THE PROBLEM OF THE ANORTHOSITES 241 



sorting of plagioclase crystals anorthosite can form only from very 

 large masses of magma that cool with great slowness, or if from 

 masses of more moderate size these must be deep-seated. The anor- 

 thosite of the large masses normally belongs below the granitic zone 

 so that, whether formed in very large bodies or in bodies of more 

 moderate size, it is an especially deep-seated rock. Peridotites and 

 pyroxenites by reason of the relative ease of sorting of these heavy 

 minerals can form from moderate masses and at moderate depths, 

 and are therefore of widespread occurrence and of general distri- 

 bution in the geologic column though never exposed in large masses. 

 Anorthosites, on the other hand, being essentially deep-seated are 

 exposed only in terranes that have suffered, locally at least, excep- 

 tionally deep erosion, the pre-Cambrian and perhaps early Paleo- 

 zoic. According to the writer's opinion there are probably large 

 masses of peridotite and pyroxenite, but these have not been 

 exposed at all for the same reason that anorthosite is exposed only 

 in the older terranes. These peridotites and pyroxenites are, as it 

 were, the complements of the granites, which in virtue of their low 

 density are so abundantly exposed. Many will, no doubt, consider 

 the opinion that there are large unexpected masses of pyroxenite 

 and peridotite a pure assumption, and it is quite true that some 

 assumption must be involved in the formation of opinion concern- 

 ing inaccessible portions of the earth. Nevertheless, an assumption 

 based on analogy with many completely accessible bodies showing 

 density stratification should surely be given a preference over an 

 assumption, tacit or otherwise, that the kind of rocks exposed in 

 any body extend downward indefinitely, which is based merely on 

 lack of evidence, one way or the other, for that particular body. 

 However this may be, it is certain that anyone who believes that 

 anorthosite is a differentiate of gabbroid magma, as most penol- 

 ogists do, must believe that there is an equivalent amount of 

 pyroxenite somewhere, and if not exposed then presumably in 

 inaccessible regions. At this point the hypothesis of crystal 

 accumulation steps in with a rational explanation of the not infre- 

 quent lack of pyroxenite in anorthosite terranes, very difficult to 

 account for on the doctrine of liquid differentiation. Being an 

 accumulated mass of crystals, pyroxenites usually remain sub- 



