14 . TJic American Geologist. •^"'^'' i^*^-- 
above and subglacial streams below, competing for the en- 
glacial drift, of which, chiefly, eskers must be formed. 
Undoubtedly, the very latest direct work of the ice-sheet, 
before its motion finally ceased, was the building of the drum- 
lins. In the lee sides of a few drumlins a limited amount of 
gravel is interstratified with the till, the product, possibly, of 
superglacial streams falling through the ice at these points, 
since crevassing would, probably, occur as early where the ice 
bends over the summit of a drumlin as anywhere. At the same 
time the ice must have tended to draw away from the lee slopes 
and leave the vacant spaces in which gravel could be deposited 
by superglacial, but not possibh^ by subglacial streams. We do 
not discover, however, any isuch general intercalation of washed 
drift with the ground moraine, either in drumlins, or else- 
where, as to suggest that crevasses were a common feature of 
the ice just before it ceased to move; and of course, none could 
be formed after it became stationary. Again, it appears prob- 
able that when the ice was thick enough to override drumlins 
one hundred to several hundred feet in hight, its thickness was 
too great to permit extensive crevassing, especially in view of 
the fact that the thickness essential to flow must increase rap- 
idly wdth diminishing slope and increasing roughness of the 
ground or basal friction, and the farther fact that even alpine 
glaciers in relatively smooth and unobstructed valleys of high 
gradient, and seldom exceeding 500 to 1000 feet in thickness, 
are but little afifected by crevasses, except in cascades, and 
laterally where they feel the friction of the valley walls. In 
this connection w^e may profitably note once more that the mar- 
ginal portion of the ^Nlalaspina glacier, with a thickness of at 
least one thousand feet and a steep frontal slope, at the base of 
lofty mountains and feeling the thrust of powerful alpine 
glaciers, and absolutely nothing to hinder its free motion sea- 
ward across the sloping coastal plain, is practically motionless. 
These considerations dearly point to the conclusion that the 
Pleistocene ice-sheet on the highly dissected and rough pene- 
plain surface of the greater part of the glaciated area, with 
only extremely low average gradients in any direction, and 
then often either to the northward or transverse to the glacial 
movement, must have been at least two thousand feet thick 
when it ceased to move, and too thick in general for effective 
crevassing, or the formation of crevasses extending from the 
top to the bottom of the ice. 
