246 Annals of the Carnegie Museum. 



cavities. Outer surface with a peculiar sculpture: heavy, oblique folds 

 run across the disk chiefly in its posterior half (it may be that these 

 folds are continuations of the posterior angle of the bars of the beak 

 sculpture). Epidermis lighter or darker, brown to blackish, without 

 distinct rays. Beak-sculpture either simply concentric, slightly 

 angled upon the posterior ridge, and disappearing toward the disk, 

 or continued upon the disk in a zig-zag pattern, much broken up, and 

 irregular. The soft parts are primitive in structure. Supra-anal 

 separated from the anal, but the mantle-connection between them 

 very short and often absent. Inner lamina of inner gills free. 11 All 

 four gills are marsupial, but there is a slight differentiation in the 

 structure of the inner and outer gills, the water-tubes of the inner 

 gill being slightly wider than those of the outer gill. When gravid, 

 the ovisacs expand a little more, so that their lumen becomes trans- 

 versely enlarged, giving to the placentae a compressed, leaf-like shape. 

 Placentae whitish, not very solid, and not persistent, and the glochidia 

 are discharged in loose masses. Glochidia small, subovate, without 

 hooks. 



Type C. plicata (Say). 12 



Crenodonta approaches the following genus more than the preceding, 

 in fact, it is very closely allied to Quadrula. The chief differential 

 character is the sculpture of the shell. 



Crenodonta plicata (Say). 



Of this species, which is commonly called Quadrula hippopcca (Lea). 

 I have investigated numerous specimens from the shores of Lake 

 Erie in Pennsylvania and Ohio. Gravid females were found in July> 

 1910, but only eggs were present in them, and no glochidia. 



Since there was at hand much more complete mateiial of the follow- 

 ing form, of which this is undoubtedly only a local race, I prefer to 

 only give particulars of the anatomy of C. undulata, here only stating 

 that C. plicata is absolutely identical with it in every respect. 



Crenodonta undulata (Barnes). 

 A large number of specimens from the Ohio drainage in western 

 Pennsylvania are at hand. This species is tachytictic, and the breed- 



11 I have never seen it connected, although it is said to be so sometimes. 



12 This species has been misunderstood hitherto. The type locality of plicata 

 is Lake Erie, and thus the only known Crenodonta from Lake Erie should bear this 

 name, but this is the form called hippopcea by Lea. The plicata of authors (incl. 

 Simpson) should be Cr. peruviana (Lamarck). 



