IOWA ACADEMY OP SCIENCES. 
85 
to a level scarcely fifty feet above the present stream, was 
built up to eighty or ninety feet above the river at that point. 
The depth of filling is found to increase upon passing down 
the valley, and becomes scarcely noticeable at Hannibal. It 
is, therefore, much like a delta, formed where a rapid stream 
emerges into a sluggish, lake-like body of water. It consists 
mainly of fine material, sand or silt, with few pebbles greater 
than one- fourth inch in diameter. A fine gravel, however, 
appears at an exposure called “Yellow Banks,” near the 
mouth of the Des Moines river. The bowlder bed in Keokuk, 
described above, received at this time a capping of sand fifteen 
or twenty feet in depth. Sand deposits are also found at a 
corresponding level in Hamilton, 111., near the foot of the 
rapids, capping a low part of the rock bluff. Another possible 
remnant of the sand filling is found at Sandusky, Iowa, six 
miles above Keokuk, immediately back of the bowlder-strewn 
slope, noted above. It there rises about eighty feet above the 
river, or to wfithin twenty-five feet of the level of the bottom of 
the channel of the temporary Mississippi, ten miles to the 
north. No remnants of the filling have been noted in this 
interval of ten miles and it is thought probable that the rate of 
fall was so great above Sandusky that but little lodgment of 
material occurred. 
In the portion of the Mississippi valley covered by the 
Labrador ice field at the Illinoian stage of glaciation, there 
appears to be no such sand filling as is found below the rapids, 
although it has nearly as low a gradient. This feature con- 
firms the above interpretation, that the sand filling occurred 
during this stage of glaciation. 
In explanatioQ of the small amount of material deposited in 
the bed of the temporary Mississippi, Professor Chamberlin 
has suggested to me that the ground in which this channel 
was excavated may have been frozen at the time of excavation, 
its situation being on the immediate borders of the ice sheet, 
and that this frozen condition of the ground may have pre- 
vented the stream from eroding more material than it could 
readily transport 
The time involved in the valley filling is a question of much 
interest, but one on which an estimate is very difficult to 
make. The filling of any given section is not a measure of 
the full work of the stream, but simply an index to the excess 
of material above the limits of transportation by the stream. 
