Ideal 9th-Order Multigrades and Letac's Elliptic Curve (Q3980482)

An `ideal \(n\)th-order multigrade' consists of two non-identical sets \(S_ 1\), \(S_ 2\) of \(n+1\) integers such that the polynomial \(\prod_{k\in S_ 1}(T-k)-\prod_{k\in S_ 2}(T-k)\) is constant; equivalently, such that \(\sum_{k\in S_ 1} k^ i=\sum_{k\in S_ 2} k^ i\) for \(i=0,\dots,n\). There exists an old method, apparently invented by Letac, of obtaining 9th order multigrades. One starts with a pair of rational numbers \(u,v\) satisfying \((u^ 2-13)(v^ 2-13)=48\) and then one writes down two sets of 10 simple expressions involving \(u\) and \(v\) which turn out to be the desired multigrade. In the present note the author shows that the elliptic curve which is defined by the relation between \(u\) and \(v\) here, has a group of rational points isomorphic to \(\mathbb{Z}/2\mathbb{Z}\times \mathbb{Z}/4\mathbb{Z}\times \mathbb{Z}\). Using that there are infinitely many rational points he proves that Letac's method produces infinitely many genuinely different multigrades; i.e., not obtainable from each other by affine transformations. The author also claims that he found generators for the group of rational points on the elliptic curve, but he omits to include a proof of that. However, his assertion is correct as can be readily proven using the method explained in \textit{J. H. Silverman}'s paper [Math. Comput. 55, 723- 743 (1990; Zbl 0729.14026)].