LATER STAGES OF EVOLUTION OF IGNEOUS ROCKS 31 



total composition (liquid and crystals) and at the same time 

 would for a considerable period contain nearly or quite 50 per cent 

 liquid. With such a proportion of liquid it would be eruptible as a 

 whole and might be injected as dykes of peridotite into adjacent 

 rocks, if we imagine these simpKhed magmas occurring in the crust 

 of the earth. The eruptibility of this mixture would be increased 

 by the fact that the injection must take place from a position of 

 higher to one of lower pressure and, as a result of the lowering of 

 pressure, re-solution of some crystals would take place. It appears 

 likely, however, that this re-solution must be considered on the 

 whole of relatively small importance and merely as aiding somewhat 

 in the eruption of parts of a magma enriched in sunken crystals. 

 We know little about the volume-changes and heat-effects involved 

 in the solution of silicates in silicate Hquids, but such indications 

 as are to be gained from the same factors for the melting of single 

 siHcates suggest that the effect of rehef of pressure in bringing 

 about re-solution would be moderate. The great importance of 

 this process assumed by Schweig^ is unlikely. It appears also to be 

 unnecessary to make this assumption in order to explain the facts 

 for which it was proposed. 



An increase in the eruptibiHty of a part enriched in sunken 

 crystals through re-solution of the crystals in the hotter liquid 

 appears to be also of limited importance. The factors limiting this 

 action will be discussed on a later page. 



CRYSTALLIZATION IN SYSTEMS INVOLVING THE PLAGIOCLASES 



In natural rocks the plagioclases are a very important mix- 

 crystal series. We shall now consider a system related to both 

 those already discussed, involving the single pyroxene diopside, 

 and the plagioclase mix-crystal series instead of the single plagio- 

 clase anorthite.^ 



CRYSTALLIZATION OF THE PLAGIOCLASES IN THE BINARY SYSTEM 



Experimental work has confirmed the opinion that the plagio- 

 clases belong to Type I of Roozeboom's classification of mix-crystal 



^ Neues Jahrh., Beil. Band XVII (1903). S(>2,- 



2 N. L. Bowen, "The Crystallization of Haplobasaltic, Haplodioritic and Related 

 Magmas," Am, Jour. Sci. (4), XL (1915), 161. 



