386 ON THE NATURE AND ORIGIN OF 



among the eskers of Maine, as well as those of Sweden. At 

 one point on the Antarctic glacier were two huge parallel 

 crevasses probably thousands of feet deep. 



Valley Crevasses. — One peculiar crevasse noted by Shackle- 

 ton was formed lengthwise in a valley near Mt. Nansen; unlike 

 those on ridges, this crevasse was closed at the top but was 

 located because of a range of ice cliffs along the valley. As 

 it was formed through the downward sag of the ice sheet, it 

 would probably be open at the bottom and become a receptacle 

 for eroded debris. These valley crevasses may account for 

 many of the eskers lying parallel with the rivers. These 

 only maj^ be termed subglacial channels until the increasing 

 temperature dissolved the ice and opened the crevasse to 

 the light of day. 



Valley crevasses have been seldom seen in Arctic and 

 Anarctic regions for plainly obvious reasons. An ice sheet 

 crossing a ridge would crack open at the top; an ice sheet 

 crossing a valley would crack open at the bottom and there- 

 fore be invisible. The one mentioned by Shackleton became 

 known only because the ice on one side of the crack sank 

 below that on the other side and left a row of ice cliffs 20 feet 

 high. 



Crevasses formed on watersheds may be forced down 

 the next slope, but the tension being relieved, no more crevasses 

 may form for some time on the same ridge. In Maine some 

 eskers are seen on hillsides far above a river valley. Having 

 evidently been formed on the ridge to the rear it had only 

 time to move slightly in its ice-walled channel when the 

 ice sheet melted. Not Vjeing in or near the bed of a stream 

 the esker remained undisturbed. 



Eskers in valleys, such as those near Hebert and Clyde 

 Rivers in Nova Scotia, must be placed in a different class, 

 owing to a distinct difference in level and position. The 

 original cracks, coinciding as they do with the course of 



