Who first noticed Somos4 sequences satisfied theta fcn recurrences--Elkies? See, e.g., the "period 8" 0,1,1,-1,-2/3,1/3,1/9,-1/27,0,... near the top of p13 of http://arxiv.org/abs/math/0703470 . I still don't understand why we never need a geometric scale factor (e.g. c^n). At any rate, ISC doesn't recognize any of the four buggerfactors b, u, y, and Q. No problem, just use Jacobi's imaginary transformation, and we get, oops, four new unrecognizables. --rwg