Abstract
The effects of small concentrations of point defects on the free energy of a crystal of formula AB are discussed. It is concluded that the maximum melting point of the compound does not occur at the composition corresponding exactly to the formula AB. The type of phase diagram arising from this fact is derived, and its practical consequences discussed. Some experimental evidence, derived from the literature, is quoted to support the conclusions of this paper. In an appendix, it is shown that a crystal which is heat-treated in presence of its vapour will not necessarily have the same concentration of defects as a crystal of the same composition grown from the melt at the same temperature.