary radius z1 is 36 cubits long which equals 6 reeds or 3 rods. c The triangles Z 31 Z 32 Z 33 and Z 31 Z 33 Z 34 are isosceles triangles composed of two mirrored Pythagorean triangles of the shape (3,4,5). The basic Pythagorean triangles of the shape (3,4,5) have concrete legs of 6, 8, and 10 cubits

123

The mathematics in the structures of Stonehenge

71

Z34

Z31 Z33 Z32

Fig. 2 Forward construction of the Z holes. The isosceles triangles Z 31 Z 32 Z 33 and Z 31 Z 33 Z 34 and the circular arc with center Z31 are computed by a mathematical exact construction. The resulting curve for the Z holes is totally smooth at the stones Z17, Z23, and Z30 compared to the one achieved by the backward construction

The vertices Z31 Z34 serve as centers of circular arcs on which the stones of the Z holes were set (Fig. 1b). The four circular arcs are composed in a reasonably smoothed curve starting with hole/stone Z7 and ending with hole/stone Z9 anticlockwise. The length of the radius of the rst circular arc with center Z31 was arbitrary and was set 36 cubits which equals six reeds or three rods. The lengths of the radii of the remaining circular arcs result from the smoothing procedure; the ring so dened was, therefore, not smoothly closed between the holes Z7 and Z9. In the backward construction (Fig. 1b), we still see minor gaps at the connection points Z30, Z23, and Z17. In the exact forward construction, the position of the starting vertex Z31 was set very slightly lower. This construction results in a totally smooth curve (in terms of mathematical continuity) at all connection points as is shown in Fig. 2. The lengths of the circular arcs could be determined by either the angle of the arc or the heading of the two legs. This procedure for dening the construction of the Z holes provides a blue print for the construction of all the ring-like structures of Stonehenge.

123

72

A. Kainzinger

2.2 Y holes As we will show later, the construction concept of the Z holes is somewhat simpler than that of the preceding phases 3ii and 3iv of Stonehenge. The Y holes show a further decline of the construction complexity. The base of the construction is only one Pythagorean triangle (Y31 Y32 Y33 ) of the type (5,12,13); this is the simplest construction concept of all the ring-like structures of Stonehenge (Fig. 3b). The concrete lengths of the legs are 10, 24, and 26 cubits, thus starting the construction with a Pythagorean triangle of the shape (10,24,26). Once again the vertices of the triangle serve as the centers of the three circular arcs on which the stones were set. Vertex Y31 again is very near to the general center of Stonehenge. The length of the radius y1 of the dominant circular arc with center Y31 seems to be set arbitrarily to 50 cubits which equals ve poles. The leg Y31 Y32 of the Pythagorean triangle and the radius Y32 Y 3 of the circular arc with center Y32 form a straight line; as the leg Y31 Y32 is 10 cubits, the length of this latter radius is exactly 40 cubits which equals four poles. The length of the radius with center Y33 results from the smoothing procedure at hole Y15. In the backward construction (Fig. 3a), these two arcs meet each other at hole Y15 with a small gap, while the arc with center Y32 misses the main arc in the segment between the holes Y2 and Y3. In general, the holes/stones of the Y ring were not set with the accuracy of the Z ring. The shapes of three holes are outside the exact circular arc (Y20, Y25, Y29), and others do not hit the arc centrically. For the position of the hole/stone Y8, we have no mathematical concept. It could be an arbitrary determination to ll the gap between the holes Y7 and Y9. This additionally reects the decline in the geometrical methods and accuracy for ring constructions in this last phase of the Stonehenge constructions. For the forward construction, the mean of the measures of the radii y1 and y2 were used; the radius y1 provides the cubit measure with the least tolerance. In this forward construction, the position of vertex Y33 by an exact construction of the Pythagorean triangle (10,24,26) becomes a bit higher compared with the backward construction. This forward construction now results in a smooth curve at hole Y15 and hole Y3 as well (Fig. 3c). 2.3 Deduced measures The above mentioned cubit measures are ideal values according to the selected mathematical conceptin these cases, the underlying Pythagorean triangles. The actual length values extracted from the plan are summarized in Table 1A. For the Z holes, the resulting ve cubit measures vary between 49.9003 and 51.4553 cm and have a mean of 50.7228 cm. The mean cubit measures of the Y holes are 52.3558 cm (backward construction) and 53.7403 cm (forward construction). The arbitrary radius z 1 of the circular arc with center Z31 with the supposed length of 36 cubits (= 6 reeds = 3 rods) results in a cubit measure of 50.4836 cm which is very close to the above established mean value for the Z holes. The radius y1 of the circular arc with center Y31 seems to be set arbitrarily to 50 cubits (= 10 poles) and results in a cubit measure of 53.9227 cm which is again very close to the above established mean value for the Y holes (forward construction).

123

The mathematics in the structures of Stonehenge

73

Y33

Y33 Y2 Y3 y3 Y31 Y32 y2 y1

10 c

Y32

Y31
Y7 Y15 Y9

Y33

Y21

Y19
Y31 Y32

Y7

10 m

20 m

Fig. 3 Y holes. a The points Y31 Y33 are the center points of the respective circular arcs provided by circular regression analysis (backward construction). The radius y1 is 50 cubits long which equals ve poles. b The triangle Y31 Y32 Y33 is a Pythagorean triangle of the shape (5,12,13) and has concrete legs of 10, 24, and 26 cubits. c Forward construction. The Pythagorean triangle Y31 Y32 Y33 and the circular arc with center Y31 are computed by a mathematical exact construction. The resulting curve for the Z holes is totally smooth at holes Y3 and Y15 compared to the one achieved by the backward construction

123

74

A. Kainzinger

Table 1 Measures of the Y and Z holes, the ditch, and the Aubrey holes. length gures are the measures of the respective segments in the computer program. length henge results from the scale in the gure. result. cubit measure is calculated as length henge divided by no. cubits Segment Length gure ( mm) Length henge (m) No. cubits Result. cubit measure (cm)

A: Y holes and Z holes measures Measure Z holes Z 31 Z 32 Z 31 Z 33 Z 32 Z 33 Z 31 Z 34 Z 33 Z 34 z1 z2 z3 z4 Y holes Y31 Y32 Y32 Y33 Y31 Y33 y1 y2 y3
d Y31 Y32

58.4667 15.0421 14.8960 17.6184 14.5884 23.8901 53.129 66.471 72.916 73.462 15.2718 36.9047 39.6949 78.8170 62.9270 114.7840 15.7101 30.9000 8.0347 7.6574 12.3383 90.528 79.408 83.412 83.624 68.9710

20.00000 5.14553 5.09555 6.02682 4.99033 8.17221 18.17411 22.73807 24.94275 25.12952 5.22410 12.62418 13.57864 26.96133 21.42313 39.26474 5.37403 20.00000 5.20045 4.95625 7.98595 58.59417 51.39676 53.98835 54.12557 44.64142 10 10 16 Mean: 120e 100 110 110 90f 52.0045 49.5625 49.9122 50,4931 48.8285 51.3968 49.0803 49.2051 49.6016 10 10 12 10 16 Mean: 36a 45 50 50 10 24 26 Mean: 50b 40c 75 10 51.4553 50.9555 50.2235 49.9033 51.0763 50.7228 50.4836 50.5290 49.8855 50.2590 52.2410 52.6007 52.2255 52.3558 53.9227 53.5578 52.3530 53.7403

B: ditch and Aubrey Holes measures Measure Ditch D1 D2 D1 D3 D2 D3 S1 S2 S3 S4 Aubrey Holes rA

123

The mathematics in the structures of Stonehenge Table 1 Continued Segment Length gure ( mm) Length henge (m) No. cubits

75

Result. cubit measure (cm)

C: Station stones, Heel stone measures 91-93 92-94 91-92 94-93 91-94 92-93 91-96 94-96 139.5024 139.0235 53.8535 52.2805 129.2255 128.2930 119.3471 120.3418 90.29282 89.98285 34.85663 33.83851 83.64110 83.03754 77.24731 77.89113 182 182 70 70 168 168 159 159 Mean:
a 36 cubits = 6 reeds = 3 rods b 50 cubits = 5 poles c 40 cubits = 4 poles d forward construction e 120 cubits = 20 reeds = 10 rods = 2 ropes f 90 cubits = 9 poles = 1 1/2 ropes

49.6114 49.4411 49.7952 48.3407 49.7864 49.4271 48.5832 48.9881 49.2467

3 Ditch and Aubrey Holes The ditch and the Aubrey Holes belong to Stonehenge phase 1 (ca. 2950 bc). 3.1 Ditch The ditch was excavated by Hawley in the early 1920s in the eastern and southern parts. As far as the original course and shape of the ditch is concerned, the excavation results have to be questioned. The excavated cross sections and the irregularities of both the width and the depth along the ditch ring are not in line with earlier and coeval rondels and ditches. In general, the bottom of the ditch should dene a comprehensible line or small band that reects the original construction concept. The eastern ditch terminal at the entrance causeway is a good example for this critical analysis. Johnson assumed that it is more than likely therefore that the workmen had mistakenly cut a little beyond the actual end of the ditch, slicing through the postholes on the edge of the causeway (Johnson 2008, p. 103). It seems that the workmen had overdone their task in the rest of the ditch also. This open question can be resolved by an adequate investigation and cautious excavation of the remaining part of the ditch including non-destructive methods. For the generation of a potential construction concept, we have to cope with an unsettled database as far as the excavated ditch is concerned. The situation for the remaining non-excavated part of the ditch is at the present time somewhat better but not ideal; here, we can assume the construction curve in the middle of the still evident depression of the ditch. However, this can be an indication only due to the unknown relling conditions. Despite this soft database, the backward construction provides a comprehensible mathematical construction concept (Fig. 4).

123

76

A. Kainzinger

For each of the two sections of the non-excavated part of the ditch, eight reading points were chosen. The reading points of the two sections of the excavated part are sampled on the basis of the sampling theorem at equidistant spaces of 5 as shown in Fig. 4c. This procedure provides 17 and 15 sampling points, respectively. The construction starts with the isosceles triangle D1 D2 D3 composed of two mirrored Pythagorean triangles of the shape (3,4,5) in this case sharing the leg 3 (Fig. 4b). This geometric gure (including the one sharing the leg 4) can be assessed as one with the highest appreciation and is therefore quite probable for the starting phase of the Stonehenge structures. These shapes of isosceles triangles were also the basis of many explanation models for the construction of stone rings in Britain and Brittany by Thom and Thom (1980). The construction of center D4 is not apparently due to the soft database for the respective circular arcs; this center could be identical either to the point D1 or the center A1 of the Aubrey Holes (see below). The lengths of the legs of the triangle D1 D2 D3 are 6, 8, and 10 cubits. The lengths of the radius s1 are the most reliable data for the deduction of a cubit measure. This length can be assumed 120 cubits which equals 20 reeds or 10 rods and results in a cubit measure of 48.8285 cm that ts to the overall measuring scheme (see Sect. 7). However,owing to the a.m. database situation, the determination procedure according Fig. 13 has to be assessed as non-successful. As both the construction concept and the deduced measuring unit t in the overall explanation model of the Stonehenge structures, the ditch results are worthwhile to be documented.

3.2 Aubrey Holes The 56 Aubrey Holes dene an exact circle (Fig. 4a, c). The circle was calculated by circular regression analysis involving all the 56 holes. The center of this circle A1 is inside the triangle D1 D2 D3 (Fig. 4b). The construction of the 56 holes/stones is a good example for the encoding of mathematical concepts in earthworks by illiterate societies; thereby an approximation of the circle constant is represented in the henge: radius of the Aubrey holes circle rA = 90 cubits = 9 poles (Table 1B) diameter dA = 180 cubits = 18 poles; circumference of the Aubrey holes circle cA = 56 poles; 28 56 poles cA = = 3.1111 . . . . = circle constant := dA 18 poles 9 The mathematical concept is based on a regular polygon of 56 sides inscribed in a circle. The respective circular area AC could be established by the method described by Johannes KEPLER (Wuing 2008, p. 439). This method also starts with an inscribed polygon composed of isosceles triangles providing the formula AC = circumference radius/2

123

The mathematics in the structures of Stonehenge

77

D3
10 c [5]

D4 D1

s3 s4 s1 D3 D4 D1 A1 D2 s2

A1
6c [3]

rA

D2

C
50

1 40

D3

D4 D1 D2

10 30

20

20 m

Fig. 4 Ditch and Aubrey Holes. a The points D1 D4 are the center points of the respective circular arcs of the ditch provided by circular regression analysis. The center of the Aubrey Holes A1 was calculated by circular regression analysis also. The radius s1 is 120 cubits long which equals 20 reeds, 10 rods, or 2 ropes. The radius of the Aubrey holes circle r A equals 90 cubits or nine poles. b The triangle D1 D2 D3 is an isosceles triangle composed of two mirrored Pythagorean triangles of the shape (3,4,5). The basic Pythagorean triangles (3,4,5) have concrete legs of 6, 8, and 10 cubits.c Forward construction of the Ditch and the Aubrey Holes. The isosceles triangle D1 D2 D3 and the circular arc with center D1 are computed by a mathematical exact construction

123

78

A. Kainzinger

This formula is equivalent to AC = r 2 but does not need explicitly. We get AC = 92 28/9 = 252 pole2 = 36 7 pole2 . We have no knowledge of the algebraic achievements of the Stonehenge people. An approximation of 7 could, however, be established by the formula a 2 b a b/2a, which was used in Babylonian, Chinese, and Indian mathematics. It seems that the [a.m.] approximations . . . were already known in pre-Babylonian mathematics (van Waerden 1983, p. 47); q.v. der (Folkerts 2006, p. II 2933). For 7, we get 7 = 32 2 2 2/3 and thereby 252 6 2 2/3 = 16. Therefore, we have a solution for the problem squaring the circle in whole numbers: the area of a circle with the diameter 18 equals the area of a square with the sides 16. This proportion 18/16, alternatively, 9/8 is also applied in problem 50 of the Rhind papyrus (ca. 1650 bc, copied from an earlier papyrus of ca. 18601814 bc) for the tasks squaring the circle and circling the square (van der Waerden 1983, p. 170). The deduced value 28/9 is the most accurate approximation for at the beginning of the third millennium bc known so far. The establishment of this gure is an admirable mathematical achievement for this time horizon. The Babylonians . . . always used = 3. This is also the value given by Vitruvius; and is found again in the Chinese literature. (van der Waerden 1954, p. 32). We will meet the approximation = 3 in the structures of Stonehenge also (see Sect. 5.1.). 3.3 Deduced measures The mean cubit measure for the ditch is 50.4931 cm (Table 1B). The arbitrary radius s1 of the circular arc with center D1 with the supposed length of 120 cubits or 2 ropes results in a cubit measure of 48.8285 cm. The radius r A of the Aubrey holes was set arbitrarily to 90 cubits or nine poles and results in a cubit measure of 49.6016 cm. Of all the measures deduced from the structures of Stonehenge, this measure is the one with the lowest range of error or the highest condence as we have a complete circle with 56 sampling points all lying on an exact circle. 4 Station stones and Heelstone The Station stones and/or holes display no circular or ring-like structure (Fig. 5a). The Station stones 91-94 determine a rectangle that is composed of two Pythagorean triangles (91-92-93 and 91-93-94 or vice versa 91-92-94 and 92-93-94); these triangles have the primitive shape (5,12,13) and are combined with the leg 13. The concrete length of the legs are 70, 168, and 182 cubits thus having a Pythagorean triangles of the real shape (70,168,182) and a multiplication factor of 14 compared to the primitive triangle. Another isosceles triangle composed of two mirrored Pythagorean triangles of the primitive shape (28,45,53) is added at the leg 91-94. This isosceles triangle has therefore the shape (56,53,53); the Heelstone 96 was placed at the top of this triangle. The concrete lengths of the legs of the Pythagorean triangles (28,45,53) are 84, 135, and 159 cubits; thus, the base leg of the isosceles triangle 91-94-96 is 168 cubits. This geometry represents the algebraic concept of the least common multiple (LCM) as 168 is the LCM of the integers 12 and 56.

123

The mathematics in the structures of Stonehenge

79

A
94

159 c [53]

c 70 ] [5
8] [2 c

5] c [4 5 13

84

93

c 8 16 12] [

[13 ] 182 c

91

5] 70

92

Fig. 5 Construction of the Station stones and the Heelstone. a The Station stones and the Heelstone form a geometric gure assembled by a rectangle and an isosceles triangle. Both the rectangle and the isosceles triangle are composed of Pythagorean triangles of the shape (5,12,13) and (28,45,53), respectively. The lengths of concrete legs are 70, 168 and 182 cubits and 84, 135, and 159 cubits. The leg 91-92-93-94 is 168 cubits which is the LCM of the integers 12 and 56 = 2 28. c Forward construction for the Station stones and the Heelstone. The rectangle 91-94 and the isosceles triangle 91-94-96 are computed by a mathematical exact construction

[53] 159 c
[2 8]

c 182 [13]

84

[1 2] 8 16 c

123

80

A. Kainzinger

Presently, the Heelstone 96 is leaning considerably toward SSE so we can assume an ancient construction point at the northwestern rim of the Heelstones present image. The exact forward construction was rotated adequately to cope with this fact (Fig. 5b). The bearing from the center of the rectangle to the Heelstone SC -96 has an angle to East of 39.65 . The deduced cubit measure is 49.2467 cm (Table 1C). 5 Sarsen stones The mathematics involved in the construction of the Sarsen stones shows a great leap forward compared with the preceding phases of Stonehenge. In this stage (ca. 2550 bc) as well as in the later phase of the Bluestones, elliptical concepts were adopted. 5.1 Sarsen circle The Sarsen circle is a completely exact circle. For the circular regression analysis, all the 17 still standing stones were incorporated. The reading points were taken at the innermost edge of the stones. This approach provided a best t of the regression circle with the exception of stones 10 and 11 which are not as exactly on the circle line as do all others (Fig. 6c). The radius of this circle with center S1 is 30 cubits which equals ve reeds. This supports the assumption that the inner faces of the stones were the basis for the construction concept as well as the fact that the inner faces of the uprights were made considerably smoother than the outer faces. By the construction of the 30 Sarsen circle uprights, we have another approximation of the circle constant encoded in the Stonehenge structures. We have identied two options of constructions: (a) if we apply the concept of the Aubrey holes circle, then the circumference of the Sarsen circle was taken as 30 reeds: radius of the Sarsen circle rS = 30 cubits = 5 reeds (Table 2A) diameter dS = 60 cubits = 10 reeds = 1 rope; circumference of the Sarsen circle cS = 30 reeds = 3 ropes; 3 ropes cS 30 reeds = = 3. circle constant is := = dS 10 reeds 1 rope The mathematical concept bases again on a regular polygon inscribed in a circle, in this case with 30 sides, resulting in an integer value for . For the area of the circle, we get AC = 52 3 = 75 reed2 . Applying the a.m. approximation for the calculation of a square root, we get 3 1 3/4 and for 75 = 52 3 5 1 3/4 = 8 3/4. The solution for the problem squaring the circle in this case is: the area of a circle with the diameter 10 equals the area of a square with the side 8 3/4. This Sarsen circle approximation for is worse than that established by the Aubrey holes circle some 400 years ago, but in accordance with the practiced standards in Babylonia. Either the Aubrey holes value of 28/9 was not communicated by oral tradition and it had fallen into oblivion, or it was a conscious decision to operate from now on with this simple integer value. The latter assumption could be motivated by having a more

123

The mathematics in the structures of Stonehenge

81

C
T8

T5 T2

D
T7 T1 T9 S1 S2

T1

rS

W
T3 T4

S1

[3] 3f

S2 S

T6
11

10 m

Fig. 6 Sarsen circle and horseshoe. a The ellipse dened by the center T1 , the major and minor axes and the tilt angle (points T2 through T5 ) are calculated by elliptical regression analysis. The center of the Sarsen circle S1 was calculated by circular regression analysis. The ellipse denes a rectangle and a regular lozenge. b The lozenge is composed of four Pythagorean triangles of the shape (28,45,53); the rectangle T6 T7 T8 T9 is composed of two Pythagorean triangles (T6 T8 T9 , and T6 T7 T8 or T6 T7 T9 and T7 T8 T9 , respectively) of the shape (28,45,53); each of these Pythagorean triangles has concrete legs of 28, 45, and 53 cubits. c Forward construction of the Sarsen circle and horseshoe. The ellipse is computed by a mathematical exact construction. d The centers S1 and T1 of the Sarsen circle and horseshoe aligned with the point of the compass dene a Pythagorean triangle S1 S2 T1 of the shape (3,4,5); this Pythagorean triangle has concrete legs of 3, 4, and 5 feet

123

82

A. Kainzinger

practical gure for circle calculations or because this change was related to ritual traditions; Seidenberg et al. showed that in Greece and India geometrical constructions were regarded important for ritual purposes (Seidenberg 1961, 1978, 1981; van der Waerden 1983). (b) In the above mentioned concept, the angle of a singular isosceles triangle with base 1 reed and legs 5 reeds has a central angle of 11.478 which is significantly smaller than the theoretical value of 360 /30 = 12 ; relating to the full circle, we get a difference by the 30 triangles of 15.65 which could doubtless be recognized by the builders of the Sarsen circle. A construction which ts the 30-side polygon better into the full circle could be based on an isosceles triangle with base 1 reed + 2 palm = 38 palms: radius of the Sarsen circle rS = 30 cubits = 5 reeds (Table 2A) diameter dS = 60 cubits = 360 palms circumference of the Sarsen circle cS = 30 38 palms = 1140 palms; 57 1140 palms cS = = 3.1666 . . . . = circle constant is := dS 360 palms 18 This value approximates better than the value 28/9 = 56/18 established by the Aubrey holes, and can be seen as a direct improvement of the latter value by raising the numerator from 56 to 57. For the area of the circle, we get AC = 52 57/18 = 475/6 = 92 11/6 92 2 reed2 . Applying the a.m. approximation for the calculation of a square root, we get 92 2 9 2/18 = 8 8/9. The solution for the problem squaring the circle in this case is: the area of a circle with the diameter 10 equals the area of a square with the side 8 8/9. To settle for one of the cases (a) or (b), we need more secured data about exact and regular stone or wood circles in coeval earthworks. 5.2 Sarsen horseshoe The seven Trilithons 5154, 57, 58, and 60 form an ellipse. Five points in a plane always form one and only one conic sectionin this convex situation exactly one ellipse. Hence, the accuracy of the seven reading points on the ellipse is remarkable (Fig. 6c). Once again the Trilithons were smoother at the inner faces, and therefore the reading points were also taken at the innermost edge of the stones. The fallen and broken Trilithon 59 was not incorporated in the computing as there are not sufcient data on the original position. However, the remaining eventually displaced base matches with the computed ellipse. The resulting major and minor axes of the ellipse generate the rhombus T2 T4 T3 T5 and the rectangle T6 T7 T8 T9 (Fig. 6a). Both geometric objects are composed of Pythagorean triangles of the shape (28,45,53). In the case of the triangle T6 T8 T9 (and the congruent triangles T6 T7 T8 , T6 T7 T9 and T7 T8 T9 alike), this fact is very impressive: these triangles have the concrete legs of 28, 45, and 53 cubits (Fig. 6b). The results of the elliptical regression analysis including the standard deviation values are provided as ESM. The Trilithon 56 was not obviously constructed at a position on the elliptical curve. This upright was straightened in 1901, and its partner stone 55 has fallen and broken.

123

The mathematics in the structures of Stonehenge

83

For a later comment, we want to state here that the depiction of Trilithon 56 in Fig. 8a shows the alignment of the smoother outer face not perpendicular to the major axis of the ellipse but rotated some 10 anticlockwise. If we assume that stone 55 has fallen in a right angle in relation to its original alignment, then this statement holds for stone 55 also. There are still reasonable and debatable arguments about the original position of both stones (Johnson 2008, pp. 136140). Therefore, the determination procedure according to Fig. 13 breaks down after step 1. The exact mathematical forward construction of the ellipse matches with that of the plan with a high degree of precision (Fig. 6c). Surprisingly, the centers of the Sarsen circle and the Sarsen ellipse are not identical. However, there is a comprehensible mathematical concept to explain this difference (Fig. 6d). The a.m. centers and the virtual point S2 congure the Pythagorean triangle S1 S2 T1 of the shape (3,4,5). The lengths of the legs are 3, 4, and 5 feet (Table 2A). The deduced foot measure is 29.8918 cm. The ancient foot measures vary considerably, even on the basis of a stable cubit measure depending on different subdivisions of the cubit and the palm (Rottlnder 1979).2 The deduced foot measure is therefore absolutely within the deduced overall measuring scheme. The displacement of the center T1 of the ellipse might be because the Sarsen horseshoe was thereby placed better in the middle of the Sarsen circle. 5.3 Deduced measures Once again, the above mentioned cubit measures are ideal values according to the selected mathematical conceptin these cases, the underlying circle and the Pythagorean triangles. The actual length values extracted from the plan are summarized in Table 2A. The cubit measure for the Sarsen circle is 48.8733 cm and that for the Sarsen horseshoe, 49.3194 cm; the mean of all Sarsen stones cubit measures is 49.2079 cm. 6 Bluestones The Bluestone circle and horseshoe were constructed some 450 years after the setting of the Sarsen stones. 6.1 Bluestone circle The present record of the remains of the Bluestone circle comprises 27 stones of which 12 have fallen. Hence, we can involve 15 stones in the determination procedure for the original construction plan. The original amount of stones generating the Bluestone circle is estimated between 44 and 62 uprights (Johnson 2008, p. 158); thus, we can imagine a very dense ringlike structure. Despite there being only 15 secured original positions, the nonlinear regression analysis provides a clear picture (Fig. 7c). We obtain four circular arcs with centers B21 B24 with 5, 6, 3, and 3 reading points,
2 Typical variations: foot on the statue of Gudea = 26.45 cm; 18 digit foot from the Nippur

cubit = 33.199 cm.

123

84 Table 2 Measures of the Sarsen stones and the Bluestones Segment Length gure ( mm) Length henge (m) No. cubits

A. Kainzinger

Result. cubit measure (cm)

A: Sarsen stones measures Measure Sarsen circle Radius T6 T9 T8 T9 T6 T8 76.279 71.222 116.204 136.2936 14.66199 13.68996 22.33618 26.19771 30a 28 45 53 Mean: Mean of all Sarsen stones: Circlehorseshoe relation No. feet S1 S2 S2 T1 S1 T1 4.7230 6.1575 7.7583 0.90783 1.18357 1.49126 3 4 8 Mean: B: Bluestones measures Bluestone circle Measure B21 B22 B21 B23 B22 B23 B21 B24 B23 B24 v1 v2 v3 v4 Measure B1 B6 B1 B11 B1 B2 B2 B6 B2 B11 B3 B6 B3 B11 B1 -79 52.0250 5.2062 14.7238 15.9908 37.3092 39.8863 66.2110 54.1300 62.6860 98.1440 65.3750 16.1440 16.3149 39.1920 42.0928 43.0451 42.6787 41.8508 39.4354 10.00000 1.00071 2.83014 3.07368 7.17140 7.66676 12.72677 10.40461 12.04921 18.86478 10.00000 2.46945 2.49559 5.99495 6.43867 6.58434 6.52829 6.40165 6.03218 5 5 12 13 13 13 13 12 49.3889 49.9117 49.9579 49.5282 50.6487 50.2176 49.2435 50.2682 12 35 37 84 91 Mean: 25 20 24b 36 50.0356 48.5167 49.8434 51.2243 50.5500 50.0340 50.9071 52.0231 50.2050 52.4022 Foot meas. 30.2611 29.5891 29.8253 29.8918 48.8733 48.8927 49.6360 49.4296 49.3194 49.2079 Sarsen horseshoe/Trilithons 52.0250 10.00000

Bluestones horseshoe

123

The mathematics in the structures of Stonehenge Table 2 Continued Segment Length gure ( mm) Length henge (m) No. cubits Mean: Circle and horeseshoe mean: B2 -79 B6 -79 55.4680 55.5642 8.48459 8.49930 Forward construction

85

Result. cubit measure (cm) 49.9506 49.9923

a 30 cubits = 5 reeds diameter = 10 reeds = 1 rope b 24 cubits = 4 reeds = 2 rods

correspondingly; the stones 40g and 47 are connecting points of the adjacent arcs. The circular arcs with centers B23 and B24 are calculated on the basis of three stone positions. As three points in a plane dene one and only one circle, we get an exact solution for the respective centers. However, this accuracy is deceptive because the regression analysis is calculated without redundancy. In this case, a defective position of one reading point (stone) results directly in a defective position of the center, while the regression calculation of a circle on the basis of more than three points compensates much better for slight position errors. Therefore, the circular arc with center B24 has the lowest probability in our assessment procedure as this arc has additionally a very small central angle from stone 47 to stone 31. Nevertheless, the results in combination with the deduced measure units pass our determination procedure according to Fig. 13with reservations as far as the center B24 is concerned. The construction starts with the Pythagorean triangle B21 B22 B23 of the shape (12,35,37) and the size 12, 35, and 37 palms (Fig. 7b). Again, this primitive form of the Pythagorean triangle is a strong support in favor of this explanation model. Another Pythagorean triangle of the shape (5,12,13) is added to the rst at the leg 35. As 35 is a multiple of 5, we get this triangle in terms of palms in the concrete shape of (35,84,91). The calculated radius of the circle with center B23 is 24 cubits which equals 4 reeds or 2 rods (see ESM Table 2). The construction of the ring on which the stones of the Bluestone circle were set started probably with this circular arc. The additional arcs were attached in a way to get a smooth curve; the ring does not close smoothly between the stones 31 and 32c (Fig. 7a). The exact forward construction (Fig. 7c) matches with the plan perfectly: the connection of the adjacent arcs at stone 40g is now totally smooth compared with the backward construction where the calculation provided a minor gap (Fig. 7a); the gap between stone 31 and 32c remains.

6.2 Bluestone horseshoe As far as mathematics is concerned, the complexity of Stonehenges construction concepts reaches its point of culmination with the Bluestone horseshoe. In combination with the holes 7378 the stones/holes of the Bluestone horseshoe create a ring-like structure. The nonlinear regression analysis provides three ellipses (Fig. 8a): a large ellipse with center B1 , a smaller ellipse in SW with center B6 and another smaller

123

86

A. Kainzinger

B23
3 [3 7 p 7]

91 [1 p 3]

] [1 2 p 84

B21 [1 p 12

2]

B22

B24

B23 B22 B21

B24

11

10 m

Fig. 7 Bluestone circle. a The points B21 B24 are the center points of the respective circular arcs provided by circular regression analysis. The arbitrary radius v3 is 24 cubits long which equals 4 reeds or 2 rods. b The triangle B21 B22 B23 is a Pythagorean triangle of the shape (12,35,37), and the triangle B21 B23 B24 is a Pythagorean triangles of the shape (5,12,13). The basic Pythagorean triangle B21 B22 B23 of the shape (12,35,37) has concrete legs of 12, 35, and 37 palms (p). c Forward construction of the Bluestone circle. The right-angled triangles B21 B22 B23 and B21 B23 B24 and the circular arc with center B23 are computed by a mathematical exact construction. The resulting curve of the Bluestone circle is totally smooth at the stones 40g and 47 compared with the one achieved by the backward construction. The gap between the stones 31 and 32c remains

123

The mathematics in the structures of Stonehenge

87

B19

B
B5 B14 B2 B8 B9 B1 B11 B18 B13

C
B5 B2
13 c [1 3 ]
d ro 1 ] = 8 c ] [2 12 [12 [13]

[53]
5] [4

B11
] [5 c 5
[53]

13 c
[53]

] [5

B1
5 c

13 c d [13] 1 ro ] = 28 c ] [2 12 [12

B16

B6

B12 B15 B10 B3 B4

B6
5] [4

[13 ] 13 c

[53]

B3

B4

B7

B17

Fig. 8 Bluestone horseshoe. a The ellipses with center B1 (large ellipse), center B6 (small ellipse SW), and center B11 (small ellipse NE) are calculated by elliptical regression analysis. b In the depiction of the construction concept, the small ellipse NE was shifted perpendicular to the major axis of the large ellipse so that the center B11 rests on this major axis; additionally, this small ellipse was rotated anticlockwise so that the axes coincide. c The basis of the construction concept of the Bluestone horseshoe is the lozenge B2 B6 B3 B11 composed of four Pythagorean triangles of the shape (5,12,13); each of these Pythagorean triangles has concrete legs of 5, 12, and 13 cubits. The large ellipse with center B1 denes the rectangle B16 B17 B18 B19 (see Fig. 13b) and a regular lozenge B2 B4 B3 B5

123

88

A. Kainzinger

ellipse in NE with center B11 . In the depiction of the construction concept in Fig. 8b, the center B11 of the small ellipse NE was shifted perpendicular to the major axis of the large ellipse so that this center rests on this major axis (length of shift 9.2 cm); additionally this small ellipse was rotated anticlockwise so that the axes coincide (rotation 5.67 ). The slight shift and rotation are within the respective condence intervals of the regression analysis due to the input data situation: the original position of the now disappeared stones 7378 cannot be established with the accuracy of the other elements of the Bluestone horseshoe. The basis of the construction concept of the Bluestone horseshoe is the lozenge B2 B6 B3 B11 composed of four Pythagorean triangles of the shape (5,12,13); each of these Pythagorean triangles has concrete legs of 5, 12, and 13 cubits. The large ellipse with center B1 denes the rectangle B16 B17 B18 B19 and a regular lozenge B2 B4 B3 B5 composed of four triangles similar to Pythagorean triangles of the shape (28,45,53). The proportion of the legs 45 and 28 is 45/28 = 1.607 . . ., and therefore also a good approximation for the Golden Ratio of 1.618 . . . (Fig. 8b, c). The small ellipse SW with center B6 denes the regular lozenge B7 B10 B8 B9 composed of four right-angled triangles being similar to a Pythagorean triangle of the shape (48,55,73) (Fig. 9a). The ellipse B7 B10 B8 B9 is significantly rotated anticlockwise by 16.362 in relation to the major axis of the large ellipse. We had stated above that the alignment of the great Trilithons of the Sarsen horseshoe might have been rotated the same way. This rotation of in the inner segment of both horseshoes may have a reason for astronomical or ritual purposes. The small ellipse NE with center B11 denes the regular lozenge B12 B15 B13 B14 composed of four right-angled triangles also being similar to a Pythagorean triangle of the shape (48,55,73) (Fig. 9b). Both small ellipses are tangential to the larger ellipsethe small ellipse SW with one tangent point and the small ellipse NE with two tangent points; hence, we can assume that the intention of the Bluestone horseshoe designers was to demonstrate this geometrical achievement. The positioning of hole/stone 79 gives another insight to the mathematics applied by the builders of Stonehenge in phase 3iv. (a) The line segment B1 -79 has the same length as the semi-minor axis B1 B2 (12 cubits = 1 rod) thus giving with segment B2 -79, the construction of the diagonal of a unit square or a geometrical representa tion of 2 (Figs. 8b, 9c). On the other hand, this diagonal equals the line segment B6 -79 with a difference of 1.471 cm as established in the forward construction (Table 2B). B1 B6 equals ve cubits, and so we get an approximation for the square root of 2 by 17/12 = 1.41666 . . . (exact value 1.41421 . . .). This approximation (1;25 in the Babylonian sexagesimal notation) frequently occurs in Babylonian texts (van der Waerden 1983, p. 47). (b) We have shown above that the proportion of the major and minor semi-axes of the large Bluestone ellipse is an approximation for the Golden Ratio; the point 79 therefore divides the major semi-axis B1 B5 in the Golden Section (Fig. 8b). We had assumed that the large ellipse was constructed on the basis of a regular lozenge based upon Pythagorean triangles of the shape (28,45,53). The construction of a Golden Section geometry as a rst step cannot be excluded totally either. A construction method for the Golden Section as also described in the Elements by EUCLID is depicted in Fig. 9d. This open question could be resolved if the remains of the respective construction pegs would ever be found in the soil of Stonehenge. (c) For the special case of an ellipse with the axes in the proportion of the Golden Ratio,

123

The mathematics in the structures of Stonehenge

89

B
B8
B2

B14
[73]

B13

5] [5

B9
[55

[73]

B11

[4

8]

B1
B1
[4 8]

B11

5] [5

B6
[55
B6

B12
[73] ]

]
B15

B10
[73]

B3

B3

B4

B7

C
B2
1 ro d
2 rod

B2

B6

ts bi cu

B1

= d ro ts 1 ubi c 17

12

cu

ts bi
B4 O2 G B1 a M b

Fig. 9 Bluestone horseshoe (II). a The small ellipse SW with center B6 denes the regular lozenge B7 B10 B8 B9 composed of four right-angled triangles being similar to a Pythagorean triangle of the shape (48,55,73). b The small ellipse NE with center B11 denes the regular lozenge B12 B15 B13 B14 composed of four right-angled triangles being similar to a Pythagorean triangle of the shape (48,55,73). The semi-major axis B1 B2 (or B1 B3 ) of the lozenge B2 B6 B3 B11 has equal length as the line segment B1 -79. c The line segments B6 -79 and B2 -79 have equal length thus having an approximation for the square root of 2 by 17/12. d Construction of a rectangle having the legs in the ratio of the Golden Section and construction of the osculating circles with centers O1 , O2 and O3

the point 79 is also the center of the minor osculating circle at the vertex (Fig. 9d). The construction of the four osculating circles at the vertices is a good rst step for an ellipse approximation. The knowledge and use of this concept is absolutely within the scope of a society constructing ellipses. Once again, traces of ancient construction pegs of the remaining three centers would resolve this option. In the exact forward construction in Fig. 10, the size of the two small ellipses based on Pythagorean triangles of the shape (48,55,73) was performed in such a way that these ellipses have the indicated tangent points with the large ellipse. This method provides the best matching with the Bluestone horseshoe plan. The geometrical

= 12 cu ts bi

79=O1

B5

B3

O3

123

90

A. Kainzinger

B 11

B 10

Fig. 10 Forward construction for the Bluestone horseshoe. The small ellipses with corresponding centers B6 and B11 are constructed similar to a Pythagorean triangle of the shape (48,55,73) and in the way that the indicated tangent points are achieved

construction of an ellipse being tangent to another given ellipse can be achieved by an approximation procedure. We can interpret the structures of ancient rings and earthworks as large mnemonic artifacts. Fortunately, there are also handy ancient artifacts discovered in the proximity of Stonehenge which underscore the layout concept of the Bluestone horseshoe: the Bush Barrow gold lozenges. The large lozenge displays the underlying lozenge of the two small Bluestone horseshoe ellipses based upon a Pythagorean triangle of the shape (48,55,73). For the investigation, we use the scaled graphic rendering in Fig. 11a. The angles 41.121 in the determining triangle L 1 L 3 L 2 matches the respective angles 41.112 in the Pythagorean triangle (48,55,73) nearly exactly (Fig. 11b). However, there is another way to conrm this explanatory model: when applying the determination procedure according to Fig. 13 to the gold lozenge also, we have to establish a consistent metrology for the manufacture of the lozenge. The lozenge is now 18.55 cm

123

The mathematics in the structures of Stonehenge

91

L2

L3

L5

B
L7

L4 L2 L6

L12

3] [7
] [53
[53 ]

L3

[55]

L1

3] [7 corn 3 7

[55]

L5

L13

[45]
[53

L11

[45]
] [53

L15

L14
mean ( L11L13L12,

L4 L8 [55] 55 corn
L9L8L6 = L1L3L2 = 41.121 Pythagorean triangle (48,55,73) = 41.112 41.112

L11L13L14, L11L15L12, L11L15L14) = 31.304 Pythagorean triangle (28,45,53) = 31.891

L9

Fig. 11 Bush Barrow gold lozenges. a Bush Barrow large lozenge. The lozenge denes the regular rhombus L 2 L 3 L 4 L 5 . b The rhombus is composed of four Pythagorean triangles of the shape (48,55,73). Hence, the triangle L 8 L 9 L 6 is a Pythagorean triangle of the shape (48,55,73) also, and this triangle has concrete legs of 48, 55, and 73 corns as well as the axes of the rhombus L 2 L 3 L 4 L 5 having the lengths of 48 and 55 corns. c Bush Barrow small lozenge. The lozenge denes the regular rhombus L 12 L 13 L 14 L 15 . d The rhombus is composed of four Pythagorean triangles of the shape (28,45,53)

long; it is considerably compressed at the corners, and the original size should be slightly larger due to the crumbling of the plate. Let us, therefore, assume as working hypotheses that the original distance between the vertices L3 and L5 was 19 cm. 19 cm divided by 55 provides 0.34545 cm; taking this measure as a unit for the corn measure, we have to multiply by 144 and get a cubit measure of 49.7455 cm. This cubit is equivalent to the established cubit for the Bluestone horseshoe of 49.9506 cm (2B). Hence, the triangles L 8 L 9 L 6 (and L 6 L 7 L 8 alike) are Pythagorean triangles with the concrete legs of 48, 55 and 73 corns and the axes of the lozenge have a length of 48 and 55 corns, respectively. The small Bush Barrow gold lozenge was not manufactured with the precision and artistry of the large one. The production of this tiny plate may date back to the early

123

92

A. Kainzinger

phases of the manufacture of gold plates. This plate could have been passed over generations to the man to whom it was attached in his tomb. The marvelous large lozenge seems to reect the culmination of that kind of plate. Nevertheless, the small gold lozenge could have played a comparable role. For the backward construction, we use a photo as no adequate scaled rendering is available (Fig. 11c); therefore, we have to take into account potential slight photographic distortions. The backward construction is very close to a regular rhombus composed of Pythagorean triangles of the shape (28,45,53) (Fig. 11d). This Pythagorean triangle is the basis of the construction of the large ellipse of the Sarsen horseshoe, and the large ellipse of the Bluestone horseshoe as well. The length of the major axis of the small lozenge is 3.1 cm; the division by 45 yields a measure unit of 0.06888 cmfar below the unit of a corn. We have no secured sources for a subdivision of the unit corn in antiquity. Following the old English unit (shoe) iron of 0.0529167 cm (1/48 inch), let us denote this potential unit an iron. A subdivision of a corn by six irons yields a cubit measure of 59.5199 cm and a subdivision by ve irons of 49.5999 cm. The latter measure matches with the established cubit measure scheme for Stonehenge, although this can only be an indication. 6.3 Deduced measures The actual length values extracted from the plan are summarized in Table 2B. The mean cubit measure for the Bluestone circle is 50.0340 cm, and the mean for the Bluestone horseshoe 49.9506 cm. As both values match to a high degree, we have established a mean cubit measure for all Bluestones of 49.9923 cm. 7 Metrology All measurements of the Stonehenge structures are based on a consistent metrology. This metrology is in line with practiced standards in Babylonia concerning both the principal system of measures and the concrete measures. Verications of mathematical constructions without written sources get an additional afrmation if they are based on a consistent and coherent metrology and correspond with coeval veried systems of measures. In the construction of the Stonehenge structures, length measures were used as we encounter them in Babylonia. A preferred unit to compare ancient measures of length is the cubit. In the common Babylonian system of measures, a cubit was subdivided into six palms, a palm into four (st. ve) ngers, and a nger into six corns; six cubits equal one reed and 12 cubits equal one rod; 10 cubits equal one pole, and 10 reeds equal one rope (Trapp and Wallerus 2006).3 The unit 10 palms was also a used measure; this unit is commonly denoted as Megalithic yard (my) (see also ESM Fig. 2). The system of measures used in the geometrical construction of Stonehenge conforms to this Babylonian system. We have identied respective cubit measures for the phases of Stonehenge under investiga3 The Babylonian corn measure differs from the old English barleycorn measure. In the latter case, the basis for the measure is the length of a barleycorn while a Babylonian corn (e) is about the third of this measure and seems to represent the width of a corn.

123

The mathematics in the structures of Stonehenge

93

cubit measure [cm]

53 52 51 50 49 48 47 46 -3000 -2800 -2600 -2400 -2200 time [year] -2000 -1800

3vi

-1600

measure units structure Aubrey Holes Station stones, Heel stone Sarsen circle and horseshoe Bluestone circle and horseshoe Z holes Y holes
1)

phase time 1 3i 3ii 3iv 3vi 3vi

cubit measure -2950 49.6016 -26501) 49.2467 -2550 -2100 -1900 -1600 49.2079 49.9923 50.7228 53.7403

currently undated, 'before the central stones' (Johnson 2008, p. 136)

Fig. 12 System of measures. The system of measures corresponds with practiced standards in Babylonia

tion. The length of the cubit has a comprehensible development scheme from phase 1 (ca. 2950 bc) to the end of phase 3vi (ca. 1600 bc) (Fig. 12). A considerable increase in length measures in times of decline is an often-noticed fact. The cubit measure 49.99 cm established for Stonehenge phase 3iv (ca. 2100 bc) corresponds with the coeval Babylonian cubit of Lagash of 49.61 cm.4 This Lagash value is identical to the cubit measure of the Aubrey Holes of 49.60 cm, which we had identied as the cubit measure with the lowest range of error. Hence, the consistency of the Stonehenge measuring system is a strong support favoring the established mathematical concepts encoded in the structures of Stonehenge.

4 The cubit of Lagash was documented on a statue of Gudea of Lagash. The current status of research does

not provide exact data for the ruling time of Gudea. New approaches vary from 2140 to 2060 bc.

123

94

A. Kainzinger

8 Methods The construction of exact circles in ancient earthworks and stone/wood rings can be veried easily. The deduction of construction concepts of noncircular, ringlike structures without written sources is a complex task with a considerable number of pitfalls. To avoid misinterpretations to a great extent and to work out a highly probable solution a six-step determination procedure was established (Fig. 13). In terms of science theory, this procedure should be taken as nal conclusions as scientic proofs for concepts of illiterate societies are hardly to be established. An explanatory model for a potential construction concept can be accepted only if all steps of the procedure are passed successfully. The procedure can break down at every stepexcept step 4. Step 1 If the plan of the present situation of the monument does not contain sufcient data, then the procedure should be canceled at this point. Secured correction can be made, e.g., if the original position of a leaning stone is obvious. Step 2 To establish circular or elliptical arcs in the backward calculations in a reproducible manner, nonlinear regression analysis methods are inevitable. Step 3 If the selected construction concept does not match with a consistent system of measurements, the procedure should be canceled at this point (exception: similar transformations). Step 4 The forward construction yields an exact mathematical construction model. Step 5 Owing to both the decay of ditches/stones/holes and the potential imprecise construction of individual elements of the original monument, the concept of step 2 is no exact model in general. Therefore, slight modications of the measure, position
original layout of the earthwork ancient secured corrections decay
1

original construction concept

present concluding t assessment of the result

plan of present situation

other sources/ research results

backward calculation potential 2 concepts of construction forward calculation potential system of measurements
3

matching 5 of plan and construction slight modification of measure, position & heading exact 4 construction model

Fig. 13 Determination procedure. The determination of the original construction concept and potential earth work layout is performed on the basis of a six-step procedure. The procedure can break down at all steps except step 4

123

The mathematics in the structures of Stonehenge

95

and alignment of the step 4 model can be made in step 5. If this forward construction model results in a worse matching compared to the backward construction, then the procedure should be canceled at this point in general. Step 6 The potential solution of step 5 should pass an overall assessment taking into account also superordinate arguments. The circles and circular/elliptical arcs in the backward calculation were computed by a nonlinear regression analysis. For the analysis of line elements (ditch, bank), preferable eight or more points of the respective section of the curve should be identied. In case of particular irregularities of curved lines, the reading points should be taken on the basis of the sampling theorem, as performed in the line determination of the excavated part of the ditch. In case of stones and holes, the number of reading points is predened. Centers of stones and holes were established by a visual approximation. The center of gravity of shapes should be taken if the shape has an irregular noncircular image and no additional information about the original center is available. The results of the regression analysis including the standard deviations are provided as Supplementary Material (ESM Tables 14). 9 Summary All the stone structures of the Stonehenge enclosure and probably also the ditch were designed, constructed, and set on the basis of mathematical concepts. We could identify only one exception: in times of an obvious decline of the mathematical traditions,

Table 3 List of the rst 12 primitive Pythagorean triangles. rope length denotes the sum of the lengths of the three legs First 12 primitive Pythagorean triangles (ordered by rope length) Primitive Pythagorean triangle (3,4,5) (5,12,13) (8,15,17) (7,24,25) (20,21,29) (12,35,37) (9,40,41) (28,45,53) (11,60,61) (16,63,65) (33,56,65) (48,55,73) Rope length Used in constructions Element

12 30 40 56 70 84 90 126 132 144 154 176

II (I) IIII

Sarsen circle/horseshoe, Z holes, (ditch) Station stones, Bluestone circle, Bluestone horseshoe, Y holes

I III

Bluestone circle Heelstone, Sarsen horseshoe, Bluestone horseshoe

II

Bluestone horseshoe (II)

123

96

A. Kainzinger

the stone indicated by hole Y8 of the Y holes was set presumably arbitrarily. The major mathematical concept in the design of the Stonehenge structures is the application of Pythagorean triangles. This is in line with earlier and coeval earthworks and stone rings. Among the outstanding Pythagorean triangle (3,4,5), specic Pythagorean triangles were repeatedly applied in the construction concepts of Stonehenge: (5,12,13), (28,45,53), and (48,55,73)the last two with considerable large rope lengths (see Table 3). The Pythagorean triangle (28,45,53) is applied for the design of both the ellipse of the Sarsen horseshoe and the large ellipse of the Bluestone horseshoe. The Pythagorean triangle (48,55,73) seems to be discovered in the forefront of the construction of the Bluestone horseshoe, and the design of the two smaller ellipses was dedicated to this discovery, while the shape of the large ellipse was designed in the tradition of the Sarsen horseshoe ellipse, indicating a practice bequeathed over centuries. The dedication of the Pythagorean triangle (48,55,73) to Stonehenge is underscored by the identical design of the Bush Borrow large lozenge. In general, Pythagorean triangles were composed to form rectangles, isosceles triangles, rhombi, and general polygons. The standard concept for the design of ringlike structures was the composition of circular arcs orin the case of the horseshoeselliptical arcs to predominantly smooth lines; we could notice the total closing of the line (mathematical continuity) in the Bluestone horseshoe only. In this phase, the application of mathematics with the involvement of elliptical concepts was at its peak. As far as geometry is concerned the construction of ellipses being tangent to a given ellipse is the highlight of the Stonehenge mathematical designs. The concrete construction method of the ellipses could not be disclosed. The encoding of approximations for the circle constant (28/9, 3 and 57/18 respectively) for the task squaring the circle and of the approximation 17/12 for the square root of 2 in the construction of Stonehenge is another important achievement as is the application of the law of similitude. The a.m. concepts are encoded in the structures of Stonehenge. It can be assumed that the general level of mathematical achievements in the Stonehenge society was remarkably of a high standardwith a distinct culmination in phase 3iv and an obvious time of decline in phase 3vi (Y holes). The used length metrology corresponds with common Babylonian systems of measurement based on the unit of a cubit, broken down till the measure of a corn and summed up to the measure units reed, pole, and rope. Systems of measurement change rarely; we can, therefore, assume a common origin of these metrologies noticeable before the time under consideration. In addition, this consistent metrology establishes high condence in the described mathematical concepts. References
Folkerts, Menso. 2006. The development of mathematics in Medieval Europe. Aldershot: Ashgate. Haarmann, Harald. 1991. Universalgeschichte der Schrift, 3850. Frankfurt-New York: Campus. Hawkins, Gerald S. 1965. Stonehenge decoded. Garden City: Doubleday. Johnson, Anthony. 2008. Solving Stonehenge. London: Thames & Hudson. Mathews, Jerold. 1985. A neolithic oral tradition for the van der Waerden/Seidenberg origin of mathematics. Archive for History of Exact Sciences 34: 193220. Neugebauer, Otto. 1969. The exact sciences in antiquity, 2nd ed. Mineola: Dover Publication. Neugebauer, Otto. 1975. A history of ancient mathematical astronomy part one-three. New York: Springer.

123

The mathematics in the structures of Stonehenge

97

Rottlnder, Rolf C.A. 1979. Antike Lngenmae, 4043. Braunschweig/Wiesbaden: Vieweg. Seidenberg, Abraham. 1961. The ritual origin of geometry. Archive for History of Exact Sciences 1: 488527. Seidenberg, Abraham. 1978. The origin of mathematics. Archive for History of Exact Sciences 18: 301342. Seidenberg, Abraham. 1981. The ritual origin of the circle and square. Archive for History of Exact Sciences 25: 270327. Thom, Alexander, and A.S. Thom. 1980. Megalithic rings: plans and data for 229 Monuments in Britain. British Archaeological Reports, Br. Series 81. Trapp, Wolfgang, and Heinz Wallerus. 2006. Handbuch der Mae, Zahlen, Gewichte und der Zeitrechnung, 193194. Ditzingen: Reclam. van der Waerden, Bartel L. 1954. Science awakening. Groningen: Nordhoff. van der Waerden, Bartel L. 1973. Science awakening II. In Birth of Astronomy. New York: Springer. van der Waerden, Bartel L. 1983. Geometry and algebra in ancient civilizations. New York: Springer. Vansina, Jan. 1985. Oral tradition as history, 4445. Wisconsin: University of Wisconsin Press. Wuing, Hans. 2008. 6000 Jahre Mathematik. Berlin: Springer.

123