> Fix HASH. Right :-) ... However, I'm not convinced yet that HASH is the real problem. Is there any reason that reg_qty of a given reg might change from one call of HASH to the next? Christian