Games with geometrical figures

fig. 3 1

16 3 2 13 fig 33 fig 34

5 10 11 8
12 27 6 16 36 8
9 61 7 12
9 15 21 12 20 28
4 151 14 1
fig. 32 24 3 18 32 4 24
X2 =4 2x9-18 2X4-8 2x15=30
fig. 35

2 x 7=14 2X5=10 2x3=6 2x15=30

2 9 4
2X6=12 2X1 =2 2x8=16 2x15=30
7 5 3
2x15=30 2x15=30 2x15=30 2x15=30
6 1 8
trance to his school: "Let no one enter who mas-
ters no geometry." fig. 36 fig, 37
The link between mathematics, geometry and
art also underlies the work of the German painter
Albrecht Durer, a contemporary of da Vinci. In 6 1 8 2 7 6
Durer's noted engraving Melencolia, there is a
magic square, often considered the first example 7 5 3 9 5 1
of one seen in the West. It is constructed so the
rows, columns and diagonals add up to 34 2 9 4 4 3 8
(Fig. 31). Moreover, the four central compart-
ments add up to 34. The second and third com-
partments in the bottom row indicate the date of
composition: 1514. Aside from the intimacy be-
tween the arts and sciences during the Renais-
sance, perhaps another reason for DOrer's inclu-
sion of a magic square in his engraving, is that
fourth-order squares were thought to possess
special therapeutic virtues. Indeed, astrologers of multiplying by 2 gives a sum of 2 x 15 = 30
the period advised wearing them as amulets to (Fig. 32). Similarly, Figs. 33 and 34 are construc-
dispel melancholy. ted by multiplying by 3 and 4 the entries of the
Let us move on to other magic squares. Take square in Fig. 29. If we exclude rotations and
the third-order square of Fig. 29 and multiply its reflections, there exists a unique magic square of
entries by a constant to produce new entries. the third order. Figs. 35 and 36 are only a single
These will again add up to a constant. Indeed, third-order square with a reflection about the cen-

46
 Games with geometrical figures

16 2 3 13 7 12 1 14 4 15 6 9
5 11 10 8 2 13 8 11 3 16
10
9 7 6 12 16 3 10 5 11 8 13 2
141
fig. 38 4 14 15 1 fig 39
9 6 15 4 fig. 40
1 12
:R~~ .::
.4:
..... .777
.

7V12 11 b C a d f e
21t38 11 C a b e d f
16 10 5 a b C f e d

fig 4 1
a9 6 154 fig. 42
b C a fig. 43
d f e

tral row, while Fig. 37 has been rotated around a
diagonal. With higher orders, the number of
arrangements increases. A fourth-order square
allows 880 different placings of its 16 numbers,
excluding reflections and rotations. This was
first discovered by the mathematician Bernard
Frenicle de Bessy in 1693. Figs. 38, 39 and 40
show some of these solutions, with the sum 34. It
is not yet clear what mathematical law governs the
disposition of numbers in magic squares. The
question remains open, and the known solutions
have only been discovered by trial and error.
How many fifth-order magic squares are there?
Until recently the estimate was about 13,000,000.
In 1973, however, Richard Schroeppel, a pro-
grammer with Information International, deter-
mined the exact number with the aid of a modern
computer (His findings were later published in
Scientific American, vol. 234, no. 1, Jan 1976).
Without counting rotations and reflections, there
are 275,305,224 different solutions.
Diabolic squares
These are even more intriguing than magic
squares, because of their additional properties.
Again consider Durer's square as it is rearranged
in Fig. 39 and repeated in Fig. 41. The sum of the
four central squares is 13 + 8 + 3 + 10 = 34, as is
the sum of the four corners and the vertical as well
as the horizontal off-diagonal squares. Such
magic squares are called "pandiagonal." The
same constant results if we add the set of four
numbers marked in Figs. 42 and 43.
Similarly we can form fifth-order magic squares
with particular properties. In Fig. 44 the sum is
always 65. Considering corner numbers plus the
central one, as well as diagonal numbers (includ-
ing the central one), we always produce the
same sum: 9+13+17+1 +25=65, 7+5+
21 + 19+ 13=65.
A square of order 5 in which any pair of num-
bers opposite the center adds up to n 2 + 1 (n

47
 Games with geometrical figures
fig. 44
115 24 8 117
23 7 16 :514
20 4 13 22 6
12 2 10p193
9182 1125
being the order) is called ''associative." Here,
n2 + 1 = 26. Thus, horizontally, 20 is opposed to
6, and 20 + 6 26; diagonally, 17 is opposed to
9 and 17 +9 = 26, 25 is opposed to 1, and
25 + 1 = 26. The Lo shu of Fig. 29 also has this
property and is therefore associative. Indeed,
with n = 3, n2 + 1 = 10, and in the square
7+3=10, 4+6=10, 8+2=10, 9+1 =10. A
fourth-order square may be either associative or
pandiagonal, but never both. The smallest square
that can be both is of the fifth order. If, as usual,
we exclude rotations and reflections, there are
only 16 fifth-order squares with both properties
according to Schroeppel's calculations.
In Medieval times, the Moslems imbued pan-
diagonal squares of order 5 with 1 at the center
with mystic significance, for number 1 is the sa-
cred symbol of Allah, the Supreme Being. The
problem of representing God and the concept of
God occurs in all religions and theologies. The
symbol that best evokes the unity of being is the
48
fig 45
0,
0-N09
09 fig 46
r0 0
0\
kV
number 1. God is one. However, the Moslem con-
ception of God is such that no sign or picture can
adequately represent Him, not even the most ab-
stract and immaterial such as the number 1.
Hence, in some magic squares the ineffable na-
ture of the Supreme Being is suggested by leav-
ing the central square empty.
Magic stars
Similar features are observed in other geometrical
figures such as magic stars. Take twelve counters
numbered from 1 to 12 (Fig. 45) and construct
a star of David from two equilateral triangles
(Fig. 46). Now place the counters on vertices and
intersections so the numbers along each of the six
sides add up to the same sum. As before this can
be achieved by trial and error. In Fig. 47, the sum
is always 26. However, if we add up the six verti-
cesweget3+2+9+11 +4+1 =30.