INTRODUCTION. 



67 



Now the ricaturalist, occupying perhaps an msiilated 

 position, and familiar Avith a given species only as it 

 exists in a single section, where its characters are uni- 

 formly the same, cannot a 'priori suppose it to assume 

 other appearances elsewhere ; and when a strongly- 

 marked variety is presented to him, he of course con- 

 siders it to be a new species, closely allied indeed to 

 the one with which he is acquainted, but yet distinct. 

 He hastens to make it known by publication, and thus 

 falls into an error similar to those which have given 

 occasion for these remarks. It is the local naturahst 

 especially, who is most likely to commit this kmd of 

 mistake ; for in the paucity of materials for comparison 

 usually at his command, he cannot have the means of 

 arriving at a more correct judgment. 



These remarks lead to certain inferences, which, if 

 well founded, are worthy of serious attention, and which, 

 combined with other considerations, ought to govern 

 the conduct of naturahsts. In the first place, we see 

 that the diagnosis of species wliich rests exclusively upon 

 external characters is, from their mere uncertamty, not 

 wholly to be rehed upon. The sheU is an extraneous 

 product of the animal, a svibstance foreign to it in 

 some respects, and formed only for shelter in times of 

 danger, and the protection of the soft parts from ex- 

 ternal injuries. It possesses no vitaUty, and its charac- 

 ters are, therefore, as we have seen, subject to change 

 and even to obliteration, to a degree that never happens 

 to organs partaking of the vitahty of the animal. Sci- 



