GL.ICIAI. GKOLOGY OF XORTH GREENLAXD 8O9 



surface of the ^-lacier there welled up a huge spring (Fig. 33). 

 The water shot up not less than ten feet above the bottom of the 

 basin whence it issued. The water was intensely red, owing to 

 the ])resence of the flour of red rock which the ice had ground 



Fig. 33. Spring on the first glacier in the valley above Godhavn, Disco. 



up. The upper part of the ice was nearly free from debris, and 

 the water must have risen from a lower horizon in the ice. The 

 inconstant character of englacial drainage is shown by the fact 

 that two months later, at the same site, there was no suggestion 

 of a spring. The water had found some other avenue of escape, 

 though the basin and the opening in its bottom were easily 

 found. The opening was about five feet in diameter, and for a 

 distance descended nearl)' vertically. 



Karnes and eskers. — No esker was seen in Greenland, nor was 

 any process observed which would at any time result in the for- 

 mation of an esker. 



Except in one situation, namely, on the north side of Olriks 

 Bay, no kame was seen. Here there were some kame-like hills 

 on a surface which had been abandoned by the ice, but no 

 process was seen in operation along the margin of the ice at any 

 point which seemed to throw special light on the origin of this 

 class of drift hills. 



The surface over which the ice has retreated. — In a number of 

 places, surfaces were seen which have been, within relatively 



