The present work is a continuation of that published in ‘Phil. Trans., Royal Society,’ vol. 214, pp. 109-146, 1914 (Parts I. and II.) and vol. 215, pp. 79-103, 1915 (Part III.). It will lead to clearness in the following development of the subject if a brief résumé of these papers is given. A t the same time, I wish to discuss one or two points in connexion with the views which have been previously advanced and the relation between mechanical and molecular theory. In Part I. the experimental evidence brought forward has justified the hypothesis of molecular distortion enunciated at the outset. We have thereby been led to regard the molecular configuration of a material medium as a distorted one, and this applies particularly to a substance which is crystalline. The extent of this distortion is small, but is sufficient to account for the observed change of specific susceptibility which occurs on crystallization. Such change will naturally depend upon the particular crystalline symmetry assumed by the substance.