426 N. L. BOWEN 



at later stages of crystallization, and in this case, while the disrupted 

 mass would not give monomineralic inclusions, it might readily 

 give inclusions strongly contrasted with the magma that would pro- 

 duce schliers and related irregularities in the rocks formed. 



In the banded peridotites of Skye, Harker describes seams of 

 nearly pure picotite and chromite. It is possible that these may 

 have been formed by some such action as that described above. 

 However this may be, the petrologist should look askance at molten 

 spinel and chromite. Molten corundum, molten olivine, and 

 molten magnetite even with volatile components in solution are to 

 be similarly regarded, and every possibility of explaining masses of 

 such minerals as the result of crystal accumulation from complex 

 magmas or of hydrothermal action should be exhausted before the 

 probability of such melts is entertained, for the chemical difficulties 

 in the way of their production are colossal. By hydrothermal 

 action is meant introduction by dilute solutions as contrasted with 

 introduction as a melt, even though it be a melt with volatile 

 material dissolved. 



ORIGIN OF THE ALKALINE ROCKS 



Throughout the foregoing pages consideration has been given 

 to phenomena exhibited only by the so-called subalkaline rocks. 

 For such rocks we can speak with considerable assurance regarding 

 the chemical effects of relative movement of crystals and liquid, 

 because systems approaching them reasonably closely in com- 

 position have been completely investigated. Concerning the alka- 

 line rocks we cannot speak with the same assurance, for not much 

 more than a beginning has been made in their experimental 

 investigation. Nevertheless some consideration will be given to 

 the alkaline rocks, particularly to the question of the relationship 

 between alkaline and subalkaline magmas. The writer has stated 

 elsewhere the opinion that the most important of the alkaline 

 magmas, foyaitic magma, is definitely related to granitic magma. 

 The equilibrium reactions which take place in granitic magma, and 

 of which the evidence is plain from the occurrence side by side of 

 feldspars, mica, and quartz, furnish a clue to the origin of foyaitic 

 magma. The principal reactions indicated are the breakdown of 



