2 The American Geologist. "^"'-^'i ^^^-■ 
ice-sheet- And in reopening this cHscussion. as one of the still 
unconverted adherents of the superglacial theory, my main 
purpose is to show that under normal conditions the deposits 
of gravel and sand formed in a superglacial channel may be 
let down upon terra firma without obliteration and without 
loss of the distinctive features of an esker. In this connec- 
tion I appeal especially to basal melting and to the principle, 
first enunciated by Upham, that the ice beneath a supergla- 
cial river will be melted downward and the channel deepened by 
the water which saturates and flows through the deposit of 
gravel and sand constituting the embryo esker. But I also 
hope to reinforce these principal arguments by others which, 
if less cogent or less vital, are yet essential to a complete theory 
of eskers. 
It is recognized by all that eskers must represent the wan- 
ing stage of the ice-sheet ; and probably the extent of basal 
melting during this stage is generally underestimated ; but 
still few will question that it is to the surface melting or abla- 
tion of the ice that we must look for the main source of water 
for the glacial streams, whether superglacial or subglacial. 
It is undoubtedly true that deposits which may be classed 
as eskers have been formed under a variety of conditions : in 
superglacial channels : in subglacial channels ; in ice-walled, 
earth-bottomed canyons, open to the sky ; and witli or with- 
out the active agency of water. And all the theories are, no 
doubt, essential to a complete explanation of eskers, the main 
question now being as to their relative importance, or as to 
which best accounts for the more typical and important exam- 
ples and has. therefore, the best claim to be regarded as tJie 
theory of eskers. 
Evidence of Existing Glaciers atid Ice-Sheets. 
It must be conceded at the outset that nowhere have ob- 
servations having any obvious bearing upon this problem been 
made upon existing icemasses w^liich realize, even approxi- 
mately, the essential conditions of the Pleistocene ice-sheet as 
it existed upon the plain country remote from mountainous 
tracts, where, chiefly, eskers are now found. In Alaska, we 
have, besides the Malaspina glacier, only aJpine glaciers in 
lofty mountain valleys of high gradient ; and the ]\Ialaspina 
