# 4900, a particularly square number

I read in "Letters to a young mathematician" that 4900 is the only square integer that is the sum of consecutive squares (I believe he meant by that "starting from 1", but maybe that's not even necessary). I did a quick run through with a python script and of course this seems totally devoid of a computational pattern. Why is 4900 (and 1 of course) the only number such that this works?

I did find out that the sum of squares is the following...

$\sum^{n}_{i=1} i^2 = \frac{n(n+1)(2n+1)}{6}$

-
What is the question? – Jonas Meyer May 31 '10 at 8:15
– KConrad May 31 '10 at 8:19
Jonas, you could just put that as the answer, that'll get me where I need to be. The CannonballProblem link doesn't help unless you have the papers. – Michael Hoffman May 31 '10 at 8:36
It seems no one has picked up on the question of whether "starting from 1" is necessary. Of course, it is; trivially, every square integer is the sum of (one) consecutive square(s). Slightly less trivially, $5^2=3^2+4^2$, $29^2=20^2+21^2$ (the first two members of an infinite sequence of examples obtainable via Pell). There's also $38^2+39^2+\dots+48^2=143^2$, again the first in an infinite sequence of square sums of 11 consecutive squares, and many, many more. – Gerry Myerson Jun 30 '10 at 2:55

I remember vaguely that the equation $1^2+\dots+24^2=70^2$ gives a construction of the Leech lattice (one has to consider one of the two even neighbours of the lattice $\mathbb Z \frac{w}{70}+\Lambda'$ where $w=(1,2,\dots,24)$ and $\Lambda'=\lbrace z\in\mathbb Z^{24}|\langle z,w\rangle\in70\mathbb Z\rbrace$) which is a rather unique and exceptional structure.
See en.wikipedia.org/wiki/Leech_lattice for an outline of this construction; it uses the vector $(0,1,2,\ldots,24,70)$ in the Lorentzian even unimodular lattice of rank 26 and signature 24. – Robin Chapman May 31 '10 at 13:17