218 The Ainerican Geologist. October, i89?. 
Profs. E. W. Claypoi.e and W. H. Niles doubted tliat an 
embayment of the ice-border could be formed in the Connecticut 
valley, and referred to the valley glaciers of alpine districts as 
suggestive that local glaciers might continue in valleys for some 
time after the departure of the ice-sheet from the adjoining high- 
lands. 
Mr. G. K. Gilbert spoke of the question whether the abla- 
tion of the departing ice-sheet caused its portion south of the 
White mountains and sov;theast of the Green mountain range to 
be at last divided by these highlands from the ice farther north, 
so that its pressure by a great thickness of ice in Canada ceased, 
leaving the southern portion to take new courses of outflow de- 
pendent only on its own mass and the contour of the land. 
Conditions of accumulation of drumlin.s. By Warren Up- 
HAM. Descriptions of the various forms of these hills of glacial 
drift or till, and of their known distribution in the United States 
and Canada, were followed by discussion of the evidences that 
they were accumulated rapidly beneath the thinned and receding 
border of the ice- sheet, being probably in large part built up by 
lodgment of previously englacial drift. In becoming lodged on 
the drumlins or on other and low deposits of subglacial till, 
bowlders and pebbles of drift that had been englacial would be 
considerabl}' striated and planed ; but the drift which fell loosely 
on the surface from an englacial or superglacial position when the 
ice disappeared would be mostly angular, not being thus sub- 
jected to attrition. At a great distance from the edge of the ice- 
sheet and within all its central area, the currents of its upper and 
lower portions probably moved outward with nearly equal rates, 
the upper movement being somewhat faster than at the base. 
Upon a belt extending many miles back from the margin, how- 
ever, where the slope of the ice-surface had more descent, the 
upper currents of the ice, unsupported on the outer side, would 
move several times faster than its lower currents, which were im- 
peded by friction on the land. There would be accordingly 
within this belt a strong tendency of the ice to flow outward with 
curved currents, tending first to carry drift upward into the ice- 
sheet and later to bear it downward and deposit it partly beneath 
the edge of the ice and partly along the ice boundary. The 
author believed the drumlins to be submargiual drift accumula- 
tions chiefly so deposited inside the course of contemporaneous 
