CROSBY : ORIGIN OF ESKERS. 
403 
higher gradient, in cutting its canyon below those of its lateral 
tributaries and making of the latter hanging valleys, in which the 
aggrading will necessarily be unimportant. Stone says (’99, p. 
324) in this connection, ^‘From whatever point of view we look, 
the difficulties are immense in accounting for the branchings of the 
rivers of the ice sheet, their directions and their relations to the 
relief forms of the land, the nature of their sediments, etc., on the 
theory that we are dealing with subglacial streams alone.” 
Double and retieidated eskers. — That the superglacial deposits, in 
the case where the bordering ice disappears before the underlying 
ice, afford a complete and satisfactory explanation of the double 
eskers has been noted ; while as subglacial deposits they are essen¬ 
tially inexplicable, whether the adjacent and parallel tunnels which 
must be postulated are regarded as contemporaneous or successive 
in their formation and aggrading. Similarly, the reticulated eskers 
and kames are most easily explained as superglacial deposits, repre¬ 
senting a delta-like branching of the superglacial stream at points 
where it was approximately base-leveled and the surface of the 
bordering ice had been reduced nearly to its level by general abla¬ 
tion. Another explanation would be that the deposit, formed in a 
lake-like expansion of the river, was split up into a network of ridges 
through the unequal melting of the subjacent ice while it was being 
let down upon the ground. It is an essential part of the supergla¬ 
cial hypothesis, that, while through subfluvial and basal melting the 
more or less perfectly aggraded deposits may be let down on terra 
Jirma without being seriously disordered, exceptions must occur of 
such character and frequency as apparently to explain every aber¬ 
rant phase of esker formation. 
Stone shows (’99, p. 465) that the reticulated eskers, like the 
broad eskers and esker terraces, are most readily and satisfactorily 
explained by supposing that the subglacial streams became, locally 
at least, superglacial. Again, he says (’99, p. 299) that subglacial 
streams crossing hills and ridges are inconsistent with the existence 
of crevasses which might divert the water, which bides its time and 
at the first eligible transverse crevasse steals off sidewise toward the 
lower ground.” Also, in the discussion which follows the tunnels 
are assumed and not accounted for. 
Topographic relations of eskers. — The main points under this 
head have been duly considered, and it remains simply to note once 
