It was recalled that a recent paper concerning a model A-B alloy which contained a high vacancy concentration, had predicted thermodynamic factors for diffusion of the 2 components. These had differed from one another, although usual treatments of chemical diffusion were expected to predict them to be equal. Here, the reasons for the difference were analyzed from 2 points of view. One was purely thermodynamic, while the other was statistical and recognized the presence of vacancies. A relationship between the thermodynamic factors for A and B was derived within the statistical framework that had been used by the paper in question. In the special case of a symmetrical model, as used in the latter paper, this relationship reflected the reciprocity that was visible in the reported numerical results.

A.B.Lidiard: Philosophical Magazine Letters, 1996, 74[1], 43-9