THE PROBLEM OF THE ANORTHOSITES 213 



position that ensues as the temperature falls. Thus, though the 

 earlier crystals of plagioclase are basic bytownite, they are, in 

 nearly all cases, gradually made over into labradorite by the liquid 

 in which they remain suspended. In the meantime the liquid has 

 suffered impoverishment in ferromagnesian constituents and 

 eventually becomes decisively lighter than the plagioclase crystals. 

 Then and then only, as a rule, does subsidence of plagioclase 

 crystals become an important factor, and masses of anorthosite, 

 anorthosite-gabbro, etc., are formed according to the degree of con- 

 centration of crystals. It is to be noted that this lighter liquid 

 from which the labradorite crystals accumulate is now, of course, 

 no longer gabbroic, but, as a result of removal of femic constituents 

 and plagioclase, it approaches syenitic composition, and with con- 

 tinuation of the process actually attains the composition of syenite 

 or granite. In the ideal case in which the process had free scope 

 the resultant mass would be stratified, and would consist of syenite- 

 granite, anorthosite, and pyroxenite in descending order with, in 

 some cases, peridotites at the base. Of all these the only type that 

 was ever liquid as such would be the syenite-granite, though liquids 

 of every composition intermediate between that of the original 

 gabbro and the syenite would be concerned in the process and might 

 occur as chilled borders or in satellitic bodies. The anorthosites 

 should be intimately related to gabbro, therefore, but as intimately 

 related to syenite also, which might occur as interstitial material of 

 late crystallization in some of the phases. By increase of this 

 interstitial material gradual transition into syenite might occur. 



Intimate field association of anorthosite with gabbro and with 

 syenite is a fact that no one will question. Some have emphasized 

 its relation to gabbro and some that to syenite, but the emphasis 

 is due as much to the personal element and to the kind of exposures 

 in any particular area as to any fundamental difference between the 

 anorthosites of one area and of another. One would expect, to be 

 sure, that the andesine-labradorite phase of an anorthosite mass 

 would be the more intimately related to syenite, and the labradorite- 

 bytownite phase to gabbro. 



Quantitative considerations. — -We can perhaps form a better idea 

 of the quantitative relations involved in the process of collection of 



