27 



earth which rocks hold in nature. The smaller, 

 are necessarily derived from mineralogical cha- 

 racters ; and as these are subject to the former in 

 the system of arrangement, an obvious species of 

 irregularity must follow ; as the same compound 

 may be found under more than one of the larger 

 divisions. But this is a question that ought to be 

 tried by the principle of utility. 



In the other departments of nature, a rigid ad- 

 herence to the logic of the adopted system is easy ; 

 because the objects are definite, constant, and 

 connected by invariable and simple relations to 

 each other: it is equally useful, as it facilitates 

 that investigation which is the object of every ar- 

 rangement. It is, or ought to be, adopted, not 

 because it is consistent, but because it is useful ; 

 and it is mistaking the means for the end, to trans- 

 fer to t'e present class of objects, those rules to 

 which, from their very nature, they are not ame- 

 nable. To sacrifice thus to the forms of an ar- 

 rangement, is to be anxious about words and neg- 

 ligent of things ; it is to recur to that philosophy 

 of the antient contemplatists which the labours of 



