The Columbia Formation in X. W. Illinois. — Jlershey. 19 
it and the bluffs in which the loess was being deposited, near 
its western end it touched the bluff on the north side for* a few 
hundred yards, there accumulating its drift and boulder belt 
and cutting off the loess. The terrace is quite distinct east 
of this point, but up the valley to the west it is lower and 
otherwise much less prominent. This I interpret as indica- 
ting that while the loess was accumulating rapidly in the com- 
paratively small and narrow but deep trough between the ice 
and the bluff' east of this point, westward it spread out over 
the entire valley and so is less strongly developed at any given 
point. Had the ice not occupied the position here supposed, 
this would not have been the case. 
The heavy deposits of Valley loess along the Mississippi 
river doubtless were derived largely from the ice-sheet which 
then stood not far back from the river in Iowa and Minnesota, 
and in smaller amount from its portion in Wisconsin — a 
northern prolongation of the ice-front here indicated in Il- 
linois. 
Just beyond the extreme border of this newer drift sheet we 
iind the ordinary Upland loess, originally 7 or 8 feet thick. 
Around the northern and western sides of the Pecatonica ice- 
lobe the loess terminates comparatively abruptly at the edge 
of the newer drift. Usually, however, it overlaps it for some 
distance. But while it is 7 feet thick beyond the newer drift, 
the overlapping portion is only from 2 to 3 feet in thickness, 
often gradually thinning until it totally disappears. This 
overlapping portion may extend one, two, or ten miles back 
upon the newer drift, but it is usually not present in any 
identifiable form more than a few miles. As already sug- 
gested, there is no stratigraphic break (if the term may be 
applied in a matter of this kind) between the newer drift and 
the Upland loess. The one grades, sometimes quite insensi- 
bly, into the other. 
Around the edge of the newer drift we find the Upland loess 
quite sandy, a feature which is rare for it away from the bor- 
der, but just what we should expect near the source of the 
sediment, where the water rushed down from the ice-fx'ont, 
carrying sand as well as fine silt, and scattering it over the 
submerged hills near by. As a final argument, I will men- 
tion that the loess-covered country lies higher than the 
