428 NEW YORK STATE MUSEUM 



Lyons; between Montezuma and Syracuse; and along the Seneca 

 river. Those with direct east and west course were occupied by the 

 latest ice border drainage but apparently were not wholly produced 

 by it. These features appear on the Macedon, Palmyra, Weedsport 

 and Syracuse sheets, and a suggestion of them is shown in plates 

 9-12. The channels all lie in Salina shales and possibly they have 

 some genetic relation to the erodible nature»of the rocks. 



The existence and location of these low, continuous passages are 

 strikingly emphasized by the remarkable windings of even the larger 

 streams, Seneca river for example. The northward turns which this 

 wayward stream makes east of Savannah, and more strikingly from 

 Cross lake (an open tract) around by Baldwinsville and south to 

 near Onondaga lake, must have been found open or the stream 

 would have taken the direct eastward passages that are almost as 

 low today even without any postglacial erosion. A smaller illustra- 

 tion is found in the case of Ganargua creek east of Palmyra, where 

 it deserts the open glacial stream course and wilfully turns north 

 around by East Palmyra in a constricted and uninviting pass, and 

 repeats the act with less excuse northeast of Newark. 



One suggestion for these open passages, which were certainly left 

 open by the ice removal and were not cut by postglacial erosion, is 

 that they were made by. subglacial drainage, either under free flow 

 or under hydraulic pressure. This seems reasonable for south- 

 leading passages like those northeast of Clyde, and even for north- 

 leading channels of gentle grade, like the valley of Dead creek in 

 plate 10; but it is not satisfactory for the east and west passes which 

 were transverse to the ice movement. 



Another suggestion is that the passes were cut by glacial drainage 

 of an earlier ice sheet, with modification by the subsequent inter- 

 glacial erosion. It seems probable that the complex history of the 

 region may involve such episodes and activities; or that perhaps the 

 oscillations of the last (Wisconsin) ice sheet were sufficiently exten- 

 sive to produce the phenomena. The difficulty under this theory is 

 to explain why the drumlin-making work of the ice did not rub the 

 channels full of drift. This difficulty is of the same kind, however, 

 as the absence of drumlins over interdrumlin tracts. In the case of 

 the deep channels around the Syracuse island masses it might be 

 suggested that possibly during the latest stage the channels were 

 occupied by stagnant ice over which the drumlin-forming layers 



