GENERAL RESULTS OF COMPARISON. 125 



only to the low percentage in the feldspar and augite, but also to an equally- 

 low percentage in the base. The high percentage in rhyolite and trachyte 

 is due not only to the feldspar, but still more to the even higher percentage 

 of silica in the base. If there has been segregation, it must, therefore, have 

 affected not only the crystals, but the base even more than the crystals. 

 Such a separation, therefore, does not seem explicable by supposing a pre- 

 cipitation of crystals. 



Gathering together now the threads of this comparison, we are led to 

 the conclusion that the constitution of the eruptive rocks forbids the belief 

 that the acid varieties, or even the intermediate varieties, can be primordial 

 masses from vesicles which separated in a liquid condition from the original 

 earthmass and remained liquid up to the time of their eruption. Chemical 

 considerations of a cogent character lead up to the inference that primordial 

 magma ought to possess a constitution similar to rocks of the basaltic group, 

 though perhaps somewhat less ferruginous (?), and that it should be nearly 

 homogeneous. And in general our inference from the nature and constitution 

 of the volcanic rocks, from their great variety, from the localization of 

 eruptive phenomena, from the intermittent character of volcanic action, 

 from the independence of the several vents, is that the lavas do not emanate 

 from an earth-nucleus wholly liquid, nor from great subterranean reservoirs 

 still left in a liquid condition " from the foundations of the world," but from 

 the secondary fusion of rocks, a part of which may have formed the primi- 

 tive crust, while the remaining part consisted of deeply-buried and meta- 

 morphosed sedimentary strata. No doubt some cautious philosophers may 

 regard this inference as specifying a little too minutely the locus of volcanic 

 activity — more minutely than a rigorous deduction from known facts will 

 permit us to regard as positively proven. But at all events there is one 

 proposition which may be laid down with no small degree of confidence, 

 and it is this: We must at least admit that tke source of lavas is among segre- 

 gated masses of heterogeneous materials. This arrangement would be well 

 satisfied by a succession of metamorphic strata resting upon a supposed 

 primitive crust of magma having a constitution approximating that of the 

 basaltic group of rocks. 



II. The second general consideration has reference to the dynamical 



