CRYSTALLIZATION-DIFFERENTIATION IN MAGMAS 429 



not proceeded adequately until a late stage in the crystallization of 

 granite, when crystal settling was no longer possible. The only 

 method of separating the liquid at this stage is the method of the 

 squeezing out of residual liquid by the breaking down of the crystal 

 mesh under the action of outside forces. If this be true, alkaline 

 magma can be produced only as a result of the action of outside 

 forces, whereas subalkaline magmas, while they too can be so 

 generated, may nevertheless be generated spontaneously by 

 crystal settling. It is possible then that the alkaline rocks might 

 show a tendency toward a definite relation to earth movements^ not 

 exhibited in the same degree by subalkaline rocks, but the writer is 

 unable to see at present what the details of this connection may be. 

 The succession of igneous rocks deduced from the foregoing 

 considerations — from basic through silicic to alkaline — has been 

 discussed in some detail elsewhere.^ An objection to this has been 

 raised on the basis of the fact that in the Pacific basin there are 

 basic rocks and alkaline rocks and none of the intermediate silicic 

 rocks. ■' It should be noted, however, that these objections are 

 based entirely on the evidence of effusive rocks and the occurrence * 

 of a magma, as an effusive mass depends on certain factors other 

 than its generation within the crust of the earth. Daly has, for 

 example, presented figures that show that granitic material is 

 twenty times more abundant than basaltic as deep-seated masses, 

 whereas basaltic material is fifty times more abundant than granitic 

 as effusive masses. There is therefore something about granitic 

 magma that limits very seriously its occurrence in effusive types, 

 even though it may be abundantly present as subjacent masses. 

 We need not go into the question of what these factors may be, 

 though some readily suggest themselves. The significance of the 

 fact in this connection is that the types that happen to be exposed in 

 an effusive sequence may give but a poor indication of the genesis of 

 their magmas, and it may be taken as a general rule that funda- 

 mental relationships are much more likely to be deducible from 

 plutonic masses. When we turn to the evidence of plutonic masses 



' Such as that suggested by Harker in the classification into Atlantic and Pacific 

 types. 



^ "The Later Stages of the Evolution of the Igneous Rocks,." Jour. GeoL. XXIII 

 (1915), Supplement, pp. 55 ff. 



^ Holmes, Science Progress, XI (1916-17), 68. 



