114
PYTHAGOREAN ARITHMETIC
as a race-course (SictvXo?) 1 formed of successive numb ers
from 1 (as start, vanXgg) up to n, the side of the square,
which is the turning-point (Kap-mgp), and then back again
through {n — 1), (77 — 2), &c., to 1 (the goal, vvacra), thus:
1+2 + 3 + 4... (tt,-1) +
n
1 + 2 + 3 + 4...(t7/-2) + (77-1) + •
This is of course equivalent to the proposition that n 2 is the
sum of the two triangular numbers ^n{n+\) and %(n—l)n
with sides n and n — 1 respectively. Similarly Iamblichus
points out 2 that 'the oblong number
71 (n — l) = (1 + 2 + 3 + ... + 77/) -f- (77/ — 2 + 71 — 3 + ... -|-3 + 2).
He observes that it was on this principle that, after 10,
which was called the unit of the second course (Sevrepco-
Sovpevr] povds), the Pythagoreans regarded 100 = 10.10 as
the unit of the third course (rpiooSovptvr] povds), 1000 = 10 3
as the unit of the fourth course (rerpcoSoopevr] pouds), and
so on, 3 since
1+24-3 + ... + 10 + 9 + 8-1-...+2 + 1 — 10.10,
10 + 20 + 30 + ... + 100 + 90 + 80 + ...+ 20 + 10 = 10 3 ,
100+ 200+ 300+ ... + 1000+ 900+ ... + 200+ 100 = 10 4 ,
and so on. Iamblichus sees herein the special virtue of 10 :
but of course the same formulae would hold in any scale
of notation as well as the decimal.
In connexion with this Pythagorean decimal terminology
Iamblichus gives a proposition of the greatest interest. 4
Suppose we have any three consecutive numbers the greatest
of which is divisible by 3, Take the sum of the three
numbers; this will consist of a certain number of units,
a certain number of tens, a certain number of hundreds, and
so on. Now take the units in the said sum as they are, then
as many units as there are tens in the sum, as many units as
there are hundreds, and so on, and add all the units so
obtained together (i.e, add the digits of the sum expressed
in our decimal notation). Apply the same procedure to the
1 Iambi, in Nlcom., p. 75. 25-77. 4. 2 lb., pp. 77. 4-80. 9.
3 Ih, pp. 88. 15-90. 2. 4 Ih., pp. 103. 10-104. 13. ’