This is a revision. Section 5 cuts omit "weakly level 1" results, since weakly level 1 => level 1 (as pointed out to me by Rainer Schulze-Pillot.)

We give upper bounds on the size of the gap between the constant term and the next non-zero Fourier coefficient of an entire modular form of given weight for \Gamma_0(2). Numerical evidence indicates that a sharper bound holds for the weights h \equiv 2 . We derive upper bounds for the minimum positive integer represented by level two even positive-definite quadratic forms. Our data suggest that, for certain meromorphic modular forms and p=2,3, the p-order of the constant term is related to the base-p expansion of the order of the pole at infinity, and they suggest a connection between divisibility properties of the Ramanujan tau function and those of the Fourier coefficients of 1/j.

To appear in "Experimental Mathematics". 25 pages.

Barry Brent <barryb@primenet.com>