LATER STAGES OF EVOLUTION OF IGNEOUS ROCKS 8i 



note the common protoclastic structure of anorthosite, and the 

 non-occurrence of an effusive equivalent. It may be noted, more- 

 over, that some of the features of monominerahc rocks pointed out 

 by Vogt,^ the enrichment of peridotites in magnesia, of anorthosite 

 in Hme, and so forth, are expHcable only on the basis of their forma- 

 tion by the collection of early crystals. 



THE ORDER OF INTRUSION 



If the change of composition of the liquid is as shown in the 

 scheme presented under I, p. 75, the tapping of the batholithic reser- 

 voir at successive stages will usually realize "the normal order of 

 decreasing basicity " in the order of intrusion. Quite commonly the 

 sequence observed in batholiths themselves is simply a sequence of 

 consolidation, differentiation having taken place practically in situ. 

 It is, however, not to be expected that this order will be universal. 

 If the hypabyssal rocks of a given area were fed from two adjacent 

 batholithic reservoirs which at any given time were at different 

 stages in their career of crystallization, it is easy to see that the 

 exposed rocks might exhibit no system in the matter of change of 

 composition in successive intrusions. 



The upward intrusion of the material in a single reservoir may 

 also give results which depart from the "normal order." If the 

 magma is forced upward at a time when the upper part is largely 

 liquid and the lower part contains a large proportion of the heavier 

 early minerals with sufl&cient liquid to render it eruptible, the more 

 "basic" type should follow after the more salic type. The case of 

 Mount Johnson where essexite follows pulaskite is apparently an 

 example of this phenomenon.^ It is certain that the predominance 

 of the order of decreasing basicity which has led to its designation 

 as the normal order is to be expected if crystallization as discussed 

 in this paper is the controlling factor in differentiation. Brogger's 

 conclusion that the Diferentiationsfolge is parallel to the Kristal- 

 lisationsfolge is correct, the reason being simply that differentiation 

 is the result of crystallization. That the same is true of the 

 Eruptionsfolge is commonly, though not necessarily, true also.^ 



^ Op. cit., pp. 19-49; also summary statement of same by Harker,o^.a7., pp. 372-74. 



2 F. D. Adams, Jour. Geol. XI (1903), 281. 



3 Eruptivgesteine des Kristianiagebietes, II (1895), 175. 



