The evidence for and against the interpretation of the ``second‐order'' transition in polymers in terms of a thermodynamic transition of the second order is reviewed and some new experimental data concerning transition in mixtures of polymers are presented. It is shown that the relaxation theory is insufficient to explain the transition, as observed by mechanical, electrical and thermal measurements. The usual objections which are raised against the assumption of a thermodynamic transition are shown to be invalid. Various molecular processes responsible for a thermodynamic transition are discussed and it is shown that the general mechanism of ``hindered'' rotation (changes in rotational configuration) is the most likely.