ORIGIN. OF THEILOESS OF THE MISSISSIPPI VALLEY 799 
side than on the west side, so far as present knowledge goes. A 
similar fact seems to be true of the Missouri valley. I think 
this is generally true, but my observations are not sufficient to 
justify its unqualified affirmation as a generalization. 
There are other difficulties attending the aqueous theory in 
its simple application, but I need not attempt an exhaustive 
recitial here as they have received emphasis in the long battle 
between the eolian and aqueous hypotheses. The foregoing 
will I trust suffice to show that there is abundant occasion to 
still cast about for a more satisfactory explanation of the loess 
puzzle. 
The supplementary hypothesis herewith proposed attempts 
to divide the honors between the aqueous and eolian agencies. 
It recognizes the tremendous force of the arguments from the 
distribution and the constitution of the loess in favor of the 
glacio- fluvial hypothesis, and it adopts that hypothesis as 
the fundamental explanation of the origin of the Mississippian 
loess. It assumes the presence of the Iowan ice at the chief 
stage of loess deposition. It assumes a very low slope of the 
land and a consequent wide wandering of the glacial waters. It 
assumes the development of extensive flats over which the silts 
derived from glacial grinding were spread. It assumes that the 
glacial waters were subject to great fluctuations; 1° as the result 
of periods of warm weather in the melting season,and 2° as the 
result of warm rains, which not only added directly to the 
volume of water, but forced the rapid melting of the ice. Gil- 
bert has acutely observed that there is no way in which the 
atmosphere can convey its heat energy to a glacier so effectually 
as through warm rains. 
Let it be imagined, therefore, that the silty waters from the 
margin of the ice-fields wandered over broad flats and constantly 
built them up by their sediments, and that at periodical flood 
stages they extended themselves widely over the plains, while 
between the flood stages they withdrew to more limited courses. 
The territory covered by the maximum extension of the 
waters would be the zone of accumulation of fluvial loess. It is 
