314 MEMOIRS OF THE CARNEGIE MUSEUM 
mantle is the essential feature of this structure, which serves to close and define 
the anal siphon dorsally, while the upper opening (supra-anal), which remains 
unclosed, is only incidental to this, so to speak a by-product. 
There are a few forms, in which no supra-anal opening is present, that is to 
say, where the edges of the mantle do not at all coalesce. In our state, we have 
only two such forms, Margaritana margaritifera (See fig. 6, p. 289) and Quadrula 
(Rotundaria) tuberculata, and in each of them, we have a large, single, undivided 
anal opening. This feature was known previously, and in my opinion is rather 
significant, as showing the preservation of a primitive condition. 
In all other forms, there is a connection of the margins of the mantle separating 
a supra-anal from the anal opening. Yet in a number of species this connection 
is very short, and sometimes deciduous; or it may be torn easily, or may be oc- 
casionally altogether absent. I have observed specimens with this connection 
missing in the following species: Quadrula subrotunda, Q. rubiginosa, Q. coccinea, Q. 
obliqua, Q. undulata, Pleurobema esopus, Unio crassidens, and U. gibbosus. But 
on the whole, such specimens are rare and in all of these species I have found 
other individuals with this connection present. In some eases it seemed that this 
connection was naturally lacking, while in others it had been apparently torn 
apart by rough handling, which may happen easily, since in all these cases the 
connection is very short. 
In other instances the connection is a little longer, but it is generally situated 
so that the supra-anal is a good deal longer than the connection, and than the 
anal opening. This condition of a rather short connection is found in all species 
not mentioned here as possessing a different conformation. 
In certain forms the connection of the mantle becomes somewhat longer, so 
that it is about as long as the anal opening, and also about as long as the supra- 
anal. I have observed this in Alasmidonta undulata, Symphynota complanata, and 
Anodontoides ferussacianus. 
In still other forms the connection becomes rather long, restricting the supra- 
anal a good deal, so that the latter is distinctly shorter than the connection, and 
hardly longer than the anal opening. These species are Anodonta grandis (fig. 4, . 
p. 289), Anodonta cataracta, and Anodonta imbecillis. 
Finally, in one case, that of Lampsilis parva, I found the supra-anal opening 
entirely closed. Of this I have only three specimens, and in one I think I have 
seen a very small supra-anal opening, while in the other two the inner edges of the 
‘8In rare and individual eases, there is a double connection. Lea mentions a few such instances, and I have observed 
the same thing in Obovaria retusa. I consider this an abnormality. 
