The Cohtmbla Formation in X. W. Illinois. — Hershen. 11 
in many cases fills tlieir lower portions with a bed of blue 
gravel and driftwood to a depth of as much as ten feet. At 
the time of the excavation of these gorges the streams were 
more vigorous and flowed at a lower level, indicating a con- 
siderably more elevated condition of the region. But later, 
when the Florence gravel and alluvial sand and clay were ac- 
cumulating, not only the fact of the gradual silting of the val- 
leys, but the passage of these flood-plains through the inter- 
glacial gorges, deeply burying their rock bottoms, indicates a 
lower altitude for the district than formerly. It is presumed 
that this depression, beginning long before the appearance of 
the first flood-plain deposits, continued at a comparatively 
even rate through tlie Florence subepoch, and was merely the 
first part of the great Columbia subsidence to which the loess 
owes its existence. 
Although this basal or fluvial member of the (Columbia for- 
mation is, as yet, only known to the writer in the basin of the 
Pecatonica river, it is doubtless j)resent in all the larger val- 
leys in this portion of the state. Lying at so low a level and 
having such imperfect and inaccessible outcrops, it can easily 
escape observation. A few wells in northwestern Illinois pen- 
etrate a black mucky stratum, containing logs and other woody 
matter, under the loess. and over the drift sheet. This occupies 
the same stratigraphic position as the Florence gravel, but it 
is rather of the nature of an upland soil than a fluvial deposit. 
The buried forest bed of northeastern Iowa is also stratigraph- 
ically equivalent to the Florence gravel, and the fossil con- 
tents indicate a similar climate. Although the study of the 
fossils of our early Columbia fluvial member is yet very 
incomplete, I feel certain, from the great abundance of shells 
belonging to species now living in this region, that the climate 
was not arctic, nor even such as would be found within 100 or 
perhaps 200 miles from the edge of the great continental 
glacier. 
Valley Loess. 
The basal member of the Columbia formation in this district 
grades upward into a deposit which, in the Pecatonica valle}', 
•is a moderately fine stratified sand, followed above by a light 
bluish gra}^ or brown loess; but in the smaller valleys, notably 
■.the valley of Yellow creek, the loess silt immediately overlies 
