ESKERS IN THE VICINITY OF ROCHESTER, NEW YORK. 193 



Streams. From the ice ridges so formed material may slide down 

 irregularly to form kames. 



This manner of formation would explain the general absence ot 

 till from esker surfaces. Large boulders on "esker surfaces and 

 within the gravels may have been floated in on icebergs, may have 

 fallen from the steep ice walls at the side of the deposit, or may 

 have been incorporated in the base of the gravels from the basal ice 

 as the deposit was slowly being let down to the ground. 



The chaotic stratification, evidences of sliding, and "pell mell" 

 structure are explained by the irregularities induced in getting the 

 ridges down upon the ground. 



Segmentation is due to the occurrence of rapids in the super- 

 glacial stream with consequent lack of deposition, or to distribution 

 of that portion of the ridge in being let down upon the ground. 



This method of origin also accounts for eskers resting directly 

 upon the till, and the absence of till erosion beneath the esker and 

 back of the point of origin of the esker. 



Objections. (1) Insuperable difficulties stand in the way of 

 getting the ridge down on the ground according to the method 

 postulated. Porosity of ice would drain ofif the water from the 

 sluggish stream's course before the esker deposit could be let down 

 on the ground. Russell's observations on the lakelets of the Malas- 

 pina glacier show that they are frequently drained even on stagnant 

 ice, their deposits becoming cones due to the protection afforded to 

 the ice beneath from melting. There is no instance recorded of the 

 material on the bottom of these lakes getting down to the ground 

 by this process. 



(2) Eskers occur that are 100 miles long or more, it is doubt- 

 ful if that expanse of glacial ice can exist without crevasses which 

 would limit length of esker ridges. The extensive crevassing of the 

 margin of the Greenland glacier in the zone of esker formation 

 reinforces this point. Further streams of the length required here 

 would be apt to have their upper courses in the region of neve and 

 not on debris covered ice. These long eskers probably could not be 

 formed close to the ice margin during its retreat for the rapidly 

 melting ice would not permit of the stream coming to grade, a con- 

 dition essential for aggradation. 



