Origin of Eskers. — Crosby. 5 
regard of the topography observed in many eskers. In so far as 
the subglacial streams follow closely the axes of the ground val- 
leys, their courses may be regarded as virtually fixed, and de- 
posits formed in their channels cannot fail of obliteration or 
burial when uncovered by the recession of the ice ; and the same 
fate will, of course, be shared by deposits formed in the ice- 
walled canyons in which the ice tunnels frequently terminate. 
In view of these considerations, it is certainly not surprising 
that truly subglacial eskers are not now coming into view 
through the shrinkage of this lake of ice. The piedmont gla- 
cier appears fairly comparable in this respect with the tribu- 
tary alpine glaciers and with valley glaciers in general, includ- 
ing the Muir glacier and other ice-streams tributary to Glacier 
bay. 
The only features suggestive of eskers yet noted in the de- 
tailed studies of the Muir glacier are the deposits described 
by Prof. Wright* as formed in certain ice tunnels near the 
thin, debris-covered margins of the glacier. These tunnels 
were abandoned by the subglacial streams which made 
them and subsequently filled by the sliding in of the supergla- 
■cial detritus through holes in the roofs. Prof. Wright says, 
"In numerous places the roof of this tunnel (which is 25 to 
30 feet high) has broken in, and the tunnel itself is now desert- 
ed for some distance by the stream, so that the debris (which 
overlies the ice to a depth in some places of 15 to 20 feet) is 
caving down into the bed of the old tunnel as the edges of ice 
melt away, thus forming a tortuous ridge, with projecting 
knolls where the funnels into the tunnel are oldest and largest. 
At the same time the ice on the sides at some distance from the 
tunnel, where the superficial debris was thinner, has melted 
down much below the level of that which was protected by the 
thicker deposit ; and so the debris is sliding down the sides as 
well as into the tunnel through the center. Thus three ridges ap- 
proximately parallel are simultaneously forming-one in the mid- 
dle of the tunnel and one on each side. When the ice has fully 
melted away, this debris will present all the complications of 
interlacing ridges with numerous kettle-holes -and knobs char- 
acterizing the kames (eskers) ; and these will be approximatc- 
ly parallel with the line of glacial motio n. The same condition 
• The Ice Age in North America, p. 62. 
