An equation has recently been derived for the specific viscosity of a dilute solution of ``randomly‐kinked'' chain molecules, which, for long chains and small concentrations, is of the same form as Staudinger's empirical ``law'': &eegr;sp=ksnc. The validity of this equation for actual solutions, and the nature and magnitude of the deviations to be expected are discussed. Among the effects considered are those due to restriction of rotation about the bonds in the chain and to the size and nature of the solvent molecules, of the repeating units in the chain, and of the end groups.