ALKALINE ROCKS 329 



islands, including examples in twelve archipelagos, as well as two iso- 

 lated islands, the alkaline rocks are associated with feldspar basalt or 

 its chemical equivalent, gabbro. With abundant repetition the "Atlantic" 

 and "Pacific" types of lava are both found in the same small island. 

 Here, as elsewhere in the world, the hypothesis that the two suites are 

 differentiation products of primordially distinct magmas meets obvious 

 and apparently fatal difficulties. In the Auckland, Fiji, Hawaiian, Juan 

 Fernandez, Samoan, and Society groups the immense preponderance of 

 the subalkaline basaltic rocks is already clear. 



In fact, the simplest explanation of nearly all the alkaline rocks of the 

 Pacific is that they have been derived from primitive basaltic magma. 

 Cross has recently adopted this view of the trachytic, nephelite-bearing, 

 and melilite-bearing rocks of the Hawaiian Islands. 6 He states that the 

 alkaline rocks in Hawaii and elsewhere are "products of the same general 

 process of differentiation as the other rocks with which they are associated. 

 They are simply the extreme phases of the group now known and are 

 connected by intermediate lavas with rocks assumed by Daly to result 

 from gravitative differentiation." Cross makes these statements without 

 giving any proofs or any suggestion as to how or why a pyroxene andesite 

 was differentiated from the primitive basalt in one place and a phonolite, 

 a melilite basalt, or a soda trachyte in another. 



However, it is right to suppose that the conditions must have been 

 different in the two cases. What, then, is the controlling factor in the 

 generation of an alkaline rock from basaltic magma? Smyth's thesis, 

 that the alkali-rich rocks are the results of concentrating alkalies by mag- 

 matic emanations or "mineralizers," is very helpful. It should specially 

 aid some penologists to emerge from the mysticism induced by that over- 

 worked term "differentiation." Too many authors have considered their 

 intellectual work done when they have concluded that a rock or a rock 

 series is due to differentiation. Seldom do petrographic memoirs contain 

 a word about the mechanism or steps of the magmatic splitting, which 

 the authors of the memoirs affirm to be the origin of the rocks concerned. 

 For the alkali-rich rocks Smyth has shown a more excellent way ; yet his 

 thesis seems to need an important supplement. His essay does not an- 

 swer the question as to why the magmatic gases were specially abundant 

 and hence specially able to segregate the alkalies of an initially sub- 

 alkaline magma. It does not explain the very general association of 

 lime-rich magmas with alkali-rich magmas. It does not explain the 

 abundance of lime minerals in many alkali-rich rocks. It does not ex- 

 plain the generation of feldspathoids in place of the usual feldspars. 



6 W. Cross : Professional Paper 88, U. S. Geol. Survey, 1915, pp. 87 and 90. 



