CROSBY : ORIGIN OF ESKERS, 
401 
another, indicate, as recognized by Stone, the formation of a normal 
esker in a narrow channel, followed by a considerable expansion of 
the channel, permitting the deposition of the finer material of the 
bordering plain. This broadening of the channel might, in the case 
of a superglacial stream, be attributed to the melting back of the ice 
walls, or, better, to localized surface ablation due possibly to water 
saturating drift which has accumulated on the ice through previous 
ablation. So great is the difficulty of explaining these broad 
deposits by deposition in subglacial channels that Stone (’99, p. 
444-448), after a careful study, reaches the conclusion that through 
the obstruction of the tunnel, perhaps by the sagging of its roof, the 
subglacial stream rose through crevasses with its burden of coarse - 
gravel and became, for a longer or shorter distance, superglacial. 
Chamberlin, on the other hand, suggests that an ice tunnel became 
locally roofless, and then by lateral recession of its walls, the ice 
canyon became a lake. But it is obvious and generally conceded 
that superglacial streams afford by far the simplest and easiest 
explanation of these and all other lateral expansions of eskers, 
including the not uncommon terminal expansion, where the esker 
merges with the frontal plain or delta. The only serious questions 
raised by any one are as to the possibility of persistent superglacial 
streams with a sufficient volume of formerly englacial drift within 
their reach ; and these difficulties disappear before the conclusions 
to which we have been forced that effective crevassing was of rare 
occurrence and high-level englacial drift abundant. 
Concerning the more usual widenings of the eskers of Maine, 
Stone says (’99, p. 415), ‘^When within about 75 miles of the coast, 
every few miles enlargements of the ridges are found which have 
are little tables only 200 to 300 feet 
wide and two or three times as long. These may be solid or may 
contain one or more shallow kettleholes. Here and there a hum¬ 
mock appears on top of the osar [esker], rising 20 to 40 feet above 
the rest of the ridge, and at these ‘ ’ the ridge is generally 
broader than elsewhere.” 
Varying height of eskers. — Woodworth’s correlation of varia¬ 
tions of height with variations of original breadth of the esker 
deposit and of the ice channel in which it was formed, applies in 
only a limited number of cases. It does not, for instance, account 
for the broad and flat cols or depressions, or for the occasional high 
various forms. Sometimes they 
