W. M. Davis—Origin of Drumlins. 413 
satisfactory. Professor J. D. Dana’s description of Round Hill 
near New Haven (this Journal, 1883, xxvi, 357), would imply 
that it is a drumlin, although he calls it a kame; he suggests 
tnat it was formed by streams plunging down from the surface 
of the ice through a knot of profound crevasses, and “ causing 
a local deposition of the stones and earth that were in the ice.” 
But the absence of stratified sands and water-worn stones in the 
hill seem to negative this suggestion; and the occurrence else- 
where of numerous drumlins at high levels, as at Charlton, 
Mass., requires some other explanation. 
The first clear reference to dramlins as directly dependent 
on glacial action for their form was made by M. H. Close (On 
the glaciation of the rocks near Dublin, Journ. Roy. Geol. 
Soc. Ireland, 1864, i, 3); they are here said to be parallel to 
the neighboring strize, and hence like these dependent on the 
ice-sheet for their present attitude and form. The same con- 
clusion is presented in the paper of 1866, above mentioned, 
when the name drumlin was first specially proposed for them. 
Still later, when describing the physical geography of the 
neighborhood of Dublin (id., v, 1877, 49), Close writes: “It is 
perfectly certain that it must have been the rock-scoring agent 
which produced the bowlder-clay ridges.” Besides this, 
Kinahan and Close, in the pamphlet above named, stated their 
Opinion that drumlins were formed in a way “similar to that 
by which a stream of water often makes longitudinal ridges of 
sand in its bed.’ This is to my mind the best suggestion yet 
given to account for them. 
Geikie wrote as follows: “The remarkable linear direc- 
tion of certain mounds of bowlder clay in some districts of the 
Lowlands, agreeing as this does with the general bearing of 
glacial markings of the same localities, induces us to believe 
that we have here, with certain modifications, the original 
contour of the till after the superincumbent ice-sheet had dis- 
appeared” (On denudation in Scotiand since glacial times, 
Trans. Geol. Soc. Glasgow, 1867, iii, 61); but he believed 
that these forms may be also in part dependent on marine 
erosion. In the Great Iee Age (1877, 76), the same author 
briefly mentioned “the series of long smoothly-rounded banks 
or drums and sowbacks, which run parallel to the direction 
taken by the ice,” and regarded them as very little modified 
from their glacial form. They are “produced by the varying 
direction and unequal pressure of the ice-sheet,” and are “the 
glacial counterparts of those broad banks of silt and sand that 
form here and there upon the beds of rivers.” Dr. L. Johnson 
says that he accepts Geikie’s explanation and applies it to the 
New York ridges which were ‘formed underneath the glaciers 
by alternations of lateral pressure” (J. c. 1882); but this form 
