Theoretical calculations suggest that the duration (10−11sec) and smallest dimension (20 atom diameters) of thermal spikes in metals undergoing irradiation may be too small to permit the thermal nucleation and growth of crystals of a second phase even though the average ``temperature'' of the spike may be above the melting point. Experiments have been performed which show that even the very energetic fission thermal spikes are too rapid and small to permit the nucleation and growth of a second phase in uranium metal containing 2 atomic % chromium. It is suggested that the terms ``heated'' and ``melted'' are not strictly appropriate to thermal spikes. Rather the region of a thermal spike should be regarded as a highly disturbed zone in which atomic interchange can occur, but which quickly returns to the structure and orientation of the surrounding crystal with a distribution of interstitials and vacancies which may then interact with each other or with other defects possibly present in the crystal.