CRYSTALLIZATION-DIFFERENTIATION IN MAGMAS 427 



polysilicate molecules into orthosilicate molecules and free silica, 

 and they may be written empirically as follows : 



KAlSijOg ^ KAlSi04+ 2 SiO„ 

 NaAlSi308^NaAlSi04+2SiO.. 



Certain other reactions no doubt occur, such as 



H.O+NaAlSi04^HAlSi04+NaOH, 

 H.O+KAlSi04^HAlSi04+KOH, 



which are simple examples of hydrolysis. That such reactions are 

 going forward in the liquid is attested by the fact that certain of the 

 products are precipitated, the polysilicate molecules as feldspar, 

 the orthosilicate molecules KAlSi04 and HAlSi04 in mica, and 

 free silica as quartz. Evidently certain other products not precipi- 

 tated are stored up in the liquid, and the principal of these is 

 NaAlSi04, the liquid becoming highly alkaline ^nd showing a 

 high concentration of water and other volatile components when 

 finally precipitation has taken place to such an extent that the 

 liquid is a very small proportion of the total mass. The compound 

 NaAlSi04, the distinctive molecule of foyaitic magma, can be 

 precipitated from this liquid only if it is removed from the presence 

 of quartz. In the presence of quartz the reaction is simply reversed, 

 and the polysilicate feldspar is precipitated instead of the ortho- 

 silicate nephelite. It is not impossible that rapakivi structure and 

 certain perthites whose albite lamellae appear to have been formed 

 by reaction may represent the stage of reversal of this reaction. It 

 is not essential that such structures be present, however, for 

 commonly the using up of this liquid would be accomplished 

 merely by continued growth of albite or closely related plagioclase 

 crystals. There does appear, however, to be a particular tendency 

 toward a precipitation of albite later than that of orthoclase 

 (rapakivi-like though not necessarily having the particular rapakivi 

 structure) in granites that are related to nephelite syenites. It 

 should be noted, however, that this highly alkaline and highly aque- 

 ous liquid that is thus formed as the crystalhzation of granite becomes 

 far advanced is a very fugitive material, and that the crystallization 

 - of a large mass of granite must proceed very slowly at this stage. 



