174 Reviews — A New Rock- Classification. 



the strict sense. Much of it is doubtless of great value ; but its 

 value can be only very partially realised, for the reason that this 

 accumulation of material has far outstripped the other side of the 

 science, the business of which is to co-ordinate the scattered 

 observations and bring to light the principles which underlie them. 

 The pearls are there, but there is no thread on which we can string 

 them. During recent years, however, we have witnessed many 

 indications of a reaction from this state of things. From various 

 quarters memoirs have been put forth in which very broad views 

 are taken of the groups of rocks dealt with, the question of their 

 mode of origin being kept constantly in sight. We may remark in 

 passing that American petrologists, including some whose names 

 appear at the head of this article, have taken their part in the 

 advance on these lines. Much stress is now made on the essential 

 mutual relationships of associated rock-types, and Brogger has gone 

 so far as to give a provisional genealogical tree for those of the area 

 which he has made his own. Such essays in the direction of 

 a comprehensive genetic treatment are necessarily tentative, but 

 we have no sympathy with those who would dismiss them as mere 

 speculation. Their influence is already very apparent. Petrologists 

 feel that they have almost within grasp a fundamental principle 

 analogous to that of descent, which lies at the root of classification in 

 the organic world ; and this warrants a confident hope that there 

 may eraei'ge in due time a truly natural classification of igneous 

 rocks. If the expectant attitude of mind here suggested is, as we 

 think, very general, the appeai'ance of an elaborate artificial sclieme 

 at the present juncture will seem to many a step in the wrong 

 direction. 



The standpoint of the authors is clearly defined in one of their 

 prefatory remarks : " The present systems are to a certain extent 

 founded on theory or hypothesis, while classification in order to be 

 stable must eschew all such bases, and be founded only on ascertained 

 facts." 



As interpreted by what follows, the stability here made a 

 desideratum must be understood as rigidity ; for the argument 

 leads to the setting up of an unyielding framework, in which, 

 between the arbitrary partitions, each rock, whether discovered or 

 undiscovered, may find its appropriate pigeonhole. Although we 

 are under the disadvantage of having no alternative scheme to offer, 

 we venture to submit that an elastic classification, embodying what 

 is known and capable of being adapted to the requirements of 

 increased knowledge, might with equal propriety be described as 

 founded on ascertained facts. At least, it may be pleaded with some 

 reason that existing systems fail, not because they involve a certain 

 element of hypothesis, but in some measure because their implied 

 hypotheses are fallacious. 



Some criticism on general grounds we have thought admissible ; 

 but the scheme before us is professedly empirical, and we cannot do 

 justice to it without adopting for the purpose the same point of 

 view. The first impression of complexity, due partly to the strange 



