Close Encounters of the Rademacher kind By Shalosh B. Ekhad The great analytic number theorist, Hans Rademacher in his posthumously published book "Topics in Analytic Number Theory", Springer, 1973 (pp. 299-302) conjectured that C_{hkl}(N), the coefficient of (x-exp(2*Pi*I*h/k)^(-l) in the partial fraction decomposition of 1 ------------------------ i Prod(1 - x , i = 1 .. N) converges to a certain number defined in his book and implemented in procedure ChklInfinity(h,k,l) in The Maple package HANS, available, free of charge from http://www.math.rutgers.edu/~zeilberg/tokhniot/HANS . It turns out (empirically), that these sequences diverge very badly, and as N goes to infinity, sooner or later they oscillate widely with peaks that go to infinity (exponentially fast!) and valleys that go to negative infinity (also exponentially fast!) But it seems that one can tweak Hans's conjecture if instead of going all the way to infinity, you only go so far. In this article we will find, for h and k, 1<=h