424 N. L. BOW EN 



should be polycomponent with respect to its sihcate content alone. 

 A volatile component enters the system merely as a component 

 that would distribute itself between the two separating liquids in 

 accordance with a definite partition coefiicient which is the ratio 

 of its solubility in the two liquids. Even if we strain a point and 

 assume that one of the liquids could be monomineralic it is necessary 

 to imagine that volatile components distribute themselves between 

 the two liquids in such a way that the monominerahc liquid gets 

 many times as much as the polycomponent liquid in order that 

 the former may remain molten at the temperature concerned. 

 In other words, it is necessary to assume that volatile compo- 

 nents are many times more soluble in liquid olivine, say, than 

 in liquid basalt. Apart from the unlikelihood of such a relation 

 the real weakness of the whole case becomes apparent when, in 

 order to explain monomineralic masses of pyroxene, it becomes 

 necessary to assume that volatile components are many times 

 more soluble in liquid pyroxene than in liquid basalt, and in 

 order to explain an anorthositic facies it is necessary to assume 

 correspondingly greater solubility of volatile material in molten 

 plagioclase than in molten basalt. Finally, if it is insisted that 

 this chain of successively dependent phenomena, each funda- 

 mentally unlikely, to say the least, may nevertheless represent 

 the facts of the case we may still turn to monomineralic rocks 

 themselves and see if they give evidence of excessive richness in 

 volatile components* or so-called mineralizers. It was with this 

 reasoning in mind that the anorthosites were examined and found to 

 give no evidence of the presence of mineralizers in excessive amounts.^ 

 Scrutiny of the other monomineralic rocks will, it is believed, yield 

 a like result. The marked serpentinization of peridotites is prob- 

 ably sometimes effected by waters of atmospheric origin (no«t by 

 weathering), and where accomplished by magmatic solutions there 

 is little reason for believing that these solutions belonged to a 

 peridotite liquid rather than to adjacent igneous masses frequently 

 congeneric with the peridotite.^ 



' N. L. Bowen, "The Problem of the Anorthosites," Jour. Geol., XXV 

 (1917), 231. 



^ Benson, Amer. Jour. Sci., XL VI (1918), 710. 



