68 SCIENCE PROGRESS 



authors, however, were at a considerable disadvantage in so 

 far as they were obliged to formulate their views in the absence 

 of adequate experimental data. Fortunately for the future 

 stability of petrological theory, data of the kind most urgently 

 needed have been accumulating in the geophysical laboratories 

 of the Carnegie Institution of Washington. The physical 

 chemistry of the cystallisation of various binary and ternary 

 systems — notably those of the plagioclase felspars, of diopside- 

 forsterite-silica, of anorthite-forsterite-silica, and of diopside- 

 anorthite-albite — has been closely investigated, and the results 

 now available are of the utmost importance. Indeed, they 

 must become the basis of all future discussion. So brilliant is 

 the light which these researches throw on the vexed question 

 of differentiation, that the patient workers to whom they are 

 due deserve not only the gratification that attends success, but 

 also the gratitude of all students of igneous geology. Already 

 they have cleared away some of the mists of speculation. May 

 they be stimulated to penetrate still further the clouds that 

 yet await dispersal ! 



Mr. N. L. Bowen, himself one of the foremost investigators 

 in the Washington laboratories, presents in the paper before 

 us a lucid statement of the experimental facts already ascer- 

 tained ; and proposes, largely on their indubitable authority, 

 a systematic petrogenic theory. As regards the relative im- 

 portance of assimilation of foreign material, and differentiation 

 of a molten magma, as rival or co-operative processes in the 

 genesis of igneous rocks, he stands firmly in undivided en- 

 thusiasm for the latter. The unit of differentiation in a 

 magma may, theoretically, be a molecule, a bleb of immis- 

 cible liquid, or a crystal. Bowen agrees with Harker that the 

 Soret effect is hopelessly inefficient as a cause of concentration, 

 but he then parts company with Harker and supports Becker 

 in the view that diffusion is equally limited in its scope. Liquid 

 immiscibility has been favoured by Daly, but Bowen clearly 

 demonstrates the evidence against it (except possibly between 

 silicates and sulphides). The most powerful factor in promot- 

 ing differentiation he finds in crystallisation. " The two pro- 

 cesses involving relative movement of crystals and liquid — the 

 sinking of crystals and the squeezing out of residual liquid — 

 aid each other in a general way in the production of an arrange- 

 ment of the various differentiates, such that the heaviest lies 



