A new theory is proposed involving a short-range interaction parameter, N0, which allows for finite chain effects. The parameter z, which was originally suggested by Yamakawa, is redefined to make it applicable to finite chains. The dependence of intrinsic viscosity on molecular mass as predicted by most of the acknowledged theories of polymer solution seem to be incompatible with the Mark-Houwink-Sakurada relation. This makes the evaluation of interaction parameters from viscosity data difficult. Therefore, an attempt is made here to show that these theories contain some basic assumptions (and/or mathematical simplifications) which may lead to erroneous conclusions.