ortmann: south amehican naiades. 465 



Most of the species arc subelliptical or subovate, and moderately elongated; but 

 others approach more or less the sul)orbicular sha])e. For the latter, Simpson 

 has introduc(>d the subgenus Cydomya, and it is true that these form a rath(>r \v(>ll- 

 defined group, although there are certain species which seem to be intermediate. 



The sub(iuadrate or subtrigonal shape of the Cadalia-gvo\x\) is imknowTi in 

 Diplodon, and the well-defined posterior ridge characteristic of this group is very 

 rarely found; the species, which have it, are always rather elongated (not trigonal), 

 so there cannot be any mistake about them. 



The hinge-teeth are also variable, but generally represent, or are easily com- 

 pared with, the normal type described above (p. 459). There are one or two 

 pseudocardinals and two laterals in the left valve, and two pseudocardinals and 

 one lateral in the right valve. The pseudocardinals may be stumpy, or more or 

 less compressed. They, as well as the laterals, may be smooth, or have corruga- 

 tions, or may be split and dissected. They never have parallel ridges. 



The beak-sculpture also varies a good deal. It is always radial, but the 

 ridges composing it may be shorter or longer, finer or heavier, smooth, or dissected 

 by growth-lines, granular, and sometimes they may be irregular. 



But all of these characters of the shell are connected with each other by nu- 

 merous transitions, and it is impossible to say that any one type is more primitive 

 than another. It is also extremely diflficult to arrange the species of Diplodon into 

 groups according to these features. Simpson (1914 pp. 1225, 1228) has divided the 

 genus into subgenera and subordinate groups, and has attempted to condense the 

 results in the shape of a "key," but this key is practically worthless, of which fact 

 anyone may convince himself, when he tries to use it for the identification of a 



species. 



However, when I studied the soft parts of the various species at my disposal, 

 I discovered that there are rather well-marked and apparently important characters 

 shown by the anatomij, especially by the structure of the marsupial pari of the inner 

 gill of the female, while the rest of the anatomy is the same in all species. 



The interlaminar connections of the gills are extremely weak in the male of 

 Diplodon, and in the outer gill of the female (which has the same structure as in 

 the male). They correspond to the description given b>- me for Hyria and Castalia 

 {Tetraplodon) (Ortmann, 1911«, pp. 115, 117). (See the figures of gills of various 

 species on Plates XLV, XLVI, XLVII; also sections of the gills on Plates XLVII 

 and XLVIII). These interlaminar connections are few and scattered over the 

 face of the gill, and do not form distinct septa (thus resembling the condition seen 

 in the Margaritanidce) . But in the marsupial part of the inner gill of th(> female, 



