CROSBY: ORIGIN OF ESKERS. 
399 
the restraining mountains; that, if the Pleistocene ice sheet had 
been effectively crevassed (that is, from top to bottom) in its pas¬ 
sage over the vast peneplain tracts of the glaciated areas, the cre¬ 
vasses would have been gradually obliterated during the extremely 
slow cessation of the movement, anterior to the period of final abla¬ 
tion of the now stagnant ice sheet when eskers were formed; that 
even on the moraine- and forest-covered outer zone of the Malaspina 
glacier, crevasses are practically wanting, the streams originating 
within this zone through surface melting being superglacial and so 
continuing as they flow down the terminal ice slope; and that glacial 
potholes are by no means conclusive proof of the existence of 
moulins or crevasses in the ice sheet at these points. Add to all 
these considerations the j^i’obability that such crevasses as might 
possibly appear in the ice sheet and survive the cessation of its 
motion would become clogged and closed by superglacial drift 
falling and washing into them, and it will be seen that the burden 
of proof fairly rests upon those who find in the crevasses of the 
Pleistocene ice sheet evidence of the non-existence of important 
superglacial streams. 
During the period of growth and maximum development of the 
ice sheet, the entire volume of the drift must have been englacial, 
and through the processes of shearing and overriding it must have 
tended constantly to rise to higher and higher levels in the ice. As 
opposed to this elevation of the englacial drift, wm have only the 
supposed fact that the velocity of the ice increases upward from the 
bottom, tending to bear down the upward-sloping shear-planes. 
Observation shows that this is true of alpine glaciers of high surface 
gradient, and doubtless it would be true, in diminishing degree, for 
the lower gradients of an ice sheet, if the viscosity of the section 
were uniform. It is in the highest degree probable, however, that, 
owing to the outward flow of the terrestrial heat or the rise of the 
isogeotherms, the temperature of the ice sheet during and after its 
prime increased downward, being highest at the bottom. This 
accords with I^ansen’s observations on the Greenland ice cap, pre¬ 
viously quoted. Now the mobility of the ice, or its tendency to 
flow through the differential melting and freezing of its component 
granules, is a function of the temperature, inasmuch as it must 
increase with the temperature and reach its maximum at the melting 
point, which, as previously noted, may be lowered as much as one 
