INTERPRETATION 239 



ice advances in this region has been discovered so far, that hypothesis 

 must still be held as a working hypothesis in future studies. 



ICE-BORN STREAM-EROSION HYPOTHESIS 



Wherever positive evidence could be discovered in this area the ero- 

 sion of streams associated with the melting of the Wisconsin ice sheet 

 was found to be very slight, in most cases amounting merely to a notch- 

 ing of the drift deposits. The ice-born streams were depositing rather 

 than eroding profoundly. That the peculiarities of the valleys described 

 above are not in any important degree due to erosion by streams flow- 

 ing at the closing stages of the last ice advance is demonstrated by the 

 fact that practically all the valleys contain undisturbed morainic deposits 

 which could not have escaped destruction if there had been marked 

 erosion by ice-born streams. 



This form of evidence does not, however, eliminate the possibility of 

 water erosion connected with the advance of the Wisconsin ice; but in 

 certain cases, like the Tiougnioga, Cayuta, and Willseyville valleys, the 

 presence of older buried tributaries does eliminate this hypothesis for 

 these par ticular cases, and, since they are analogous to the others, 

 weakens the hypothesis that they are due to water erosion during the 

 advance of the Wisconsin ice sheet. 



The great number of instances, extending in all directions, and of all 

 sizes, would call for a vast amount of water erosion under very variable 

 conditions. I find it exceedingly difficult, for example, to postulate 

 any conceivable set of conditions by which ice-born stream erosion could 

 possibly cut the complex of channels near Elmira and Horseheads, de- 

 scribed above. Moreover, it does not seem possible that water erosion 

 of overflow or marginal type could form the doubly flaring condition of 

 Texas Hollow valley, nor the Chemung valley west of Elmira, nor a 

 score of other similar valleys. 



Even if several ice advances have been involved, and a large number 

 of ice front positions are postulated to account for these peculiar valleys, 

 the efficiency of glacially supplied streams to form such topographic 

 features is questionable. While the ice was advancing and receding 

 from this hilly region its front could not at any time have long stood 

 on the divides, and while it did, judging from the history recorded at 

 the closing stages of the last ice advance, deposition would seem to have 

 been the rule, not erosion. When the ice fronts stood north of the 

 divides there were marginal lake conditions, and the inefficiency of such 

 a river as the Niagara or the Saint Lawrence to form a rock gorge does 

 not lend much support to the hypothesis that the overflows of the 



