442 MEMOIRS OF THE CARNEGIE MUSEUM 



in front of this morainic system. The fall of the creeks running northward to Lake 

 Erie from this divide is much more considerable than that of those running south- 

 ward, and thus it is clear that erosion on the northern slope must have been more 

 efficient than on the other side. The consequence is that the tributaries of Lake Erie, 

 at least some of them, have worked back through the original divide, and have cap- 

 tured parts of the original Postglacial drainage of the Ohio. This is most evident 

 (see PL XLIII) in the cases of Conneaut and Elk C'reeks, and it is just in these 

 creeks that I found C. obscurux associated with (.'. pnijiiiujuus/'' while in Walnut 

 Creek, which has apparently not entirely cut through the original divide, C. oh- 

 seurus is not found. 



Thus it is possible that the presence of ('. ohscvrus in the Lake Erie drainage is 

 due to stream-piracy. Both species, ('. (ihscuruf! and j)roj)inquus, are associated here, 

 but it seems that they are antagonistic to each other to a certain degree. In the 

 tributaries of Conneaut Creek I found C. propinquvs exclusively, while (Jonneaut 

 Creek itself contained l)oth, but C. obscunts prevailed, and it appears as if the latter 

 had driven out the other species, which took refuge in the smaller tributai-ies. 



We might expect to (jbtain S(Miie light upon the question, whether C. obscurns 

 reached the Lake Erie drainage in consequence of stream-piracy or by the help of 

 the canal, by the analogy offered in the Genessee drainage, but conditions seem to 

 have been not entirely identical here. The type locality of C. obscurvs (see PI. 

 XLII, Fig. 3) is the Genessee River at Rochester, Monroe County, New York, where 

 this species also is found associated with C. pro^nnquns. Mr. W. 1'. McConnell has 

 discovered C. obscurns in the upper Genessee drainage near Ulysses, Potter County, 

 Pennsylvania. The material consists of numerous males of the first and second 

 form and of females, and there is not the slightest question that this is the true C. 

 obscurus, no trace of C. propinquus being present here. How did this species get 

 from the Alleghany drainage into that of the Genessee ? 



The drainage of the Genessee River lying entirely within the glaciated area, this 

 must have happened in Postglacial times. Fairchild (1896, p. 423) has shown that 

 during the recession of the ice the Genessee basin was occupied by a lake, which had 

 its outlets in different directions successively, draining either to the Susquehanna or 

 to the Ohio. He distinguishes ten stages, and the sixth was the last in which the 

 water flowed to the Susquehanna; in the seventh and eighth stages Genessee Lake 

 became connected with Lake \Varren, which drained to the west into the Missis- 

 sippi basin (but not into the upper Ohio), and finally the St. Lawrence drainage was 



'■'The sources of Elk Creek are in a tamarack swaiDii, which also drains to the south, to French Creek, so that some 

 kind of a direct connection may be present. I have not visited this swamp. 



