798 T. C. CHAMBERLIN 
place either in a strictly fluvial fashion, or in a fluvio-lacustrine 
fashion, or in a true lacustrine fashion, or in an arm of the sea. 
In the first place, the vertical distribution of the loess cannot 
easily be explained. The extreme vertical range is not far from 
a thousand feet. The range within a score of miles is frequently 
from 500 to 700 feet. The loess sometimes seems to the field 
observer to have a special fondness for summit heights. It 
sometimes mantles topography of a pronouncedly rolling type. 
It does not then appear to be a deposit which once had a level 
or even a smooth surface out of which the rolling surface has 
been eroded, but to be a mantle laid down upon a previously 
undulatory surface. Such a mantle might perchance be laid 
down from water, but I am not aware that we have any demon- 
strative deposition of the kind which closely simulates the 
mantling of the loess in some of the upland territory. To sup- 
pose that the Mississippi, Missouri, Illinois, Wabash and lower 
Ohio rivers were so swollen that they united over their divides 
and threw down a mantle of fine silt over the southern and 
western half of Iowa and the southern parts of Illinois, Indiana 
and Ohio, is a somewhat severe tax upon belief. It is difficult 
to imagine the conditions which should have maintained such a 
body of water. This has been so much discussed that I need 
not dwell upon it. But even if such a body be supposed, it is 
difficult to imagine how the deposition could have been precisely 
what we find in the case of the loess. It is futhermore difficult 
to account for the presence of the land shells which abound in 
it; for if this great flood had the ice-sheet for its northern 
border, it is extremely difficult to imagine how it could have 
been peopled so widely with the terrestrial mollusks. 
The limit of the loess does not appear to be a strictly topo- 
graphic one. It is difficult to bring its border into strict accord 
with a horizontal plain as required by the lacustrine and marine 
phases of the hypothesis, or even into a consistent gradient as 
required by the fluvial phase, without an arbitrary warping of 
the surface. The spread of the loess in the lower Mississippi 
valley is more extensive and reaches greater heights on the east 
