8 46 THE AMERICAN NATURALIST. [VoL. XXXV. 
Tennessee River, which may represent the northwestern point 
of our boundary. 
* These limits are : necessarily imperfect, but, nevertheless, 
include at least three-fourths of our species within an area of 
three hundred miles extent, either north and south or east and 
west." By reference to Fig. 4 (after Hayes and Campbell), one 
will see that these limits are practically those of the Appala- 
chian valley, and thus enclose a natural physiographic region. 
Again, Tryon remarks: * Assuming the Ohio River as a 
dividing line, we find Vat ninety-five per cent of all the species 
originate south of it.’ 
So much for these general facts of the distribution of this 
family which are necessary for an understanding of the influ- 
ence of the migration of the divides upon them. 
With the above facts of the distribution of these shells fresh 
in mind, about two years ago I learned from Hayes's (95) paper 
of the view that the upper Tennessee River formerly flowed by 
way of the Coosa-Alabama Rivers into the Gulf. I saw at once 
in this an adequate explanation for the peculiar distribution of 
many of these shells in these two river systems. Since the 
data for this paper have been collected Mr. C. T. Simpson ('00) 
has published similar conclusions, drawn from the Unionidz. 
According to Hayes's and Campbell's ('94), and also Hayes's (99), 
opinion, the Tennessee River above Chattanooga was captured 
by a westward-flowing stream, and thus led away from the 
Gulf by the Coosa-Alabama to the Gulf again by way of the 
Ohio and Mississippi. Thus we see that the region in which 
the maximum development of this family is reached was broken 
into two parts (Fig. 5). The bearing of this upon the shell 
fauna is very marked. 
This family is divided into two sections, and as to their dis- 
tribution Tryon (/c. cit., p. xl) says: “ While the Trypanosto- 
moid forms attain their maximum development in size and 
number in the Tennessee River, they are to a very great 
extent replaced by the Goniobasic forms in the Coosa River, 
which is undoubtedly the metropolis of the latter. The most 
striking genus of each of these groups is absolutely confined 
to the respective streams in which the groups had their origin. 
