PHYLOGENY OF THE PELECYPODA. 377 



coiulitioiis of the case.^ In support of this view, the bivalvular crustaceans may again be 

 cited as they have an analogous adductor muscle, developed, of course, on an entirely 

 different line of descent, but nnder closely similar mechanical conditions. In the com- 

 pleted prodissoconch of all Pelecypods there are two adductor muscles, PL xxiv, figs. 

 1-2, PI. XXX, fig. 4,^ thus arriving at what may be called in this respect the typical con- 

 dition of the class. During later life the anterior, the posterior, or both adductors, may 

 be retainefl, reduced or lost, according as the persistence or changes in correlated feat- 

 ures of anatomy retain in use or bring into disuse the muscles in question. 



The preceding studies show the existence of an embryonic shell to which I have given 

 the name of prodissoconch (p. 311) in many widely separated genera of Pelccypoda. 

 We will now consider this early shell and see if any close connection can be traced be- 

 tween it and ancestral forms of which it is the representative in early stages of develop- 

 ment. 



Two classes of prodissoconchs and succeeding shell structure are described in the 

 preceding pages. In the first division the prodissoconch has umbos directed posteriorly 

 and is not prismatic, but is succeeded by a dissoconch having an extei-nal layer of pris- 

 matic cellular tissue which is moi'e or less developed, but exists at least in the early 

 nepionic stages of one valve. The?e features are characteristic of the Aviculidae, Os- 

 treadfe, Pectiuid;\i and Anomiada?, as shown by representative genera. The second di- 

 vision is less well defined; but with one exception can be sepai'ated from the fii-st. The 

 prodissoconch of this second division has umbos which are directed more or less ante- 

 riorly. It has no prismatic cellular layer and is succeeded by a dissoconch which has no 

 pi'ismatic layer. Such features are characteristic of genei'a repi'csentingthe families My- 

 tilidte, Arcicbe, Chamid;B,Petricolid», Cycladida3,Veneridie,Glycimericla3, Scrobiculaiidaj 

 and Myida?. The 3'oung of the Unionidye as described (page 3L)5) does not come 

 under this second section but in some features approaches the first. Only a part of the 

 families of Pelecypods are included under these two divisions. Whether other families 

 will come under one or the other, or yet new divisions, must be proven by future inves- 

 tigations (see note, p. 375). 



The first of the divisions above considered embraces a group of families (the Avicu- 

 lidse and allies) intimately bound together as I have attempted to show in this paper. 

 It will probably prove to be a group of ordinal value and I shall limit myself to this 

 gwap in considering the derivation of the j^rodi.'^socoyich. The prodissoconch of the sec- 

 ond division has doubtless a different ancestral kinship or perhaps several lines of ances- 

 tral kinship as the case may be. 



The features by which a possible connection may be ti'aced to the radical ances- 

 tor from which the prodissoconch form is inherited includes features of shell structure 

 and anatomy, for we Ivuow both in a measure from the results of the studies in this pa- 



'The mantle of molluscs often has a highly developed 'Also figures by Lovfin, Schmidt and other investigators 



muscular system, and it seems probable tiiat the adductors of marine and fresh- water Pelecypoda The fact of the 



of the Pelecypod arose from a modification of mantle two adductors is also attested by the form of the prodis- 



mnscles. A somewhat similar modification of these mus- soconch in the genera figured on Plates xxvi-xxx inclu- 



cles is seen in the retractor muscles of the siphon, charac- sive. 

 teristic of many Pelecypods, as the clam, PI. xxv, flg 9. 



