374 ORTMANN— THE ALLEGHENIAN DIVIDE. [April 18, 



But it is highly interesting to notice that the distribution of Ancu- 

 losa dilatata in the Greenbrier on one side, and in the Tygart and 

 Cheat on the other, points to stream capture in this region, and the 

 theory is suggested that the Monongahela drainage encroached upon 

 and robbed the Greenbrier drainage. The opposite way is not pos- 

 sible on account of the limitation of this form northward, and this 

 also speaks against the possibility of passive transport. If this as- 

 sumption is correct, it also explains the fact that the Youghiogheny, 

 which also heads in the same general region, did not receive this 

 species. The upper Youghiogheny flows in a high synclinal valley, 

 is more nearly an old consequent river than, for instance, the upper 

 Cheat, which has cut down way below the level of the upper Yough- 

 iogheny. Thus it is impossible that the latter ever robbed the Cheat, 

 capturing its fauna ; rather the opposite has happened, and probably 

 is happening now. 



The Atlantic form, Ancid osa carinata, after having reached the 

 Roanoke and James, and after having become established on the 

 eastern side, had the same tendency to spread as the rest of the 

 Atlantic forms. But it did not go so far as many others, reaching 

 only the Susquehanna drainage. In this case northward migration 

 probably was due to the crossing over divides (by stream capture) 

 in the mountain region. Ancid osa is a shell characteristic for rough 

 water in mountain streams and goes possibly farther up than any 

 other of the forms discussed here. In the lowlands, it has never 

 been found, and it is also less frequent in the Piedmont section of 

 the streams, although present there. Thus its migration very likely 

 took place chiefly within the mountains, and I think that its limited 

 range northward is due to this fact. 



The genus Ancid osa is represented in the uppermost Tennessee 

 drainage by the species Anadosa gibbosa, which is to a certain 

 degree related to the dilatata-carinata-group. In fact, the Tennessee 

 drainage is the only other region where relations of this are found. 

 This makes it clear that New River must have received its Anculosa- 

 stock from the upper Tennessee. It is hard to say how this was 

 accomplished. We have seen (p. 352 f.) that stream capture was rare 

 in this region ; at any rate, if there was any, it was rather in the 



