88 R. A. DALY ORIGIN OF THE ALKALINE ROCKS 



other hand, liparite and dacite, with average alkalies of, respectively, H.Q 

 and 6.6 per cent, are placed in the snbalkaline group because of their com- 

 mon independence of the nephelite syenites, phonolites, leucitic rocks, 

 etcetera, in field occurrence. No single principle compels the separation 

 of the two groups, and the difficulty of making it sharp is enhanced by the 

 fact that there are plenty of rock species transitional between ideal alka- 

 line types and standard types of the subalkaline group. Nevertheless, the 

 separation so far actually made by Eosenbusch, Iddings, Loewinson-Les- 

 sing, Harker, and many others is about as definite as some of the divisions 

 among the subalkaline rocks. For this paper the question as to how far 

 the distinction is objective and justifiable need not be answered, though it 

 will be seen that there are grounds for believing in a very close genetic 

 connection between the two groups. 



General Statement of the Investigation 



This major classification of igneous rocks immediately raises the ques- 

 tion of origins. Why are the alkalies concentrated in certain types of 

 igneous rocks? Why are many alkaline rocks so low in silica that feld- 

 spathoids form instead of feldspar ? In many petrographic memoirs their 

 authors have assumed that each alkaline comagmatic region is underlain 

 by one or more reservoirs of magma which was rich in alkalies from, "the 

 foundation of the world." The visible eruptives are regarded as differen- 

 tiates from such primeval segregations. A few writers have suggested 

 the derivation of alkaline types from subalkaline magma by unusual 

 processes of differentiation. No one has yet stated the conditions for this 

 special differentiation. The succeeding pages bear an outline of a hy- 

 pothesis intended in some degree to declare those conditions. 



It will first be pointed out that no alkaline province can be described as 

 free from subalkaline eruptives, especially those of basaltic or granitic 

 types. Emphasis will be laid on the indisputable fact that the visible 

 volume of all alkaline rock bodies is a very minute quantity as compared 

 with the visible volume of subalkaline eruptive bodies. An inductive 

 study shows that most alkaline rocks cut thick masses of limestones, dolo- 

 mites, or other calcareous sediments. This fact suggests the hypothesis 

 that the absorption of carbonate disturbs the chemical equilibrium of sub- 

 alkaline magma in such manner that alkaline fractions are produced by 

 differentiation. Most of the alkaline species are ascribed to the inter- 

 action of basaltic magma and limestone, but more acid magma is also 

 sensitive to the solution of carbonate. The hypothesis explains the con- 

 centration of alkalies; the desilication shown by the crystallization of 



