510 TRANSACTIONS OF THE WAGNER FREE 



nal characters. But the presence of a distinct pleural ganglion, and of an 

 cesophageal dilatation with lateral, possibly salivary glands, which is affirmed 

 for both by recent morphologists, are more deep-seated and important char- 

 acters. As a trifling makeweight it maybe mentioned that in the whole group 

 of Anomalodesmacea there is no instance of such an epidermis as that of 

 Solemya, while in the Prionodesmacea it is frequent ; and Carpenter has shown 

 that the shell-structure is not only externally prismatic but " its internal layer 

 also exhibits in many parts a cellular arrangement that strongly calls to mind 

 the prismatic cellular structure of Pinna and its allies." 



For these reasons I have decided to return to my first arrangement of 

 1889 and place Solemya with the Prionodesmacea. The typical form of the 

 genus goes back to the Carboniferous, and there is a variety of forms of the 

 same and later age which, though different, seem to be connecting links be- 

 tween the Solemyacea of that time and other groups. 



While the imperfections of the following provisional classification are only 

 too evident to its author, and will justly be criticised by malacologists, it has 

 claims to one merit, which is that the groups, whether well or ill, have been 

 comparably defined. If those who have hitherto proposed new systems for 

 this class, had been obliged to differentially define their groups, with strict re- 

 gard to the known facts, the synonymy would be much smaller and the sys- 

 tems fewer. It must be expected in any classification that specialized excep- 

 tions will occur in nearly every large group. That they do is no reflection on 

 the general accuracy of the grouping. But in several recent attempts at 

 dividing the Pelecypods, the so-called definitions are chiefly remarkable for the 

 little they offer that affords a sound basis for an effort to estimate their value. 



It should be understood that, in the following definitions, characters stated 

 as diagnostic of larger groups are not recapitulated in the diagnoses of their 

 smaller subdivisions, and are then referred to only when specially modified or 

 exceptionally absent. Thus, to obtain all the characters of a family, those of 

 the ordinal, superfamily and other groups, which include the family, must be 

 taken together and modified by the specifications of the family diagnosis. 



