394 ALFRED HARKER 



composition, an essential and close relationship between the two 

 rocks thus associated. Instead of using the above terms in an 

 altered sense, it will be better to coin new ones, and we shall 

 accordingly recognize two kinds of xenoliths, accidental and 

 cog?iate. This distinction is based, not on difference or likeness 

 in composition, but on the existence, in the latter kind, of a 

 genetic relationship between the enclosed and the enclosing 

 rock, which is wanting in the former kind. A like distinction 

 will apply to the permeation of a solid rock-body by a fluid 

 magma. Now although both permeation and the incorporation 

 of xenoliths are known in instances which fall under the acci- 

 dental category, they are known thus only as quite local phe- 

 nomena. If new rocks of any considerable extent or importance 

 are actually produced by admixture, it is by admixture of two 

 cognate rocks. One reason for this is doubtless to be found in 

 the consideration that reactions between a solid rock and a fluid 

 magma will be promoted by the former being still at a high 

 temperature when the latter comes into contact with it. There 

 may be other reasons of a chemical nature. 



Without discussing at once whether admixture is a factor of 

 prime importance in the genesis of igneous rocks, we may 

 inquire what kind of rocks are to be expected from such a mode 

 of origin. We take first the simplest case, that of admixture 

 between two members of the same rock-series. The chemical 

 composition of the resulting product will be the same whether 

 both or only one of the two rocks be fluid at the time when 

 they are brought together. If the series be a linear one, the 

 admixture will produce a rock having the composition of a pos- 

 sible member of the series. This is Brogger's principle already 

 quoted, which, however, requires to be limited by the condition 

 here imposed. In the more general case the mixed rock will 

 not correspond with a possible member of the series, but will 

 differ more or less in composition from that possible member 

 which has a like silica-percentage. This is clear when expressed 

 graphically (Fig. 2). If OM and OM' represent the silica- 

 percentages of the two component rocks, and their proportions 



