82 • The American Geologist, August, 1894 
drift from the land, carried forward subglacially while the 
drumlin was being amassed, would be commingled with its 
previously englacial drift. Especially where such a drumlin 
appears to have been formed upon a rock knoh. thereby con- 
cealing it, as instanced by Chamberlin from observations of 
Mr. Buell in Wisconsin, the englacial drift and many frag- 
ments from the knob on which it was accumulated must be 
intermixed in the growing drift hill. The idea of Prof. Davis 
that the englacial drift in becoming superglacial by ablation 
must be mostly or wholly washed, assorted, and stratified on 
the ice surface by the water and streams produced in its melt- 
ing, seems not to be supported by the prevailing character of 
the drift covering the border of the Malaspina ice-sheet; and 
it was not so supposed by me for the waning border of the 
Pleistocene ice-sheet in the process of its concentration of the 
drift to form these hills. After reading Prof. P. S. Tarr's re- 
cent paper on this subject,* I still believe the usual drumlins 
consisting wholly of till to have been formed from englacial 
drift which had become superglacial and was afterward 
enclosed as a stratum of drift in the ice-sheet. By this view 
I think that all the peculiarities of distribution and group- 
ing of the drumlins may be best explained. 
Upon a large region extending eastward from Madison 
drumlins are very abundant, so that Mr. I. M. Buell, assisting 
Prof. Chamberlin in the glacial field work of the United States 
Geological Survey, has mapped nearly 2.500 of them in south- 
eastern Wisconsin, finding in some tracts an average of 
about seventy- live for each township six miles square. f The 
country is a moderately rolling or hilly but nowhere very ele- 
vated expanse of the Cambrian and Silurian bed rocks, upon 
which the drift is spread as a somewhat uniform sheet. Above 
the general drift sheet its drumlin hills comprise usually only 
a small part of its entire amount. They occupy, even where 
grouped most closely, perhaps a quarter or third of all tin- 
area, rising 50 to 150 feet above the intervening low grounds. 
It will be a most interesting question to determine whether 
the Madison type of drumlins has any large representation in 
this region. Conversations with Profs. Chamberlin and Sal- 
* Am. Geologist, vol. xm. pp. 393-407, June, 1894. 
I Am. Geologist, vol. xir. pp. 172, 176, Sept., 1893. 
