204 BULLETIN OF THE 
In some cases, notably in the somewhat continuous kame belt which 
extends up the Genesee valley from its mouth to its source, we are 
compelled to believe that the stream flowed beneath ice which probably 
had, at least over a part of the path followed by the current, a depth of 
half a mile or more. It is impossible to believe that the water supply 
of this stream could have been furnished by the descent of-the fluid 
from the surface through crevices to the bed rock, for owing to the ease 
with which ice moves under pressure the weight of the superincumbent 
materials would make the maintenance of such passages for the descent 
of the waters impossible. We are therefore driven to the conclusion 
that these subglacial channels occupied by the rivers which flowed upon 
the bed rock were fed from the melting which occurred at the base of 
the glacier. 
The foregoing considerations make it appear likely that the water 
lying at the base of a continental glacier would be subjected to the very 
great pressure of the overlying ice. With ice having a depth of three 
thousand feet this pressure would amount to near ninety tons to the 
square foot. 
It appears tolerably evident that, notwithstanding the pressure of 
the superincumbent glacier, the water which flowed beneath it passed 
through continuous arches leading from the interior portions of the ice 
field towards its margin. We may, therefore conceive that this water, 
for doubtless it had to wrestle with the ice for passage, moved irregu- 
larly towards the margin of the glacier, making its way in a violent 
manner by the obstructions which it encountered. The position of 
the kames of serpentiform outline which mark the position of these 
ancient subglacial streams clearly indicates that the course of the ice- 
covered waters was liable to much change; as is shown by the eskers, 
their paths shifted in a somewhat sudden manner from one side of the 
valley to the other. Near the margin of the ice where the depth of 
the sheet would permit the formation of crevasses which for a time at 
least would remain open, the erratic movements would doubtless be 
greater in amount than in the sections where the ice was so deep as to 
prevent the formation of fractures. We may fairly conceive that the 
motion of these subglacial streams, urged as they were by a pressure of 
many tons to the square foot, would, at least near the margin of the ice 
sheet, take place with paroxysmal violence. Quantities of the débris 
which came in the path of these vigorous currents would doubtless be 
urged laterally away from the axis of motion of the glacier, or borne up 
from near the bed rock into higher lying parts of the ice. Each con- 
