CRYSTALLIZATION-DIFFERENTIATION IN MAGMAS 415 



plagioclases accumulate at a certain place in excess of the amount 

 that could have separated from the liquid at that place, the liquid 

 there surrounding them is incapable of completely making them 

 over into the average plagioclase of the mass.^ It makes no differ- 

 ence how the excess of plagioclase arrived at that place, the laws of 

 physical chemistry will not be suspended in favor of those arriving 

 by the convection route. The fact that those portions of the mass 

 that are particularly rich in plagioclase do not have significantly 

 more basic plagioclase than other parts may therefore be regarded 

 as a very serious objection to accumulation of plagioclase by con- 

 vection and indeed to any process involving the accumulation of 

 plagioclase by a positive active motion on its part relative to the 

 liquid from which it separates. 



The theory of crystal settling, however, does not demand any 

 positive action of plagioclase, at any rate not until a quite late 

 stage of crystallization. Just as settling is not regarded as possible 

 for a crystal that does not yet exist, so it is not regarded as possible 

 for a crystal unless it is heavier than the liquid. Grout is able to 

 refer to a place (p. 33) where I have discussed the effect of the sink- 

 ing of plagioclase, but reference to that page will show that there 

 simple mixtures of albite and anorthite are being discussed. Mani- 

 festly basic plagioclase crystals will sink in a liquid having the 

 composition of more acid plagioclase, but if femic material is 

 dissolved in the liquid the problem is not so simple, and on another 

 page of the same paper (p. 79)^ reference is made to this fact. There 

 the following statement is made: ''It appears, indeed, that plagio- 

 clase crystals may at the earlier stages of crystallization be only 

 very slightly heavier than the liquid at most and possibly even 

 somewhat lighter. " In another paper this suggestion is elaborated 

 and the behavior of plagioclase is discussed at some length.'' 

 It is pointed out that plagioclase "remains practically suspended in 

 the liquid with probably a very slight tendency to rise at first and 

 the whole of the liquid is available for the production of the change 

 of composition that ensues as the temperature falls. Thus, though 



' N. L. Bowen, "The Problem of the Anorthosites, " Jour. Geol., XXV (1917), 212. 

 * These page numbers refer to those in "The Later Stages of the Evolution of the 

 Igneous Rocks," Jour. Geol., XXIII (1915), Supplement. 



3 N. L. Bowen, "The Problem of the Anorthosites," Jour. Geol., XXV (1917), 211. 



