Much of what we think we know about crystallization in polymers is obtained from studies carried out under isothermal conditions. From extrapolations of equilibrium melting temperature to crystallization regimes, interpretations in the literature all assume that the measured temperature of a polymer film is the temperature at which the crystallization occurred. Recently it has been possible to study crystallization at very high supercoolings, even in rapidly crystallizing polymers, through the use of a rapid cooling technique in which a thin film contains a microthermocouple to measure actual temperature. In adequately thin films an isothermal temperature is set up through a balance between the heat of fusion released and the cooling rate applied. This technique has been applied successfully to polypropylenes, and to polyethylene homopolymers and copolymers. The results have told us much about crystallization at very high supercoolings. Recently, a study of the effect of nucleating agents on quench-crystallization in PET and in nylon 6 was conducted, in which it was discovered that films containing nucleating agents generated the expected temperature plateau when crystallization occurred. In this paper, we wish to report the results of studies of thin films of nylon 66 and PET which did not contain nucleating agents. The results are totally unexpected and indicate that a steady state condition exists at the growth face of the polymer, which is not reflected in the macroscopic temperature of the thin film, as measured using a microthermocouple. In other words the measured temperature of a film is not the temperature at which crystallization occurs. The consequences of this finding to our general understanding of crystallization in polymers will be discussed. (c) 2005 Elsevier Ltd. All rights reserved.