Dr Anderson's paper (Proc. Roy. Soc. Edin., vol. Ivi, 1936, pp. 128–157) is of considerable interest as a contribution both to geology and to the theory of elasticity. One of his assumptions about fracture, however, has little support, and I am inclined to think that some difficulties that he mentions towards the end would be lessened if it were removed. This assumption is that the complete form of a fracture surface can be inferred directly from the distribution of stress before fracture, intersecting the surfaces of principal stress at some constant angle. It is often to be found in both geological and mechanical literature; but it rests on no theoretical grounds and is, I think, untrue.