80 The American Geologist. August, 1894 
posit of till contains notably fewer boulders and less gravel 
than on the hill and close to its base, and its thickness is de- 
creased td 5 feet or in some places only 3 feet. At Prof. 
King's house. 24 feet above lake Mendota, the till is 8 feet 
deep, underlain by 12 feet of sand and fine gravel, to the bed 
rock at the depth of 20 feet. Likewise on the lowland east 
of the University hill, a well at Prof. Birge's house, about 10 
feet above lake Mendota, went S feet in till, and several feet 
into sand below. 
One reason for an exceptionally large proportion of sand 
in the drift of this area is found in the underlying Potsdam 
sandstone, which reaches five miles northeastward and twelve 
miles northward from Madison. Prof. Irving suggested that 
the basins filled by the lakes north and south of this city were 
probably made chiefly by glacial erosion of this soft sand- 
stone. 
When the border of the ice-sheet had retreated so far as to 
uncover the land here, much further deposition of sand and 
gravel as hillocky and ridged low kames and eskers. occasion- 
ally enclosing boulders but not overlain by till, took place on 
the area extending from a half mile to one and a half miles 
south of the University hill. The most noteworthy of these 
deposits is an esker a half mile or more in length from south- 
east to northwest and 30 to 60 feet high, which forms part of 
the northeastern shore of lake Wingra and reaches southward 
beyond Fitchburg street. Excavations in the southern part 
of this gravel and sand ridge show an irregular stratification 
with a prevailing northwestward dip. varying from 15° to 
\') . The glacial river by which the esker was formed, walled 
by ice on each side occupying the present areas of lakes Win- 
gra and Monona, flowed to the northwest, transverse to the 
previous direction of the glacial movement over this tract, 
but toward the Black Earth col and the avenue of the some- 
what earlier glacial drainage to which I think the accumula- 
tion of the sand and gravel cores of the Madison drunilins 
was due. 
A< I I MDLATION OF THE OVERLYING TILL. 
A.bove the sand and gravel in these drunilins there is so 
scanty a veneer of till that we may readily assign nearly all 
of it to the probably somewhat uniformly thick sheet of finally 
