TRANSITION TO CLASSIFICATORY SCIENCE. 199 



allowed to be a metal, is silicium one, a body which 

 wants the metallic lustre, and is a non-conductor of 

 electricity? It is clear that, at least, the obvigus 

 application of a classification by physical charac- 

 ters, is attended with endless perplexity. 



But since we cannot even begin our researches 

 without assuming a classification, and since the 

 forms of such a classification which first occur, end 

 in apparent confusion, it is clear that we must look 

 to our philosophy for a solution of this difficulty; 

 and must avoid the embarrassments and contradic- 

 tions of casual and unreflective classification, by 

 obtaining a consistent and philosophical arrange- 

 ment. We must employ external characters and 

 analogies in a connected and systematic manner; 

 we must have Classificatory Sciences, and these 

 must have a bearing even on Chemistry. 



Accordingly, the most philosophical chemists 

 now proceed upon this principle. "The method 

 which I have followed," says M. Thenard, in his 

 Traite de Chimie, published in 1824, "is, to unite 

 in one group all analogous bodies ; and the advan- 

 tage of this method, which is that employed by 

 naturalists, is very great, especially in the study of 

 the metals and their compounds 2 ." In this, as in 

 all good systems of chemistry, which have appeared 

 since the establishment of the phlogistic theory, 

 combustion, and the analogous processes, are one 

 great element in the arrangement, while the dif- 



2 Pref., p. viii. 



