8o JOURNAL OF GEOLOGY— SUPPLEMENT 



Though the separation of immiscible liquid fractions has already 

 been fully discussed and reasons given for rejecting it in silicate 

 liquids, a certain phase of the subject may be pointed out briefly in 

 this connection. Some have supposed that individual minerals 

 may separate from the magma in the liquid state with the conse- 

 quent possibility of the formation of a monomineralic rock. Thus 

 the formation of anorthosite has been considered due to the local 

 collection of immiscible plagioclase portions. If plagioclase did 

 separate from the magma as a distinct phase, it would no longer be 

 in solution in anything and therefore must acquire its own individual 

 properties, one of these being a definite temperature or temperature 

 of range of melting. Thus the plagioclase AbiAnz could separate 

 from the magma as a liquid, if at all, only above 1490°, and common 

 olivine only above some such temperature as 1 700°. These temper- 

 atures need only be mentioned in order to prove the impossibility 

 of such a process. Any immiscible liquid phase separating from a 

 magma would necessarily be a solution of minerals capable of 

 existing alone as a liquid at the temperature concerned, for the 

 presence of other liquid as a separate phase affects its properties not 

 at all. 



Winchell has described the collection of crystals for the formation 

 of anorthosite masses in all its stages in the Duluth gabbro of 

 Minnesota,^ and the writer has seen the incipient stages of the 

 same process in neighboring regions of Canada.^ 



There are, moreover, certain features of anorthosites which 

 strongly confirm the conclusion that they are formed simply by the 

 collection of plagioclase crystals. When plagioclase crystals have 

 collected locally in a magma in sufficient concentration to render the 

 rock an anorthosite, the amount of interstitial liquid, which is of 

 mixed composition, must be reduced to a minimum. The move- 

 ment of such a mass involved in its reintrusion into surrounding 

 rocks would necessarily involve much breaking up of the plagioclase 

 crystals. Likewise one would hardly expect a rock so generated to 

 form effusive masses. Corresponding with these expectations we 



' N. H. Winchell, Geol. and Nat. Hist. Survey of Minnesota, Final Rept., V (1900), 

 66. 



^ N. L. Bowen, Ontario Bur. Mines, 20th Ann. Rept., 191 1, p. 127. 



