CROSBY: ORIGIN OF ESKERS. 
383 
Probable Status of the Ice Sheet during the Formation 
OF Eskers. 
That eskers were formed by glacial streams in, on, or under the 
ice sheet and during the waning stage of glaciation, w^hen this 
border area was rapidly disappearing through superglacial and sub¬ 
glacial ablation, are propositions accepted by all and requiring no 
argument here. Further, we may postulate, with Davis and other 
able students in this field, the stagnant condition of this marginal 
zone during the esker-forming period. In fact, a stationary ice mar¬ 
gin is a condition highly favorable, if not absolutely essential, to 
every theory of esker formation ; and, as Davis has insisted, it is 
demanded by the highly irregular and fragmentary border of the ice, 
of which we have conclusive proof in the distribution and outlines 
of the deposits formed upon or against it and in the absence of evi¬ 
dence of glacial thrust. 
In my paper on englacial drift (’96) , I have accepted and elab¬ 
orated IJpham’s idea that over the vast plain or peneplain tracts of 
the glaciated area, the ice sheet was developed primarily by accum¬ 
ulation and not by invasion, existing at first as a sedentary ice cap, 
which gradually acquired motion as it gained in thickness and also 
as it was progressively overridden by the older and already active 
ice to the northward. Again, I see now, as five years ago, no escape 
from the conclusion, also first enunciated by Upham, that, during 
the periods of ice accumulation and of maximum glaciation, the 
entire volume of the drift, including both the preglacial detritus and 
all that due to glacial rending and abrasion, was englacial, or firmly 
frozen in the basal portion of the ice. This conclusion is absolutely 
demanded by the universal rectilinear striation of the bed rock sur¬ 
face, and is inconsistent, so far as I can see, with no established 
facts. 
I have also shown in the paper cited above, that the observations 
of Chamberlin and others in Greenland and elsewhere indicate that 
an ice sheet, in its progress across even a comparatively smooth 
surface, is subject to oblique shearing movements which tend to 
transfer the englacial drift to higher levels in the ice. I hold now, 
even more strongly than when writing the paper on englacial drift, 
that the tendency to the elevation of the drift in the ice must have 
