BASIS OF THE CLASSIFICATION OF VOLCANIC KOCKS. 85 



assumed correlations are, so far as we can discern them, seen to be only 

 very partial and imperfect. SLill we may hold that, for the time being, the 

 best classification is the one which expresses the largest number of facts 

 and relations hitherto ascertained, and we may advantageously adopt such 

 a classification in preference to any other, though conscious that it fails to 

 bring into recognizable order some outstanding facts and relations which 

 we are compelled for the present to look upon as anomalies. 



In proposing a system of classification of volcanic rocks, I shall endeavor 

 to conform to the foregoing conceptions as to the purposes and scope of 

 any or all classifications. Strictly speaking, I can pretend to nothing more 

 than the most convenient and accurate expression which the nature of the 

 case may admit, of the state of my own knowledge and convictions con- 

 cerning the properties and relations of volcanic rocks. Holding that all 

 classifications are ephemeral, merely indicating the instantaneous phases 

 of advancing knowledge, it is fully admitted to be an artificial one for the 

 most part, and is natural only so far as nature has been truly discerned 

 and expressed. The object in presenting a new classification instead of 

 selecting and adopting an old one is to give precision to the terms employed, 

 and to lay down from the beginning a systematic statement of the views 

 entertained regarding the affinities of the various kinds of eruptive rocks 

 so far as known and understood by the individual writer. Not only does 

 there seem to be no impropriety in any or every writer expressing as accu- 

 rately and systematically as possible his own views of such relations and 

 affinities, but it is rather incumbent on him to do so, and in no way can 

 this be accomplished so compendiously as by a scheme of classification.* 



In a classification of volcanic rocks, the facts which it is desirable to 

 formulate and arrange are, first, those having reference to the physical con- 



*I may advert here to a malpractice of some writers, who take advantage of slight pretexts to 

 coin new names for slightly-altered divisions of old groups. A new name is always au inconvenience, 

 even though it may be necessary ; unless, indeed, it he a purely descriptive one, conveying at once its 

 significance or giving some conception of its meaning to one who hears it for the first time. Thus, the 

 introduction of such names as protogene, elvanite, nevadite, miascite, &c, entails the necessity of 

 much labor and effort to fix in the memory their meaning, all of which might have been avoided and 

 every useful purpose subserved by using the terms hornblendic granite, quartz porphyry, granitoid 

 rhyolite, nephelin syenite, &c. Irrelevant terms like the first may be very convenient to the. writer or 

 speaker, hut they are very inconvenient to the reader or hearer. Inasmuch as all classifications are 

 evanescent and constantly shifting, it is manifestly desirable to make them as easily intelligible as 

 possible. 



