TnK Crawfishes of Wesikrn Pennsylvania. 405 



eral direction (south to north ) of the Monongahela and Youghiogheny 

 Rivers, and was stopped in its progress by the first large cross-valley 

 encountered, that runs in an east-west direction. 



The distribution of C. i//o<^{'//('s is not yet fully understood. Its 

 area in western Pennsylvania somewhat resembles that of C. mpiioni::;a- 

 lensis^ although it goes beyond the limits of the latter in a northerly 

 and northeasterly direction, crossing the Allegheny River in Alle- 

 gheny county, and the Loyalhanna in Westmoreland county. At 

 present, no physical features of the country have been discovered that 

 might furnish a barrier to the dispersal of this species northward, 

 although it is certainly not found in the northern half of western 

 Pennsylvania. Toward the east, the Chestnut Ridge limits this spe- 

 cies. The presence of this species in eastern Pennsylvania further 

 complicates this question. 



^^'ith respect to the river-species, C. ohscunis and prof^iiiqiius, at 

 the first glance, conditions appear simple, the first belonging to the 

 Ohio drainage, the second to the Lake Erie drainage. We have seen, 

 however, that parts of the Ohio drainage, namely, the upper Youghio- 

 gheny system, do not possess C. obsciirus, it apparently finding its 

 boundary where this river is rushing over falls and rapids through the 

 narrow gorge of the Laurel Hill and Chestnut Ridge. This fact 

 clearly proves that at least in the Youghiogheny the dispersal of this 

 species was directed up stream. On the other hand, C. obscunis \% 

 found, together with C. propinqiius, in certain parts of the Lake Erie 

 drainage. Here, I think, we have to deal with a case of stream 

 piracy, the capturing of the headwaters of one system by another one. 

 In our case, we know, that the drainage features of Conneaut Creek 

 and French Creek have entirely changed from what they were in pre- 

 glacial times : but this part of their history does not bear upon the 

 present case, the immigrartion of both river species into these parts un- 

 doubtedly belonging to postglacial times. Thus we are led to be- 

 lieve that quite recent changes of drainage have taken place here, 

 which may be generally explained as a tapping of the drainage of 

 French (and possibly also of Shenango) Creek by Conneaut and Elk 

 Creeks cutting back through the original divide, which was formed by 

 moraines. Together with the waters, certain elements of their fauna 

 were thus deflected. In this respect it needs special attention, that, 

 while C. obsciirits has thus become a member of the fauna of the Lake 

 Erie drainage, the opposite has not taken place with reference to C. 



