^l^^- THE SEQUENCE OF PLUTONIC ROCKS. 289 



of geological time. He also supposed that in some cases eruptions 

 might take place from local basins in the earth's crust, and not from 

 the universal concentric zones. 



The theory of the two magmas possesses, at present, little more 

 than an antiquarian interest. It was a bold attempt to solve, with 

 wholly insufficient data, the problem of the sequence of volcanic 

 rocks. The principal interest for us lies in that part of the paper 

 dealing with what the author terms " liquation." He says : — " The 

 magmas which have produced the igneous rocks are comparable to 

 fused masses of several metaJs which separate in cooling into different 

 alloys, according to the circumstances of solidification." Here we 

 have a clear and distinct recognition of the important fact that a 

 homogeneous magma may, in the process of cooling, become diff"eren- 

 tiated, and so give rise to igneous rocks of varying composition. This 

 is the mode of viewing the matter which is at present receiving a 

 considerable amount of attention. 



The attentive study of any crystalline igneous rock is sufficient to 

 convince one that the consolidation of a molten magma is accompanied 

 by differentiation. The original magma was homogeneous ; the 

 resulting rock is heterogeneous. As the molten mass cooled 

 molecular groups of magnetite, augite, felspar, and the like were 

 built up, became locally concentrated, and finally coalesced to form 

 solid crystals. The whole rock did not consolidate at once. The 

 minerals commenced to form at different stages. There was a 

 sequence in this process of crystal building. If, after consolidation 

 had progressed to a certain extent, the crystals had been filtered 

 away from the mother liquor, the two portions thus obtained would 

 have been found to differ widely in composition from each other and 

 from the original magma. Suppose the original magma to have had 

 the composition of an andesite or diorite, then the two portions 

 would differ from each other as an acid rock differs from a basic 

 rock. The crystalline aggregate would be poor in silica, rich in lime, 

 iron and magnesia, and poor in alkalies, especially potash ; the 

 mother liquor would be rich in silica, poor in lime, iron and magnesia, 

 and rich in alkalies, especially potash. That this is no mere specula- 

 tion, but an actual fact, can be proved by observing the order in 

 which the minerals form, or by separating the ground-mass from the 

 porphyritic crystals in some typical andesite, and determining its 

 composition by direct analysis. Thus, the ground-mass of one of 

 the Cheviot andesites has been shown by Petersen to have the 

 composition of a quartz-felsite or liparite. 



Differentiation by crystal-building in magmas of intermediate 

 composition may therefore give rise to secondary magmas of 

 increasing acidity. Are there any facts which such a differentiation 

 would explain ? 



Again, the phenomena of crystal-building seem to imply that 

 differentiation must precede the actual formation of a crystal. 



