1913-] ORTMANN— THE ALLEGHENIAN DIVIDE. 373 



After S. tappaniana had reached the James drainage (it has not 

 been found in the Roanoke, but only the headwaters of this are 

 known), it had a chance to spread on the Atlantic side and to attain 

 its present wide range, exactly as the majority of the Atlantic forms, 

 favored by the same causes. It always remains a small-creek-form, 

 but just in these small creeks the best opportunities were given to 

 cross from one system into the other. 



Anciilosa dilatata and carinata. 



Ancnlosa carinata is the Atlantic form and is known to me from 

 the Roanoke to the Susquehanna, where it goes up into New York 

 state. In this restriction (not being found in the Delaware and be- 

 yond) it is dififerent from Strophitiis and Symphynota tappaniana, 

 which go to New England. West of the divide we have Ancnlosa 

 dilatata, first of all in the same region where Symphynota tappaniana 

 is found (Greenbrier and New rivers) ; but in addition it is also 

 in the upper Monongahela drainage, in Tygart and Cheat rivers ; 

 in the latter it goes down below the canyon, as far as Cheat Haven, 

 Fayette Co., Pa., and further it is found in West Fork River. Re- 

 markably enough, it is absolutely absent in the upper Youghiogheny, 

 although the conditions appear favorable for it. 



With exception of these localities in the Monongahela drainage, 

 the distribution fairly well agrees with that of Symphynota tap- 

 paniana, and we won't make a mistake if we advance the same expla- 

 nation for it : stream capture on the part of certain Atlantic streams 

 (Roanoke and James), which robbed the water and the fauna of 

 certain parts of the old New River drainage. Thus only the pres- 

 ence of this form in the Tygart and Cheat needs explanation ; into 

 West Fork River it undoubtedly got from the Tygart. 



The headwaters of these rivers interlock in a very complex way 

 in Pocahontas and Randolph Cos., W. Va. (see PI. XII.), and there 

 is no objection on general principles to assume that there has been 

 intercommunication of these rivers by stream capture. But condi- 

 tions are rather obscure in this region and have been so little inves- 

 tigated from a physiographical standpoint that it is practicallv im- 

 possible to draw any positive conclusions as to the history of the 

 development of the headwaters of these systems. 



