4i6 REGINALD A. DALY 



dotitic, limburgitic or picritic magma resulting from the remelting of 

 the settled phenocrysts, would form at the base of the lava-column and 

 would be considerably smaller in volume than the correlative andesite ; 

 fourthly, that the heavy, ultra-basic magma would oftener form intru- 

 sive than extrusive rocks, and would fill fissures either in the basal 

 beds of the cone itself or in the older formations on which the cone 

 rests ; fifthly, that the primary basaltic magma is directly represented 

 on the earth's surface where the lava was erupted rapidly in great 

 volume, and at such temperatures that the phenocrysts of large size 

 had not yet formed in abundance; and, sixthly, that augite andesite 

 would be most often developed in and about the great volcanic 

 cones, in the building of which the physical conditions and the 

 duration of eruption were appropriate for pronounced fractional 

 crystallization. 



These various implications seem to be fairly matched by the facts 

 of rock-distribution and rock-occurrence. The whole group of rocks 

 here considered — basalts, andesites, peridotites, picrites, limburgites, 

 etc., are consanguineous and several of them are commonly associated 

 in the field. Where some of the derivatives are lacking in the existing 

 outcrops, as in some of the British Columbia cases, the others may have 

 been eroded away, or, on the other hand, have not yet been revealed 

 by erosion. The ultrabasic rocks never form bodies of batholithic 

 size but form dikes, sheets, "chonoliths,"^ etc., and are thus always 

 in such development as is explicable on this hypothesis. They form 

 injected, not subjacent bodies. The olivine basalts are most largely 

 developed in the vast lava-fields of the fissure- eruption type, where 

 the emission of the primary lava has been too rapid for differentiation 

 in the fissures. Augite andesite is most characteristically found in 

 and about the greater cones, like those of the Andes mountain system 

 or like most of the hundreds of very lofty cones rising from the floor 

 of the deep sea. It has been already pointed out that the Hawaiian 

 volcanoes carry superheated lavas, which, on account of the high 

 temperature of the main body of each lava-column, cannot usually 

 undergo pronounced fractional crystallization; yet the studies of 

 J. D. and E. S. Dana show that some separation into olivine-free 

 basalt and ultra-basic basalt has taken place before some of the: 



I Journal of Geology, Vol. XIII, 1905, p. 485 ff. 



