420 REGINALD A. DALY 



the chemical nature of an individual and typical olivine basalt, its 

 phenocrysts and its base ; by the fortunate occurrence of simultaneous 

 eruptions from the crater and low-lying lateral fissure at the Reunion 

 volcano where Velain shows the differentiation has taken place ideally; 

 by the abundant proofs that heavy crystals do and must slowly sink and 

 light crystals slowly rise in lavas, and in rock-melts; by the well-known 

 facts of field-association and bodily development of augite andesite, 

 olivine basalt and ultra-basic rocks throughout the different continents. 

 If the hypothesis correctly represents the facts and the augite 

 andesites of the sea-floor volcanoes and of so many continental 

 volcanoes are truly derivatives of olivine basalt, we have one more 

 important link in the chain of argument leading to the belief that 

 basaltic magma forms the universal substratum of the earth's crust 

 today and has formed that substratum since Keewatin (early Archean) 

 times. With this conception as a working theory the writer also con- 

 nects the view that the crust overlying this basaltic substratum is 

 stratified by density, so that the lower layer of the crust is crystalHzed 

 basaltic magma (gabbro, possibly merging upward into anorthosite) and 

 the upper layer is a composite of dominantly acid material. This latter 

 is considered as made up largely of original granite or gneissic rock, 

 similar to the staple fundamental gneiss of the pre-Cambrian — a layer 

 probably less than thirty kilometers thick. This layer was crystallized 

 in pre-Keewatin time; through it the basic Keewatin lavas were 

 erupted ; and through it basaltic magma has, from place to place and 

 from time to time, ever since been erupted. The universal basaltic 

 layer has thus been the effective source of the heat involved in the 

 eruption of post-Keewatin igneous magmas. By the spontaneous 

 differentiation of the primary basalt through fractional crystallization, 

 the few rock-types specially discussed in this paper, have been locally 

 derived. Most of the other eruptive rocks are, on this same working 

 hypothesis, regarded as derived from the formation and differentiation 

 of magmas which are the product of the solution of the acid, original 

 gneissic shell and of its sedimentary veneer in the primary basalt. In 

 other words, both the "syntectic"^ (assimilation) theory and the frac- 

 tional-crystallization theory seem to the writer to be essential and 

 principal elements in the final solution of the genetic problem of the 

 igneous rocks. 



I F. Loewinson-Lessing, Congres geol. internal. Compte rendu, 7th session, St. 

 Petersburg, 1897, P- 375- 



