THE PROBLEM OF THE ANORTHOSITES 215 



of the liquid changes along the boundary curve DE toward D. 

 Both diopside and plagioclase continue to separate, and the plagio- 

 clase crystals, not only those separating at any instant, but also 

 those which had formerly separated, continually change in com- 

 position, becoming richer in albite. At 1220 the whole mass is 

 made up of 37 per cent diopside crystals, 25 per cent labradorite 

 crystals of composition L (Ab x An 2 ), and 38 per cent liquid of 

 composition K. As the temperature falls still lower the liquid 

 gradually decreases in amount and continually changes in compo- 

 sition until at 1 200 it is all used up, the last minute quantity having 

 the composition M. In the meantime diopside and plagioclase 

 crystals have been separating, and the plagioclase has been changing 

 continuously in composition until at 1200 , when the last of the 

 liquid disappears, the composition of the feldspar is A^An^ The 

 whole mass now consists of 50 per cent diopside and 50 per cent 

 AbxAni. 



Crystallization takes place according to the foregoing outline if 

 no sinking of crystals occurs. If diopside crystals sink, no effect on 

 the composition of the liquid results. We should have then at 

 1220 a mass in which the diopside crystals were of increasing con- 

 centration toward the bottom, and in which a certain upper portion 

 was free from diopside crystals, consisting of 60 per cent of liquid 

 of composition K and 40 per cent labradorite crystals of composi- 

 tion L (Ab z An 2 ). Let us imagine that at this stage appreciable 

 sinking of plagioclase crystals begins, and that it increases in 

 importance as the liquid changes toward M, and therefore becomes 

 lighter. It is necessary to imagine also that the plagioclase crystals 

 sink very slowly, and are outstripped by the heavy diopside crystals 

 which are forming simultaneously and which increase in size more 

 rapidly since the liquid is becoming relatively impoverished in 

 diopside. It seems possible, then, that, locally at least, plagioclase 

 crystals might accumulate in a mass free from diopside crystals and 

 containing only a little interstitial liquid whose composition would 

 lie between K and M. If the mass had 20 per cent interstitial 

 liquid of composition V and 80 per cent crystals slightly more calcic 

 than AbjAn^ the final rock formed on solidification of the interstitial 

 liquid would consist of 95 per cent A^Am and 5 per cent 



