Pirsson — Petrography of Tripyramid Mountain. 429 



group, then they should not be associated with norite and gab- 

 bro, for these are members of the alkalicaclic (or sub-alkalic) 

 rock series. It is recognized, of course, that rocks consisting 

 of ferromagnesian minerals and lime-soda feldspars occur in 

 the alkalic rock series, and to one of these with certain min- 

 eral and chemical features the name of essexite has been given. 

 It is also understood that these rocks, of which essexite stands 

 as a good example, show by the kinds of pyroxene and 

 hornblendes they contain, by presence of more or less nephe- 

 lite, or sodalite, and similar peculiarities, which may be regarded 

 as tribal markings, their allegiance to the clan from which they 

 have sprung. 



At this point the writer desires to say that with this view, 

 recognizing the genetic relations existing between rock series, 

 he is in hearty accord and that he believes the demonstration 

 of it, made during the last fifteen or twenty years, to have 

 been one of the most important results of petrologic research. 

 But the question which the Tripyramid assemblage brings to 

 mind is, how far can this be carried, and how invariable are 

 the associations of given rock types ? There are some perhaps 

 who, insisting upon the law of invariability, would say that in 

 this case the gabbro and norite are not really such, but are 

 certain types, or aberrant forms, of essexite. But if we hold 

 this view, then the definition of essexite becomes so broad and 

 general as to lose any specific value. So far as one can see 

 there are no chemical, or mineral, peculiarities about the Tri- 

 pyramid gabbro which would differentiate it from the usually 

 accepted types of that rock, or any grounds on which we could 

 call it an essexite beyond the fact of its association. But if we 

 do this and thus push the genetic view of classification to its 

 limit, we shall have to admit that rocks (in some cases) are not 

 to be classified as kinds according to the inherent properties 

 which they possess, but by these and their associations. Thus 

 we shall find rocks whose inherent properties are similar in two, 

 or perhaps more, places in our classification. But what if we 

 find such a type alone, by itself, without any associations, a 

 not uncommon occm'rence ? Then we shall be unable to 

 classify it, or else we shall have to adopt the procedure 

 of sometimes classifying rocks by one method and sometimes 

 by another, a process which can scarcely commend itself as an 

 orderly or logical one. This, to the writer's mind, is one of 

 the chief difficulties of rock classification as based on genetic 

 association and descent, and at present no practical way of 

 solving it appears. It has also indirectly been brought for- 

 ward by Whitman Cross* in a recent paper treating of this 

 subject. 



* Natural Class, of Ign. Bocks, Quart. Jour. Geol. Soc, lxvi, p. 481, 1910. 



