PHYLOGENY OF THE PELECYPODA. 363 



era now taken np, excepting the Unionidiie, two marked features of difference ocenr as 

 compared with those pi-evionsly studied. The 2>rodissoconchs are directed more or less an- 

 teriorly, and the succeeding dissoconchs are not characterized by imsmatic cellular 

 striictwe. 



The Pelecypoda is a class in which it is very difficult to find hinding characters on 

 which to malve systematic divisions. The direction of the prodissoconch umbos and the 

 condition of pi-ismatic sti-ucture in the succeeding dissoconch, I am aware, is not wholly 

 confined to the Aviculidge and their alHes, but these features are characteristic of the 

 whole of that group and are not together characteristic of any other group so far as 

 known, excepting, possibly, the Unionidge. A thin, but well-defined, prismatic layei- ex- 

 ists in Pandora rostrata, Lam.; a prismatic layer also exists in the deeper-seated portion 

 of the periostracnm of Solemya mediterranea, Lam., and it has been observed in limited 

 patches in young Ensis viridis, Say. Cai-penter observed the prismatic layer in the two 

 former genera and mentions finding ti'aces of it in other genera, but he was impressed, 

 with the fact that it is a mai'ked feature in the group of the Aviculidje and allies. 



Mytilas edidis, L. This mollusc belongs to one of the oldest families of the Pelecy- 

 pods extending back to early palaeozoic times. It is therefore interesting to know the 

 conditions existent in the young shell. In PI. xxLs:, fig. 9, it is shown that a well- 

 defined prodissoconch,^, exists which is yellowish-brown and marked by fine concentric 

 lines of growth. The form of the prodissoconch is evenly rounded and it has well de- 

 veloped umbos which are directed anteriorly. The succeeding dissoconch growth of 

 Mytilus is sharply marked off from the prodissoconch stage by its changed form and 

 color. In form the dissoconch has already acquired features characteristic of the group 

 to which it belongs, though not of the genus itself. The hinge line is produced poste- 

 riorl}' in an arched line and antei'iorly the lines of growth are cmwed inwards toward 

 the umbonal area, but they are not crowded closely as in the adult or any species of the 

 genus. The shell at this age reminds one strongly of forms seen commoidy in the 

 Modiomorphas, palaeozoic kindred of the Mytilidae. In coloi- the dissoconch is distinct 

 from the prodissoconch. AYhereas the latter as stated is yellowish-brown, the dissoconch 

 is characterized by the violet color characteristic of the adult. The transition from brown 

 to violet color in Mytilus is as sharply defined as is the change fi'om homogeneoiis lime 

 in the prodissoconch of Ostrea to the prismatic structure of the succeeding dissoconch, 

 fig. 27, p. 314. Specimens of Mytilus edalis of the light horn colored variety of course 

 do not show this change to violet with the introduction of the dissoconch period. No 

 trace of prismatic cellular tissue has been found in the shell of Mytilus or other member of 

 this family at any period of growth, but a well defined epidermis exists.^ Lacaze-Duthiers 

 (41) published an anatomical figure of young Mytilus edulis comparable to the prodisso- 

 conch stage and he notes the striking difference between the shell of that age and the 

 adult. Loven also published an anatomical figure of developing Mytilus which compai-es 

 favorabl}^ in its shell form with PI. xxix, fig. 9. 



An older specimen of Mytilus edidis, PI. xxix, fig. 10, shows no important differences 

 from the younger specimen, fig. 9. The prodissoconch valves on account of the increased 

 arcuateness of the umbos are not seen in side view as well as in the earlier stag-e, a 



'Prismatic structure is described in tlie inner layers of Dreissena, but this genus lias not been studied. 



