Value of the Shells of Mollusca. 263 



If we modify Mr Gray's system by dividing the Brachio- 

 poda into two orders, and by regarding Argonautidae as the 

 shell of an Octopod, we have (including the Tunicata) seven- 

 teen orders, of which two are naked, ten are wholly testa- 

 ceous, and five are partly testaceous. But of the fifteen 

 which are more or less testaceous, only five are easily sepa- 

 rated by ordinal characters in the shells, and the others are 

 distinguished with more or less difficulty. In descending 

 to ihe families we find the value of the shells much greater 

 than in the orders. Here will I limit my remarks to the 

 families in Mr Gray's list of genera, both to avoid unneces- 

 sary detail, and because they are constituted with more re- 

 gard to the soft parts than those of any other classification. 



In the class Cephalopoda the two testaceous genera (Argo- 

 nauta being excluded from the class) constitute as many 

 families, which are, therefore, amply indicated by their shells. 

 In the first order of Gastropoda, Pectinibranchiata, Mr 

 Gray makes five families, of which three, Strombidse, Volu- 

 tidae, and Cypraeadse, are easily distinguished by their shells ; 

 but the family MuricidsB embraces some species of Pleuro- 

 toma and of Conus, whose shells are not easily distinguished 

 by family characters from some of the Buccinidae. We are 

 not aware, however, that the soft parts could be distinguished 

 with any greater facility. This is an example of the grada- 

 tion of types into each other. 



The second order, Phytophaga, contains twenty-five families, 

 of which one, Atlantidse, is naked; Tecturidse (= Lottiadae) 

 is not distinguishable by ordinal characters of the shells from 

 Patellidse in another order ; the Littorinidae, containing Sola- 

 rium, cannot be distinguished from the Trochidae, which con- 

 tains Phillippia ; and Vermitidae and Dentalidae so closely re- 

 semble certain Annelides that it is doubtful whether any family 

 characters can distinguish them. The remaining nineteen 

 families may be distinguished by their shells with greater 

 or less facility ; the difficulties, where they exist, arising from 

 gradation of types, not from a want of correspondence be- 

 tween the shells and the soft parts. 



The third order, Pleurobranchiata, contains five families, 

 with a remarkable intermingling of testaceous and of naked 



82 



