(O NEW YORK STATE MUSEUM 



The main valley was only partially filled by glacial deposits, and 

 was reoccupied as the main drainage channel on the retreat of 

 the ice. The main change effected was the shifting of the posi- 

 tion of the divide between the easterly flowing stream in one, and 

 the westerly moving drainage in the other end of the valley. The 

 present divide is at Rome, is of' the most trivial description and is 

 composed of glacial deposits which manifestly could not have- 

 formed a preglacial divide. The preglacial col was in all proba- 

 bility at Little Falls, as urged by both Chamberlin and Brigham. 

 Here the valley is narrowest, here is the most resistant rock mass 

 anywhere in the valley, brought up on the west side of the fault, 

 and here the drainage adjustment of the long, preceding time of 

 wear would inevitably locate the divide. 



After the ice had disappeared from the Mohawk valley but was 

 still blocking that of the St Lawrence, the waters of the Great 

 lakes went to the sea by the Mohawk valley route, and this great 

 rush of water must have been very efficient in cutting away the 

 rock obstruction at Little Falls. On the other hand, it is obvious 

 that at this time the divide could by no means have had the hight 

 of the present valley walls, nor even that of the pre-Cambrian 

 surface at the fault line (600 feet), the latter being more than lOO' 

 feet above the present divide at Rome. Chamberlin has suggested 

 that the outer and wider gorge at Little Falls was cut during 

 Lnterglacial times, and this is very probable; at all events, it is 

 certain that the inner gorge represents the total amount of cutting 

 since the ice retreat. This interglacial erosion of the col, to- 

 gether with the heavy drift deposition about Rome, shifted the 

 divide to that point, so that between the two points there is now 

 easterly, where formerly was westerly, flowing drainage. 



Brigham has argued that the preglacial course of West Canada 

 creek was by way of Holland Patent, where now is a broad, open 

 valley occupied by a small stream. Certain it is that from Pros- 

 pect to its mouth the stream is not in its old channel, and that 

 from Prospect to below Trenton Falls it is not in an old channel of 

 any sort. 



From Middleville to Herkimer the course of a small preglacial 

 stream is apparently followed, whose source was at Middleville, 



