PRIMARY CIRCLE. 



261 



Romans. Some of the American species are remarkably 

 thick, particularly at their bosses, which, in proportion as 

 they are rubbed, the animal thickens internally; and this 

 is very remarkable in old shells. In regard to the ana- 

 tomy of the animals, we must refer the reader to Poli's 

 account of the European species, and to the scattered 

 notices of the American in various other publications. 



(246.) The natural arrangement of this family, or 

 rather the principle of its variation, in regard to the shell, 

 is precisely the same as in all the other groups of the 

 DITHYRA. The most typical groups are those two 

 which stand between the elongated and the cuneated 

 types ; the most aberrant being intermediate with the 

 two latter, and having no teeth. The typical groups, 

 consequently, possess the two sorts of teeth, cardinal and 

 lateral, in the highest perfection ; while, in the three 

 aberrant divisions, only one of these sorts is apparent, 

 or none. As a whole, this is the most perfect family, 

 perhaps, of the bivalve Testacea. Like the Psittacidae 

 and the Picidce among scansorial birds, the shells of the 

 UnionidtB have such a stamp of identity upon them, 

 that they cannot possibly be confounded with those of 

 any other family ; while their amazing diversity, on the 

 other hand, offers the very best facilities for determining 

 the natural succession of their forms. Under the belief 

 that this would long ago have been done by those Ame- 

 rican conchologists who have especially studied this 

 portion of their native fauna, we suspended our labours 

 towards this object some years ago ; but as nothing has 

 yet been done to supercede our own views, they will 

 here be detailed with as much brevity as the subject 

 will admit. 



(24?.) Presuming that the station of the UnionicUs, 

 as a family, has been determined ; and that Iridina, 

 from possessing true siphons, is that aberrant group 

 which comes nearest to the Chamadce in the tribe of 

 MacrotracMa, we thus get a sure point from which to 

 start : our first object is to show that the whole family 

 form a circular group, with no chasm, hiatus, or inter- 

 s 3 



