EARTH'S CRUST — WASHINGTON. 305 



It has been suggested by several prominent penologists ss that the 

 comagmatic regions may be referred genetically to two large types of 

 magma or " provinces," called " alkaline " and " subalkaline " by 

 Iddings, or "Atlantic " and " Pacific " by Becke. The latter goes so 

 far as to attempt to ascribe all the comagmatic regions to two areas, 

 the one dominantly alkalic and surrounding the Atlantic Ocean, and 

 the other more calcic and surrounding the Pacific. Harker, further- 

 more, would connect these two main types of comagmatic region with 

 two main types of crustal movement or stress, such as are recognized 

 by Suess, which give rise to different types of coast, mountain forma- 

 tion, etc. In the opinion of the writer such recognition of but two 

 types is not consonant with what we know of the general distribution 

 of the igneous rocks. The whole subject is very complex, far too much 

 so for proper discussion here, and the data available seem to the writer 

 to be inadequate for very broad generalizations at present. 



CHEMICAL COMPOSITION AND ROCK DENSITIES. 



The above outline of comagmatic regions leads us to the considera- 

 tion of two subjects with which we may close this sketch of the chem- 

 istry of the earth's crust ; that is, the relation between the chemical 

 composition of rocks and their density, and that between these and the 

 theory of isostasy. 



In the preceding pages we have considered igneous rocks almost 

 only from the chemical point of view. As we know, however, they are 

 actually aggregates of definite chemical compounds, minerals, mostly 

 silicates. Furthermore, we know that magmas of the same general 

 chemical composition may crystallize as diverse aggregates of differ- 

 ent minerals, according to the conditions that obtain during solidifica- 

 tion. If we know the chemical compositions of the various rock- form- 

 ing minerals, the quantitative mineral composition may be readily 

 calculated from the chemical analysis of the rock. But from what has 

 just been said, it is evident that the particular mineral aggregate to be 

 calculated will depend on the conditions controlling solidification. 

 It is also obvious that if we know the mineral composition and the 

 densities (specific gravities) of the minerals, that of the rock as a 

 whole may be readily calculated. 



In the conception and elaboration of a system of classification of 

 igneous rocks that was proposed some years ago by some American 

 petrologists, the chemical composition of igneous rocks was regarded 

 as their most fundamental character, and therefore that on which their 

 classification was primarily based. But, in order to recognize the fact 



83 For some general discussion of this and related topics, the reader is referred to Har- 

 ker, The Natural History of Igneous Rocks ; Iddings, Igneous Rocks, Vol. I ; and Daly, Ig- 

 neous Rocks and Their Origin. 



42803°— 22 20 



