﻿Vol. 66.] CLASSIFICATION OF IGNEOUS KOCKS. 477 



A large proportion of the dominant alkalicalcic rocks of California 

 are very rich in soda as compared with potash — a much larger pro- 

 portion than among the rocks of the Rocky Mountains. The Pacific 

 coastal zone appears to be a soda-rich province, as compared with 

 the Rocky Mountain portion of the United States. 



Another distinction which seems notable from a study of Wash- 

 ington's tables is that on the Pacific coast high silica goes with 

 high soda, so that foyaitic magmas, nepheline-rocks, are rare, 

 especially among persalanes and dosalanes. In the salfemane 

 class there are some nepheline-bearing rocks, as illustrated by three 

 cases of the foregoing table. These belong, of course, to the Atlantic 

 kindred, with the essexite of Salem ISeck, which is included in the 

 table for purposes of comparison. 



These facts regarding the rocks of the Pacific zone 

 of North America indicate, I think, that they possess 

 provincial peculiarities of interest; but these are not by 

 any means identical with the features emphasized by 

 Becke and others as characterizing the Pacific kindred. 



The dominance of soda over potash is also very strong in all 

 rocks from the Hawaiian Islands thus far analysed ; and several 

 of these belong to the * Atlantic kindred ' of Becke, including 

 trachyte, 1 melilite-neph elite basalt, and others usually but not well 

 called basalt, such as the one from the crater of Haleakala, collected 

 by myself, which has in part the composition given in the fore- 

 going table. This rock contains but 2*49 per cent, of magnesia. 

 It is evidently very uear the typical essexite in chemical character, 

 for which reason I apply to it the name ' essexitic andesite ' in the 

 table (=trachy dole-rite). 



The broad underlying significance of these facts of geographical 

 distribution postulated by Harker, Becke, and independently by 

 Prior, 2 has been touched upon by these authors. Becke points out 

 that the Pacific kindred of young volcanic rocks seem to be con- 

 nected with mountain-chains representing folding under tangential 

 stress, according to Suess, while the Atlantic kindred occur on 

 lines of fracture through radial contraction. Whether this is 

 true or not (and it certainly needs confirmation and elucidation) 

 the relationship involved is not that responsible for the 

 chemical differences of the magmatic series ; nor does it 

 appear that the generalization of distribution applies 

 to the older rocks. Thus in the Great Rift Valley of East 

 Africa, the rocks of which have been so well described by Prior, 

 the younger foyaitic-theralitic series of magmas of the Atlantic 

 kindred are underlain by old crystalline granites, gabbros, para- 

 gneisses and other rocks, which from Prior's descriptions seem to 

 belong to the Pacific kindred. The same relations probably exist 

 in the Bohemian Mittelgebirge, the volcanic rocks of which are 



1 W. Cross, ' An Occurrence of Trachyte on the Island of Hawaii ' Journ. 

 Geol. Chicago, vol. xii (1904) pp. 510-23* 



2 'Contributions to the Petrology of British East Africa' Min. Mag. vol. xiii 

 (1901-03) pp. 228-63. 



