ORIGIN OF THE LOESS OF THE MISSISSIPPI VALLEY 801 
taking place on the bluffs adjacent to the Mississippi and Mis- 
souri valleys. Observation seems to clearly indicate this. But 
such accumulations are relatively scant in amount and limited in 
extent, and it is difficult, if not impossible, to believe that the 
great loess mantle had its origin from the wind drift of flood 
plains no more extensive than those of today. It must be con- 
stantly borne in mind that the eolian deposits are measured, not 
by the quantity of silt borne by the winds and lodged on the 
surface, but by the difference between such lodgment and the 
erosion of the surface. Under most conditions with which we 
are familiar the erosion is more than a match for the dust 
accumulations. The conditions must then have been extraor- 
dinary which would give a dust deposition sufficient to supply 
erosion and still leave so large a residuum as the loess mantle 
implies. The unleached and relatively unweathered nature of 
the body of the loess is specially in point here. These con- 
siderations warn us of the theoretical danger of too greatly cir- 
cumscribing the fluvial action. 
On the other hand, if we attempt to extend the fluvial 
hypothesis too greatly we fail to leave sufficient feeding ground 
for the molluscan life and we encounter the topographical and 
physical difficulties which have been previously urged against 
the pure aqueous theory. A Janus-faced hypothesis is here 
offered in the hope that by a judicious reference of a part of the 
loess to one class of action and a part to the other, a joint 
explanation may be found to afford a true elucidation of the 
perplexing formation. At any rate, it has seemed worth while 
to propose the hypothesis for trial. It will doubtless be 
extremely difficult to find a line of demarkation between the two 
classes of deposits. Such attempts as have been made in this 
line justify this apprehension. This supplementary theory has 
been in mind for several years and was briefly suggested in my 
paper on the Genetic Classification of the Pleistocene Deposits, pre- 
sented at the Fifth Session of the International Congress of 
Geologists at Washington in 1891." An effort has been made 
x Compte-Rendu of the Fifth Session of the International Congress of Geologists, 
Washington, 1891, p. 192. 
