THE PROBLEM OF THE ANORTHOSITES 243 



rock, a rather large proportion of minerals other than plagioclase 

 being necessary before such an occurrence would become possible. 

 Anorthosite should be intimately associated with gabbro, but per- 

 haps as intimately with syenite or granite. Anorthosites should 

 commonly be labradorite rocks rather than bytownite or anorthite 

 rocks. 



A consideration of anorthosites with special reference to the 

 Adirondack and Morin areas gives some reason for believing that 

 anorthosites do show the requisite characters. For the Adirondack 

 area especially, evidence is adduced favoring the possibility that 

 there anorthosite and syenite may still occupy the relative positions 

 in which they were generated by the process outlined, the Adiron- 

 dack complex being interpreted as a sheetlike mass with syenite 

 above and anorthosite below. 



Other monomineralic rocks present essentially the same problem 

 and are restricted in their occurrence in substantially the same 

 manner if we consider especially those that approach most closely 

 to the strictly one-mineral character. All of the monomineralic 

 rocks do occur, however, as dikes and dikelike masses in essentially 

 contemporaneous, congeneric, igneous rocks, a fact that may be 

 interpreted as due to the intrusion of a heterogeneous, partly 

 crystalline mass. 



On the whole the inquiry gives considerable support to the 

 belief that the monomineralic rocks, of which the anorthosites are 

 perhaps the most important representatives, are generated by the 

 process of collection of crystals under the action of gravity. 



