Geographical Distribution of Certain Fresh- Water Mollusks. 321 



Tennessee, and to the southern and eastern limits of the Ohio drainage. 

 Here it is that we begin to meet with those forms about whose specific 

 rank there can be little reasonable question, though it is a self-evident 

 fact that nearly or quite fifty per cent, of the so-called " species" of 

 this region, are local varieties of the central types mentioned above. 



In considering the fauna of those streams presenting the greatest 

 number of forms claiming specific rank, we can always separate them 

 into two or more groups having a distinct facies ; and in all cases, one 

 of these groups will be that of tjq^ical Ohio river forms. 



I do not, in this consideration of the matter, refer to special cases of 

 form, such as "oval," "quadrate,'' "wide," etc., terms used by Mr. 

 Lea in his grouping of the TJnionidoi, but to a certain general stamp 

 or character, which belongs to larger groups, holding often many of 

 these "forms." It is impossible for the collector, w^ho has waded 

 through these interminable variations, as well as the streams contain- 

 ing them, not to be struck with the force of this fact, and to have it 

 continually brought before him, as if there was a commingling of 

 faunas, widely enough separated to lead to comparison with different 

 areas and s^^stems of drainage, or to suggest the mingling of species 

 from such systems. In some of the streams, even comparativel}^ small 

 ones, a predominance of these abnormal groups exists ; and there may 

 even be in very small streams, an entire absence of the Ohio t3^pes. 

 In considering the present distribution of these mollusks, w^e find an 

 infinite variety of conditions as we pass from stream to stream, and we 

 discover, as the result of this, many local varieties that doubtless owe 

 their origin to these causes ; but there are other groups that are 

 evidently related to some remoter source. These are those which, 

 in the Unionidcn, may have no separate generic characters, but which, 

 in the Strejoomatidce^ have been stamped as having higher value than 

 that of mere varietal distinction. 



In this connection we have such problems to deal with as the pres- 

 ence of the JJnio spinosus in the Altamaha river alone, at the southern 

 end of the Atlantic slope af the Appalachians, and of the U. collimis in 

 New river of Virginia, on the western slope, far to the north. These 

 are, I believe, the oul}^ spinous species of the family known. Neither 

 has any distribution, yet discovered, beyond the stream in which it 

 was first found. The case of the 31. margaritifera has already been 

 cited. Other cases equally remarkable exist, which it is not necessary 

 to quote here. There is, also, this further fact to be observed; that 

 many streams contain species not found in contiguous ones, a fact 



