Page:The Journal of geology (IA journalofgeology11893univers).pdf/357

minerals—are naturally more than two, and therefore the liquation must become very complicated, being not only a function of temperature but also dependent on the original proportions. Therefore, in other places, where the original magma had another composition, relatively stable andesitic magmas might be formed, but this was evidently not the case in Iceland.

Liquation is no doubt also a function of the pressure, but experimental data are wanting. Still it may be considered as probable that, if liquation would augment the volume of the magma, then pressure would act the same as increase in temperature, and inversely. The first is most frequently the case with liquid-mixtures.

The purpose of this communication is to give to liquation and not to diffusion its place as the working hypothesis, upon which the theory of differentiation is to be constructed. How far this theory may differ from the approximation to it, given by Rosenbusch in his "Kern"-theory, the future will show.

In conclusion, I wish to express my best thanks to my friend and colleague Dr. S. Arrhenius for much valuable information furnished me in numerous discussions on this and other subjects which lie on the border between petrology and physical chemistry.