‘ 
8 PROCEEDINGS OF THE DETROIT MEETING. 
is at present so great that there is a zone running north and south over the plains, 
where the western tributaries are leaving a part of their load and building up the 
land. Again, the comparatively slow denudation east of this zone is itself dis- 
tributed unequally, being largely confined to the drainage channels and only to a 
very small extent, as sheet erosion, affecting the flat surface of uplands away from 
creeks and gullies. The lateral slopes of the smallest as wellas of the largest drain- 
age channels in the loess region soon merge into the nearly horizontal upland 
plain. Krosion decreases with the slope and practically ceases with it. Rain water 
on a level surface appears to soak into the ground as rapidly as it falls, even in the 
heaviest rains. This is especially the case where the surface is covered by vegeta- 
tion. By far the greater part of the land area in the region of the loess consists of 
such flat land. It is believed that the greater part of the sediments of the Missis- 
sippi are taken from its own bed and bluffs and from the beds and immediate 
slopes of its tributaries, large and small. hese constitute mere narrow belts, 
which dissect the much more extensive plains, and they occupy only a small part 
of the entire land surface. The greater part of the land denudation in the central 
region of the Mississippi valley is hence confined to a comparatively limited area, 
from which the greater part of the river’s load is taken. The removal of only a 
small quantity of materials from the much more extensive level uplands must 
make their planing down by sheet erosion exceedingly slow. Lvyidently this rate 
would be still further reduced if the drainage were more sluggish. It might be so 
slow as not to equal the secular accumulation of atmospheric dust on the land sur- 
face, in which case this would of course accumulate. 
SrmILariry or Composition or Dust anp Loxss 
In their mechanical composition fine wind sediments and loess are largely 
identical. The bulk of each consists of particles from one-sixteenth to one-sixty- 
fourth of a millimeter in diameter, with two nearly symmetrically decreasing series 
of admixtures above and below these sizes. An aqueous deposit, spread over hun- 
dreds of miles of a broken topography and reaching a thickness of a hundred feet, 
could not very well be as uniform in its mechanical composition as the loess is. It 
would more frequently contain coarser materials. In particular it seems improb- 
able that a water deposit as fine as the loess should be without thin seams of fine 
silt, such as are generally to be observed in aqueous sediments. hese are more 
or less conspicuously laminated. in a wind sediment, on the other hand, such a 
sorting and lamination is impossible, owing to the smallness of each sorted load, 
to the less constancy of the depositing current, and to disturbing agencies which 
are at work thoroughly mingling successive deposits on the surface of the land. 
OTHER FEATURES SUSI'AINING EoLIAN HypornEsis 
There are other features of the loess which appear easy to explain if it be re- 
garded as a terrestrial deposit. One such feature is the relation which has been 
shown to exist between the border of the lowan drift and the loess. At the time 
of the deposition of these two terranes there was a low drainage gradient. The 
land to the south of the ice was probably a low swampy plain, where surface ero- 
sion was nearly at a standstill except along the water-courses, and where, as a 
consequence, atmospheric dust may have accumulated. Such accumulations may 
have reached up over the margin of the ice fields and may thus have caused oc- 
a 
