504 PEOCEEDINGS OF TRE NATIONAL MVSEUM. vol. xxii. 



gills begiu to be crimped as they are being filled with embryos, until 

 when full, they become a series of marveloas folds. In U. irroratus 

 several ovisacs in the center of the outer gills grow out to a great 

 length, become filled with young, and are closely coiled. In TJ. cor- 

 ntitus a few central ovisacs develop so as to project below in a long, 

 straight flap. In all the South American and Australian Unios, so far 

 as is known, the inner gills alone, as a rule, are filled with young, and 

 this is probably the case with the species of the Ethiopian region and 

 most of those of southeastern Asia. After the young have passed out 

 into the water the gills of all the species change back into their ordi- 

 nary condition, and when not gravid there is great similarity in those 

 of most of the sr)ecies formerly classed as Unios. 



These peculiar evanescent characters, assumed when the gills are 

 transformed into marsupia, seem to be quite constant, and I believe 

 they can be used as a basis for the foundation of genera. When these 

 are once discovered and understood it will be found, on careful exami- 

 nation, that there are minor shell characters that correspond with those 

 of the marsupia, and which help us in i)lacing certain forms of whose 

 anatomy we know but little. Thus the principal shell characters of 

 such species as Unio trigomis, metanevrus, and plicatus are much like 

 those of TJ. gihhosus, crassidens, and hucMeyi. But in the former set 

 all four gills will be found to be filled with young in the gravid female, 

 while in the latter only the outer ones are full. And the former have 

 generally shorter, solider, more inflated shells and deej) heal: cavities, 

 while the latter have longer, lighter shells, which have shaJloiv healc 

 cavities. In Anodonta edentnla the ovisacs are short and run across 

 the gills 5 in the very similar looking A. ferussaciana the outer gills are 

 filled with oblique ovisacs, and the inner have (at least in some cases) 

 more or less ovules. The beak sculpture of these two species is very 

 'different. 



From the earliest period in which the ISTaiades have been studied to 

 the present time it has been claimed by some that the sexes were sej)a- 

 rate, by others that the animals were hermaphroditic. Eecent careful 

 studies by such men as Sterki, Taylor, Kelly, and others appear to 

 demonstrate that in the more highl}^ organized Unionid;e (those which 

 have two forms of shells, and have the ovisacs in the hinder part of 

 the outer gills) the sexes are always separate. In the more simply 

 organized Union idte (those with but one form of shell and with the 

 embryos occupying the entire gill) the sexes may or may not be 

 separate. 



Further study has shown me that the provinces established in 

 the paper I have quoted hold good as there laid down. There is evi- 

 dently a very close relationship between many of the unionoid forms of 

 southeastern Asia and tropical Africa. So close is this relation that 

 the two regions might be united if it were not for the fact that a large 



