SIGNIFICANCE OF THE VARIABLE COMPOSITION 413 



and there concentrated. Many of these portions, rich in dark constitu- 

 ents, show no evidence whatever of having been intruded or in any way 

 thrust into the positions they now occupy, as Bowen suggests, but rather 

 they appear to be differentiates practically in situ. In any case there 

 must have been enough liquid in much of the rock to have allowed mag- 

 matic differentiation and flowage, and the liquid in the foliated zones of 

 pure plagioclase rock must have been molten plagioclase. 



I do not wish to imply, however, that the anorthosite as such was neces- 

 sarily intruded in the form of a true magma to its present position, hav- 

 ing been differentiated at a much lower level. Eather, it is probable that 

 a gabbroid magma was the original intrusive which, either during the 

 process of intrusion or after the magma came nearly to rest, or both, 

 differentiated to give rise to the anorthosite which was then, at least in 

 very considerable part, really molten. This matter is taken up more fully 

 beyond under the caption, "Origin of the anorthosite by differentiation 

 in a laccolith of gabbroid magma." 



Though I believe the anorthosite as such to have been molten to a very 

 considerable degree at least, it is by no means necessary to assume that it 

 was ever completely molten with a high degree of fluidity, or even only a 

 moderate degree of viscosity. The field facts best harmonize with the con- 

 ception that the typical anorthosite as such may never have been com- 

 pletely molten, but that crystals of plagioclase originating in a gabbroid 

 magma gradually developed, while the crystallizing femic minerals tended 

 to separate themselves either by concentration locally within the general 

 mass or by settling toward the bottom. Accordingly, as the anorthosite 

 developed, it contained a decreasing amount of liquid, and this steadily 

 became more viscous. On the one hand, then, it is quite possible, or even 

 probable, as Bowen argues, that those portions of the great anorthosite 

 body which are nearly pure plagioclase may never have been completely 

 molten as such; but I maintain that even such portions must have con- 

 tained at least considerable quantities of interstitial liquid. On the other 

 hand, none of the field facts necessarily preclude the hypothesis that the 

 whole mass of anorthosite may once have been completely molten, though 

 not with a high degree of fluidity. 



Before leaving this consideration of the signiflcance of the variable 

 composition of the anorthosite, emphasis should be placed on the fact that 

 in many places its mass shows unmistakable evidence of having differen- 

 tiation phases of anorthosite-gabbro, or even gabbro, while there is no 

 positive evidence for differentiation phases of syenite or granite, as should 

 be the case, according to Bowen's hypothesis. Gushing-^ says : "The gen- 



H. p. Gushing: Jour. Geol., vol. 25, 1917, p. 505. 



