LATER STAGES OF EVOLUTION OF IGNEOUS ROCKS 59 



displace silica from its combinations. Thus is brought about great 

 concentration of the molecules which go to form biotite at the 

 stage of the normal granite, and biotite, together with quartz, is 

 precipitated. From similar causes results the precipitation of the 

 olivine fayalite in the alkaline granites, the felspathoids, and even 

 free alumina in foyaitic rocks. If the question is attacked from 

 this point of view, that is, if the various minerals precipitated at 

 any stage are recognized as those least soluble (in proportion to 

 their abundance) at that stage, if the various equilibrium reactions 

 indicated by these minerals are sought out, and if the effect on these 

 reactions of the removal of minerals is considered, it seems likely 

 that a good understanding of the alkaline rocks might be arrived 

 at from field studies with a certain amount of guidance afforded 

 by experimental work. 



Discussion of the question whether alkalic and sub-alkalic 

 rocks may be derived from the same magma has not been included 

 in the foregoing because it is believed that field facts leave no 

 possible doubt in the matter.' Crystallization proceeds, appar- 

 ently, with the precipitation of the constituents of normal granite, 

 then of those of the alkaline granites, through those of the alkaline 

 syenites, to those of the nephelite syenites, with no sharp lines of 

 demarkation anywhere in the series. It is to be remembered that 

 the series will be maintained only when the continual removal of 

 crystals from that part of the liquid in which they formed is like- 

 wise maintained. 



The genetic relation and continual passage from sub-alkaline 

 to alkaline types is now recognized as true by most petrologists and 

 is, in fact, well illustrated by the difficulty encountered by every 

 petrologist in setting up a boundary between them. There appears 

 to be, however, a peculiar subconscious persistence of the older idea 

 of a genetic separation even in the minds of many who admit a 

 genetic connection. This tendency is apparent especially in 

 nomenclature. A perfectly normal diabasic or gabbroic rock, for 

 example, is likely to be termed an essexite if it occurs in a region 

 of dominantly alkaline rocks and especially if associated with true 



^ See H. S. Washington, "The Volcanic Cycles in Sardinia," Compte-rendu, XII™^ 

 session, Congres Geologique International, p. 235. 



