THE PROBLEM OF THE ANORTHOSITES 217 



lization are in large part contemporaneous. If one could assume 

 that plagioclase follows pyroxene in the crystallization of gabbro, as 

 some geologists appear to do, the sorting would be a simple matter, 

 but chemical considerations will not permit such an assumption. 

 Yet the difficulties do not seem insurmountable, especially in com- 

 parison with those connected with other processes. Diffusion is 

 hopelessly incompetent even if it is assumed that its tendency is in 

 the proper direction. Liquid immiscibility, whose operation in the 

 case of silicates has nothing to support it, would certainly not tend 

 to produce pure liquids in any case, but only to produce liquids of 

 contrasted composition, all being, nevertheless, mutual solutions of 

 minerals. Added to these is the temperature objection to which 

 reference was made on an earlier page, and still others might be 

 mentioned. On the other hand, it does seem reasonably probable 

 that a mass of gabbroid magma might cool . sufficiently slowly to 

 permit the necessary amount of sorting of crystals especially if it 

 was a large mass, or if it was very deeply buried. 



CHARACTERISTICS OF ANORTHOSITE CONSEQUENT UPON THE 

 SUPPOSED METHOD OP FORMATION 



It must be admitted, however, that opinion as to whether the 

 process can take place is not a very decisive matter. More impor- 

 tant is the deduction of its consequences followed by a survey of 

 the characteristics of anorthosites in order to determine to what 

 extent they agree with the requirements. If anorthosites are 

 generated only by the accumulation of crystals, then the more 

 nearly a rock mass approaches an exclusively plagioclase composi- 

 tion, the more nearly it should have approached the completely 

 solid condition when that composition was attained. In discussing 

 artificial melts we have seen that if we have a portion with 80 

 per cent plagioclase crystals and 20 per cent interstitial liquid it 

 would, on crystallization, have 95 per cent plagioclase and 5 per cent 

 diopside. In other words, a rock containing only 5 per cent diopside 

 could have had, after that total composition had been attained by 

 the process we are considering, not more than 20 per cent liquid. 

 A rock containing 10 per cent diopside could have had a maximum 

 of 35 per cent liquid, and one containing only 2 per cent diopside 



