This paper reviews recent experimental work on phase transitions in rubber—particularly the change from the amorphous to the crystalline form. Several possible explanations are considered to account for the vagueness of the melting point of rubber. Molecular inhomogeneity does not completely explain the phenomenon, and a more satisfactory approach involves a size-distribution of effective kinetic units. It is suggested that the reason for such a variation in size of the kinetic units lies in the mutual interference of growing crystallites. Crystallization of rubber by stretching is discussed, and it is shown that the crystallization process, which takes place mainly in the middle part of the extension, accounts in a reasonable way for the S-shaped nature of the stress-strain curve.