1895. 



PROCEEDINGS OF THE NATIONAL MUSEUM. 297 



Since it brings together many unrelated species and widely separates 

 others that have strong affinities. In justice to Dr. Lea it should be 

 said that he regarded it as merely a temporary expedient, to be super- 

 seded by a more natural method when a better knowledge of the soft 

 parts could be obtained. ., .• , ,• ^i % 



II von Ihering has recently proposed a new classihcation ot the 35 

 Naiades, taking the form of their larv* as a distinctive character. While 

 the species of Europe and North America have a larva (Glochidmm) 

 furnished with a bivalve shell, which can completely inclose it, a cer- 

 tain number of forms of South America pass through a pecuhar stage, 

 named Lasidimn by v. Ihering, in which the larva is Wed of three 

 segments, carrying only a small single shell on the middle part The 

 same stage is probably passed through by the young of several Atn- 

 can oenera. In consequence he divides the Naiades into two fami- 

 Hes-the Mutelida? (genera Leila, Gray; Fosmla, Lea; Mi/cetopus, A. 

 d'Orbignv; Glaharis, Gray; Aplodon, Spix; P/ar/iodon, Lea; Solenma, 

 Conrad ;^^/^<ie/^^ Scopoli; Iridina, Lamarck; Spatha, Lea) and the 

 Unionid* (genera Hyria, Lamarck; Oa.toiia, Lamarck; Castahna v. 

 Ihering; CTTiio, Philipsson ; Margaritana, Schumacher; Cristaria, Schu- 

 macher; Pseudodon, Gould, and Anodonta, Lamarck). 



The foregoing sketch of the classification of the Naiades is taken in 

 part from the admirable work of Fischer and Crosse on Mexican and 

 Central American mollusks.^ , , , i i p 



In the present state of our limited, and in some cases total lack ot 

 knowledge of the anatomy of several of the genera of Naiades, any classi- 

 fication must be move or less tentative. The division of these mollusks 

 bv most authors into two families, Unionidai and Mutelid*, or two sub- 

 families, Unionina. and Mutelin^e, founded upon the incompleteness or 

 completeness of the development of siphons, can not stand. This has 

 been shown by the researches of Lea and d'Orbigny into the anatomy 

 of Glaharis many years ago; for while some species of this genus have 

 the mantle closed posteriorly so as to form siphons, mothers, which are 

 evidently closely related, the mantle is free. More recently v. Ihering 

 hasshownnhat a given species of his genus Castalina may have an 

 animal which has the two siphons completely developed, thus placing 

 it with the Mutelida^ or it may be that of a perfect Unio, having no 

 siphons at all, thus belonging with the Unionid.^. The same thmg is 

 true to some extent in the well-known genus Castalia, and it is quite 

 probable that this character will be found to vary in other genera ot 



Naiades. -, ., ^ t , a ■4-\ 



So far as conchological characters are concerned, Castalia (and with 

 it Castalina, which has been separated from it) and Hyna, though 

 hitherto placed with the Mutelid*, are evidently memb ers of the 



'ArohivfiirNaturgeschichte, p. 52, 1893. . tt „ r,n-. 



:Mi8Si(.n Scientifique an Mexiqiie et dans rAmerinue Centrale, .th part, II. V-oVo, 



1891. 



^■Zuol. Auzeiger, Nos. 380 and 381, 18Ul-til>, pp. 1-14. 



