14 SURFACE GEOLOGY. 
channels were deposited materials gathered by the streams from the 
melting glacier. By the low water of winter layers of sand would be 
formed, and by the strong currents of summer layers of gravel, often very 
coarse, which would be very irregularly bedded, here sand and there 
gravel accumulating, and without much order interstratified with each 
other. Sometimes the melting may have been so rapid that the entire 
section of a kame may show only the deposition of a single summer, 
which would then be very coarse gravel without layers of sand. When 
the bordering and separating ice-walls disappeared, these deposits re- 
mained in the long ridges of the kames, with steep slopes and irregularly 
arched stratification. Very irregular short ridges, mounds, and enclosed 
hollows resulted from deposition among irregular masses of ice. 
The glacial rivers which we have described appear to have flowed 
in channels upon the surface of the ice, and the formation of the kames 
took place at or near their mouths, extending along the valleys as fast as 
the ice-front retreated. Large angular boulders are sometimes, but not 
frequently, found in the kames, or upon their surface. They appear to 
have been transported by floating ice. Their rare occurrence forbids the 
supposition that these deposits were formed in channels beneath the ice- 
sheet, from which many such blocks would have fallen upon the kames. 
The necessity of referring the formation of the gravel ridges to glacial 
rivers became apparent during the exploration and study of our modified 
drift in 1875; and in August, 1876, this was announced in a paper “On 
the Origin of Kames or Eskers in New Hampshire.”* In the revised 
edition of Geikie’s Great Ice Age, published in London in the winter of 
1876-77, this distinguished glacialist retracts his former opinion that the 
kames were heaped up by marine currents, and attributes their formation 
to sub-glacial rivers.| This may be the true explanation in some cases, 
for such rivers probably existed through the glacial period; but more 
commonly it would appear, as already shown, that the kames were depos- 
ited at the final melting of the ice-sheet in channels formed upon the 
surface of the ice. 
% Proceedings of American Association for the Advancement of Science, vol. 25. 
+ Great Ice Age, second edition, revised, pp. 217, 239, 243, 469,478, etc. By page 414 it appears that this theory 
was first proposed by Mr. D. Hummel of the geological survey of Sweden, in 1874; and on page 415 allusion is 
made to a recent paper by Dr. N. O. Holst, also of Sweden, in which the kames have been explained in the 
same manner as in this chapter. 
