CANCROIDEA. 143 



1. The Xanthidee; carapax broad, with an arched front; inner 

 branch of first maxillipeds dilated at apex. 



2. Gancridce; carapax broad, with an arched front; inner branch 

 of first maxillipeds narrow at apex. 



3. EriphidcB; carapax subquadrilateral; inner branch of first max- 

 illipeds narrow at apex. 



The groups here indicated are of real importance. But the charac- 

 teristics laid down do not affix to them their true limits. There are 

 true Xanthidse in which the inner branch of the first maxilliped is not 

 dilated at apex; for there are those that have as narrow a carapax 

 as any of the Eriphidse, and a similar branch to the first maxillipeds. 

 According to the characteristics mentioned, the Eriphida? properly 

 include Pilumnus; and not only Pilumnus, but also the narrow 

 Chlorodii, some of which are nearly as narrow as long, and which are 

 widely removed from Pilumnus in important characters. The dis- 

 tinctions of narrow and broad, happens in the family to be of little 

 general value in classification, except when viewed under certain 

 restrictions required by qualities of higher importance. 



In the study of these species, there are actual difficulties in the 

 way of arriving at natural subdivisions with conveniently circum- 

 scribed limits. The difficulties arise mostly from the fact, that no 

 such limits exist as the systematist often looks for. Nature has 

 made her fields without fences; and although there are some moun- 

 tain ranges, in general, the blendings among the lower subdivisions in 

 the kingdoms of life are by gentle gradations. The true object 

 of classification, consists in tracing out gradations and inter-reticula- 

 tions among groups. Keeping this in view, we shall not be dissatis- 

 fied if the groups laid down are found to shade into one another, 

 instead of standing apart in bold relief. Such trenchant subdivisions 

 are necessarily artificial, and although the simplicity with which they 

 are characterized may gratify, they are to be looked upon with dis- 

 trust, and generally as sure evidence that but a small portion of the 

 field of study has actually been surveyed. 



We have already (page 68), stated the reasons for including Tel- 

 phusa and the allied genera with the Cancer group ; and have alluded 

 also in the same place, to the relations of Acanthocyclus to this group. 

 The Telphusa family are fresh-water Cancroids, and mark the transi- 

 tion to the Grapsoidea; while the Acanthocyclus is related to the 

 Corystoidea. We are thus led to arrange the Cancroidea in three 



