THE CLASSIFICATION OF IGNEOUS ROCKS. 485 



ment of opinion since chemical and microscopical researches 

 have afforded data for attempts in the direction of a natural 

 system. 



No allusion has been made above to the question of the 

 nomenclature of rocks. This is to some extent affected by 

 the difficulties which beset the problem of rock-classification, 

 and indeed many of the names now in use are little better 

 than " Sacknamen," as Brogger has expressively described 

 them. Names covering groups and families of rocks cannot 

 become fixed until the groups and families themselves have 

 been accurately defined, and their limits have become a 

 matter of common agreement. Much confusion has been 

 gratuitously introduced by the practice of arbitrarily ex- 

 tending, restricting, or changing the meaning of classifi- 

 catory or descriptive names. Some, like syenite and 

 granophyre, are used in a sense which contradicts their 

 etymology ; others, like pyroxenite, are applied by different 

 writers to totally different groups of rocks. One fruitful 

 source of confusion is the extension to a family of rocks of 

 a name, such as foyaite or tonalite, belonging originally to 

 a single type. 



Names of rock-types must clearly be numerous, and in- 

 creasingly so. To this there can be no objection, provided 

 the names are strictly reserved for rocks of the types in 

 question, and provided a new name is given to a new type 

 of rock only when the latter has been fully described. The 

 names least likely to cause difficulty in the future are those 

 derived from the localities of the typical rocks, and thus 

 embodying information which is independent of any theory. 

 To take the type rock from one locality and the name from 

 another, as in the monchiquite of Hunter and Rosenbusch, 

 is a proceeding to be deprecated. If ever a satisfactory 

 classification is arrived at, it will not be difficult to embody 

 it in a new nomenclature, possibly a binomial or trinomial 

 system as suggested by Wadsworth (17); but any such 

 systematic terminology presupposes a knowledge of the 

 true relationships of igneous rocks which is still a de- 

 sideratum. 



