AKT. 2. PETROLOGY AT GOOSE CREEK SHANNON. 35 



phyre. He takes it as the starting point for a discussion of the 

 physical chemistry of sihcates with the following introduction: 



Diabases with micropegmatite interstices are ver\' common. Sometimes the 

 micropegmatite (granophyre) is separated as a distinct body, a granite, grano- 

 diorite, or quartz diorite in composition. This association is of fundamental 

 importance in petrogenic theory and will be made the starting point for a dis- 

 cussion of the geologic evidence supporting crystallization differentiation. It is, 

 in man}' cases, clearl}' shown that when the diabasic (basaltic) magma was 

 intruded as a small body and was therefore quickly chilled, it crystallized as a 

 normal plagioclase-pyroxene diabase without quartz. On the other hand, large 

 bodies usually show micropegmatitic interstices and often a similar salic dif- 

 ferentiate. This contrast between the larger and smaller bodies has led some 

 petrologists to the opinion that the more slowly cooled, large bodies has an 

 opportunity denied the quickly cooled bodies — the opportunity to assimilate 

 siliceous material, whence the siliceous differentiate. Direct evidence of ade- 

 quate assimilation is seldom if ever clear; its accomplishment is nearly always 

 inferred from the existence of the acid differentiate. * * * 



Following a discussion of the equilibrium relations of the several 

 rock forming silicates, in the light of the results obtained from the 

 investigations of various systems, the following conclusions are at- 

 tained: 



Crysiallizaiion with zoning. — When the cooling is too rapid to give crystalliza- 

 tion of the perfect equilibrium type and yet not rapid enough to give the great 

 degree of undercooling referred to in the foregoing, the formation of zoned cr3's- 

 tals of plagioclase will result. According as one or the other of the above- 

 named rates of cooling is approached the degree of zoning is reduced to a mini- 

 mum. With a certain intermediate rate of cooling maximal ^ioning results. In 

 this case, a crystal once separated suffers thereafter no change of composition, 

 the liquid disregarding crystals which have already formed, so that the crystal- 

 lization of the liquid may be regarded as beginning anew each instant. 



The effect of this action may be reahzed by considering that during the crys- 

 tallization of the liquid F, as already outlined, the liquid portion is separated 

 from the crystalline portion at a temperature of, say, 1,220°. At tlais tempera- 

 ture the liquid has the composition K and we shall imagine that this separated 

 liquid is crystallized under perfect equilibrium conditions. Instead of becoming 

 completely crystalline at 1,200°, as it would if the crystals had not been re- 

 moved, it now becomes completely crystalline only at 1,178°, and the final 

 liquid, instead of the composition M, has the composition S; i. e., is much 

 richer in albite. If the virtual separation of liquid from crystals is a continuous 

 process accomplished through the intervention of zoning, it is plain that the off- 

 setting in the composition of the final liquid is limited only by the eutectic al- 

 bite-diopside which it actually attains in the case of maximal zoning. This 

 fact is true, not only of the special liquids to which reference has been made, but 

 of any mixture of anorthite, albite, and diopside whatsoever. 



* * * The sinking of crystals of jjlagioclase in a mass of hquid which is 

 very slowly cooled will obviously affect the upper layers from which the 

 crystals have settled in the same manner that zoning affects the residual 

 liquid. * * * 



* * * When the hquid is very quickly cooled it crystallizes quickly, if at 

 all, and with little or no tendency to an offsetting in the composition of tiie 

 liquid. If it is cooled moderately slowly, zoning of the plagioclase causes a 



