110
PYTHAGOREAN ARITHMETIC
Then the gnomon
BG' D = BB'. BG+DD'. G'D'
= BB' (BG + G'D').
Now BG = \n (n +1),
G'B' = 1 + 2 + 3 + ... + {n~ 1) = ±n{n— 1), BB' = n\
therefore (gnomon BG'D) = n ,n~ = n 2 .
Similarly (gnomon B'C"D') = {n — l) 3 , and so on.
Therefore l 3 + 2 3 + ... + n z = the sum of the gnomons round
the small square at A which has 1 for its side plus that small
square; that is,
l 3 + 2 3 + 3 3 + ... + n z — square AC = n (n+ l)} 2 .
It is easy to see that the first gnomon about the small
square at A is 3 + 5 = 2 3 , the next gnomon is 7 + 9 + 11 = 3 3 ,
and so on.
The demonstration therefore hangs together with the
theorem stated by Nicomachus. Two alternatives are possible.
Alkarkhi may have devised the proof himself in the Greek
manner, following the hint supplied by Nicomachus’s theorem.
Or he may have found the whole proof set out in some
Greek treatise now lost and reproduced it. Whichever alter
native is the true one, we can hardly doubt the Greek origin
of the summation of the series of cubes.
Nicomachus passes to the theory of arithmetical proportion
and the various means (cc. 21-9), a description of which has
already been given (p. 87 above). There are a few more
propositions to be mentioned under this head. If a-^h = h—c,
so that a, h, c are in arithmetical progression, then (c. 23. 6)
h- —ac = (a—h) 2 = (h—c) 2 ,