CRYSTALLIZATION-DIFFERENTIATION IN MAGMAS 405 



was well advanced, would form throughout the gabbro, and some of 

 it would have to rise through several thousand feet. Granting 

 that this was possible at a certain stage and complete accumulation 

 of the Kquid took place, what has become then of the further 

 quantity of like liquid that must have separated when the conditions 

 causing its first separation became intensified, that is, when crystal- 

 lization of the gabbro was carried farther ? Unquestionably clots 

 of such material should occur everywhere throughout the gabbro. 

 When it is realized too that experimental studies have shown beyond 

 peradventure that a material rich in alkalic feldspar and free silica 

 is a possible crystallization residuum of a gabbro magma and espe- 

 cially of an olivine gabbro the assumption of immiscibility is seen 

 to be quite gratuitous. This is more especially true for that great 

 number of examples where the acid material occurs in part as 

 granophyric interstices in the gabbro phase, for there it is plainly a 

 crystallization residuum occupying crystallization interstices and 

 not occurring as the clots that one would expect of an immiscible 

 liquid. 



CRYSTALLIZATION AS THE CAUSE OF DISCONTINUOUS VARIATION 



This association of gabbro and granophyre without intermediate 

 types, of which Grout has collected a number of examples from the 

 Hterature, is a feature of igneous rocks well recognized by the writer.^ 

 It is true that in discussing crystallization-differentiation it was 

 shown that a complete series might be formed, extending from 

 gabbro to granite with every intermediate step, but there is no 

 necessity that the series be complete. Only in a very large mass is 

 such a series to be expected, where not only the great size but con- 

 ditions of comparative quiet concurred to produce great freedom 

 in the settling of crystals. In masses of moderate size the differen- 

 tiation, if it can be so called, may be a localized affair occurring 

 about each individual crystal, where the separation of the earliest 

 crystalline material from its mother-liquor is accomplished not by 

 its sinking but by the formation of zones about it. In this case the 

 late differentiate occurs merely as interstitial material. If this 

 zoning, action is combined with only a limited amount of sinking of 



' N. L. Bowen, "The Problem of the Anorthosites, " Jour. Geol., XXV (1917), 227. 



