A well-known empiricism in the rheology of polymer melts is the Cox-Merz rule, which relates the linear dynamic moduli as functions of frequency to the steady shear flow viscosity as a function of shear rate. Theoretically, however, these functions are quite independent, and there is no general reason for such a correspondence to exist. In this paper, some rational hypotheses on a constitutive theory are formulated which imply that a qualitative correlation exists.