124 GEOLOGY OF THE HIGH PLATEAUS. 



But not only should we infer that the primordial masses of the earth 

 (or "primitive crust") were basic like the basalts or dolerites, but that 

 they were very neai'ly homogeneous. If we are at liberty to speculate at 

 all upon the physical condition of an all-liquid planet, its molten sui-face 

 exposed to radiation and to the action of its immense atmosphere, we 

 should be led to infer that it would be agitated by disturbances similar 

 in nature, though inferior in magnitude, to those affecting the sun, thus 

 producing a thorough and homogeneous mixture of the compounds of 

 silica with alumina, the earths, and alkalies. This admixture once formed 

 would, so far as we can now see, remain unaltered until it cooled suffi- 

 ciently for the reactions of the atmosphere. We know of no natural 

 processes capable of separating the more acid parts of such a magma 

 except the chemistry of the atmosphere acting at temperatures far below 

 the melting-points of the silicates. We have the results of that process in 

 the quartzites, granites, gneisses, and syenites among the siliceous rocks; 

 and the limestones and dolomites among the basic rocks ; with argillaceous 

 rocks as the residuum of the decomposition. Yet if these rocks could be 

 remelted together they would form one homogeneous magma. Every iron- 

 smelting furnace is an experimental demonstration of the tendency of silica 

 to take up and hold at fusion-temperature alumina, lime, magnesia, potash, 

 and soda in proportions exceeding those which occur in nature. No facts 

 are known to me which justify the conclusion that segregation into two 

 magmas could occur in such a state of fusion. Nor would it be of any 

 service in this connection to establish the possibility of such a segregation.* 

 It is suggested by Mr. King that crystals might form in the liquid and sink 

 by reason of their superior specific gravity. Although I hold it to be 

 extremely doubtful whether any crystals are formed while the rocks ai'e 

 melted, and very probable that the greater part of them are formed during the 

 viscous stage of cooling (especially the hornblendes and pyroxenes), there 

 is one consideration which would prevent us from using this view to predi- 

 cate a theory of a single magma separating into two or more of very differ- 

 ent degrees of acidity. The low percentage of silica in basalt is due not 



* Iron, however, might separate from such a compound, either as a regulus or as magnetic oxide, 

 if the conditions were favorable and the oxide in excess. 



