------start "referee" report---------------------
The authors devised a recursive procedure for enhanced computation of the
coefficients C_{0,1,k}(N) which produced more specific data. The latter
formed the basis of a conjecture on the C_{0,1,k}(N) (Conjecture 2.1 in the
paper). However, this conjecture does not ring with enduring significance
because it is neither based on general premises nor constructive. There are
usually many ways to say that a single analytic property fails, especially
as more extensive data become available. It is useful only to the extent
that future researchers on the topic  might build on it.
[A better approach (though not to be imposed on the authors presently)
would be to return to Rademacher's original analysis and attempt to modify
his derivation with the aim of obtaining a conjecture that is likely to be
true.]

The second conjecture (Conjecture 3.1)  is more useful for further work
since it relates to a rule for yet another class of the C_{0,1,k}(N).

------end "referee" report---------------------

----start editor's message------------

Based on the referee comments, we regret that we cannot accept your paper
at this time in its current form. I would be willing to consider a new
version,
which takes into account the remarks of the referee. In particular, you
should revise your text around Conjecture 2.1, or eliminate this conjecture.

The new version would be treated as a separate submission.

We appreciate your consideration of Experimental Mathematics.


Sincerely,

Yuri Tschinkel
Managing Editor, Experimental Mathematics
tschinkel@cims.nyu.edu
-------------------------------------end editor's message-----------