218 The American Geologist. April, 1896 
of detritus; and also that this, as he supposes, scanty engla- 
cial drift is now distinguishable from the ground moraine on 
which it rests. Having granted this much, Prof. Chamberlin 
is. apparently, obliged to ground his argument for the essential 
scantiness of the englacial drift upon the assumption, no- 
where explicitly stated, that but little drift was dragged or 
carried by the ice up the stoss -lopes and over the crests of 
the ledges and hills, for this material would obviously have an 
equal or better chance than that worn from these elevations 
of becoming truly englacial. Put the validity of this tacit 
assumption cannot be admitted. The severe glaciation of the 
stoss slopes is against it; and, aside from this consideration, 
it is difficult to understand how. except in valleys trending 
with the glacial movement, any considerable amount of detri- 
tus that was transported or dragged a good fraction of a mile 
or more from its source could help crossing one or more eleva- 
tions. Certainty its course would need to be very devious to 
avoid them. 
Again, Prof. Chamberlin* has given us the important prin- 
ciple that, during the closing stages of the Ice age at least,, 
the surface of the ice-sheet must have been depressed over 
highlands and elevated over valle3 T s, and that, movement be- 
ing determined by surface gradient alone, the ice would flow 
toward rather than from the highlands. This principle would 
thus operate to increase the ground moraine on the hilly 
tracts at the expense of that in the valleys and lowlands: and 
it has occurred to me that perhaps we have here an explan- 
ation of the long ridge-like accumulations of till which often, 
on the plain country, border or separate the north-south val- 
leys. Just as a river does not make its chief deposits in the 
deepest part of its channel, where the current is strongest, 
but along the margin, building up the flood-plain, so the ice- 
stream tends to crowd the detritus out of north-south valleys, 
from the line of swiftest to the lines of slowest motion. From 
these till ridges we pass easily and naturally to drumlins. 
Every glacialist knows that the drumloid slopes of till, which 
may, in my opinion, be regarded as embryo drumlins, although 
most characteristic of stoss slopes, occur abundantly on lee 
*Seventh Ann. Report U. S. Geol. Sur., 184, 185. 
