IOWA ACADEMY OF SCIENCES. 
83 
This bed has been discussed at some length by Mr. Gordon 
in the Geology of lowa,*^ and three interpretations for its 
origin are presented. 
First . — That it was formed by river action alone, i. e., as an 
alluvial bar. 
Second . — That it is due to the cutting down of a till sheet, 
the coarse material being left as a residue. 
Third. — That it is a bowldery moraine dropped at the edge 
of the ice sheet at the Illinoian stage of glaciation. 
Of the three interpretations the second seems to Mr. 
Gordon, as well as to the present writer, the most applicable. 
Dr. Kellogg’s suggestion of a fall as the cause seems, at best, 
to be poorly sustained. A similar bowlder bed occurs near 
Warsaw, 111. It there forms a capping for an eroded till sur- 
face and bears clear evidence of removal of the fine material 
by a stream, with the retention of the bowlders as a residue. 
A bowlder bed is also found along the face of the west bluff of 
the rapids near Sandusky, about six miles above Keokuk, at a 
level forty to sixty feet above the stream, that probably was 
derived from the erosion of a sheet of till, though the exposure 
is scarcely extensive enough to show clearly the relationship. 
It seems referable to the period of erosion that produced the 
beds at Keokuk and Warsaw. 
The amount of erosion effected is so great that the begin- 
ning of this new channel seems to date from near the close of 
the Kansan stage of glaciation. This becomes more evident 
as we study into the later stages of the history of the river. 
Even if the river had been forced into a channel farther east 
than the lower rapids, it seems scarcely probable that it 
remained long in that course. It apparently began its work of 
opening the course across the rapids long before the Labrador 
ice field had reached the region. 
PILLING AT THE ILLINOIAN STAGE OP GLACIATION. 
Following this great erosion there came a partial filling of 
the part of the valley immediately outside the limits of the 
Illinoian drift sheet. It is well displayed below the rapids, 
and some remnants are to be seen along the borders of the 
rapids. This filling appears to have occurred at the Illinoian 
stage of glaciation: Evidence of this relationship is to be 
found in the connection, or close association, of this filling 
* Geolog-y of Iowa, Vol. HI, 1893, pp. 252-255. See also PI. XV. 
