216 N. L. BOW EN 



diopside (/). If the mass had only 10 per cent interstitial liquid 

 of composition W and 90 per cent crystals slightly more calcic than 

 At^An!, the final rock would consist of 98 per cent AbjA^ and 2 

 per cent diopside (Z). This degree of concentration of plagioclase 

 is ample for the production of nearly all anorthosites, and is more 

 than sufficient for most of them. To understand the formation of 

 anorthosites of extreme purity it is necessary to follow what has 

 been happening to the liquid in the meantime, that is, to that part 

 of the liquid from which crystals have subsided. Instead of becom- 

 ing completely crystalline when the composition M is attained, it 

 continues to change its composition toward 5 and may go even 

 beyond 5, that is, it becomes very rich in albite (haplo-syenitic) 1 

 and the diopside becomes a vanishing quantity. Now it can be 

 considered that locally the interstitial liquid occurring between the 

 plagioclase crystals was of this type, which it might be if the rate of 

 interchange of material did not quite keep up to equilibrium require- 

 ments, a very likely possibility. On complete solidification of such 

 a mass anorthosite of extreme purity would result. This would be 

 more likely to give a rock made up of acid labradorite or even of 

 andesine-labradorite. It is because plagioclase crystals may 

 accumulate, under certain circumstances, in a liquid which is itself 

 nearly pure plagioclase (though very different from the crystals in 

 composition) that we can get plagioclase rocks of such extreme 

 purity. 2 For the case of natural rocks the interstitial liquid is 

 enriched, not merely in albite, but also in orthoclase and to some 

 extent in quartz. In those anorthosites that run very low in 

 bisilicates we therefore commonly find 5 per cent or more ortho- 

 clase, and occasionally some quartz. 



In the foregoing the writer has done his best to picture a process 

 whereby plagioclase crystals may accumulate in sufficient force to 

 give a mass of anorthosite. Unquestionably there are some 

 difficulties, the gravest being that connected with the nearly com- 

 plete sorting of plagioclase and pyroxene, whose periods of crystal- 



1 N. L. Bowen, "The Crystallization of Haplobasaltic, Haplodioritic and Related 

 Magmas," Am. Jour. Sci. (4), XL (1915), 161. 



2 Another possible method of obtaining extreme monomineralic composition is 

 suggested later (p. 238). 



