161 



develops an hypothesis of differentiation of volcanic rocks in 

 large part based on the lowering of freezing point accom- 

 panying reduction of pressure due to ejection of magmas. 

 The phenomenon of resorption, common in phenocrysts of 

 hypabyssal and volcanic rocks, has also been attributed by 

 some writers to a reduction in pressure consequent on injec- 

 tion or eruption. There can be little doubt that this lowering, 

 which never exceeds a few degrees per 1,000 atmospheres, < 12 > 

 is enormously overweighed by the decreased fusibility conse- 

 quent on removal of volatile components. < 13 )] 



Owing to the loss of volatile constituents and to a minor 

 degree of changing temperature, equilibrium would be 

 violently disturbed, and the residual magma conditions would 

 become unstable. Some of the components of the fluid pre- 

 viously near or at their freezing point would then become 

 undercooled,( 14 ) and with a magma of aplitic composition the 

 spontaneous crystallization of quartz and felspar would ensue. 



(12) Uniform pressure, of course, is postulated here. The 

 differential effects of uniform and non-uniform pressure are dis- 

 cussed in detail by Johnston and Adams. Amer. Jour. Sci., 35, 

 1913, 205. 



(13) G. Morey: Jour. Amer. Chem. Soc, pt. i., 36, 1914, 

 215. The influence of water at a high temperature on the melting 

 point of silicates is well exemplified in the work of Morey on "New 

 Crystalline Silicates of Potassium and Sodium, their preparation 

 and general properties." The case is instanced of potassium 

 silicate which when dry melts at 1015° C, yet yields in the pre- 

 sence of water in a closed vessel at temperatures of 500° C. to 

 400° C. perfectly fluid (liquid) solutions containing 8-25% water. 

 The results of these experiments at high temperatures and 

 pressures are definite, and in full agreement with the existing 

 physico-chemical theory as applied to solutions at ordinary 

 temperatures and pressures. There can be little doubt, therefore, 

 that the melting depression is dependent on the concentration of 

 the volatile component. 



(14) This undercooling is clearly 

 shown diagrammatically in the 

 temperature-concentration freezing 

 point curve of a binary solution 

 water -silicate, where P = original 

 composition of the magma and Q = 

 changed composition of the magma. 

 P 1 represents the magma of com- 

 position P starting to crystallize, 

 or near the point of crystallization. 

 The point Q 1 represents the temper- 

 ature and composition of the magma 

 which is thus undercooled, with 

 respect to the silicate, although the 

 temperature has but slightly 

 changed. The illustration is, of 

 course, purely diagrammatic, and 

 is not complete for the solution in 

 question (water-silicate). 



