ortmann: south American naiades. 467 



There arc other differences in the marsupium of the species of Diplodon, and 

 these depend upon its location within the gill. Tn the vVustraHan Ilijridella nearly 

 the whole inner gill is marsui)ial, only very small i)ortions at the anterior and 

 posterior ends remaining non-marsupial, a very common condition in the Unionidm 

 where the whole outei- gill may be marsui)ial. In Diplodon we sometimes observe 

 a similar arrangement (0. picetifi, Plate XLVI, fig. 2), but generally a more con- 

 siderable i)art at the anterior as well as at the posterior end of the gill is non- 

 marsupial, so that the marsupium is restricted to the middle portion (about half) 

 of the gill. The marsupial part may become still more reduced in size, and may be 

 located not in the middle of the gill, but more toward the front, or toward the pos- 

 terior end. In either case the posterior or anterior half of the gill is nearly or 

 quite non-marsuiiial. It should be noted that I have observed this shifting of 

 the marsui)ial part only in cases where the marsupial structure is comparatively 

 primitive, with the interlaminar connections arranged in a reticulate way or as 

 interrupted septa, but never in forms with solid septa, where the marsujiial part 

 always extends over a large section in about the middle of the gill. (Compare 

 figures on PI. XLV, XLVI, XLVII.) 



All these differences described in the structure and location of the marsupimn 

 are constant loithin the species. In some forms, indeed, my material is rather scanty ; 

 but in quite a ninnber of other cases I have a sufficient number of individuals 

 with soft parts, and I have invariably found that all females of the same form and 

 from the same locality agree with each other, with the only qualification that in 

 young females the marsupium is generally less extended, and occupies a smaller 

 section of the gill (Compare PI. XLVI, figs. 5b and 5c; PI. XLVI, figs. 7a and 7b). 



It should be emphasized that these marsupial differentiations are only found 

 within the genus Diplodon. The other genera of which I have anatomical material 

 stand generally upon the stages which I have described as the more primitive in 

 Diplodon. Firstly the structure of the marsupium is of the reticulate type in 

 the midtUe, with more or less development of interrupted septa towards the edges 

 of the gill in Hyria (Ortmann, 1911a, p. U5) and Ca.stalia undosa (^ibid., p. 117), 

 while in Castalina nehrvngi the structure showing perforated septa prevails. Ca.s- 

 talia acuticosta seems to agree with C. undosa, but my material consists of only two 

 gravid, rather young specimens, in which the structure cannot b(> clearly seen on 

 account of the mass of glochidia filling the marsupium. Secondly the location of 

 the marsupium in Hyria and Ca.<^(alia is in the middle of the gill (one-fourth at 

 anterior end, less than that at posterior end, non-marsupial). In Castalina nehringi 

 (Plate XLVII, fig. 2) the marsupium has moved a little moi-e backward with about 



