SOURCES OF LOESS. 9 
of accumulation. The extremes of this process were, therefore, accumulation of the 
dust in the form of mantling loess on the borders of the ‘“peripheral’’ zone, and 
barren stony steppes in the interior, with seas of sand-dunes following between. 
The dense dust-storms that during the seasons of predominant winds from the 
interior darken the air of northern China and southern Turkestan, are proof that 
the process outlined above, in the sense of Richthofen’s theory, was at least one 
source of the loess of China and Mongolia, and his conception must stand as a 
beautiful and well-conceived explanation. 
The writer, in reviewing the first volume of Richthofen’s great work “‘China,”’ 
in 1876, recognized the validity and strength of theargument and was led to suggest 
additional sources of wind-borne loess.* These were: first, the silts (chiefly glacier- 
ground) brought by rivers into desert regions, where after desiccation they became 
the prey of the winds; second, another important source would be where a region 
which had long been protected by heavy vegetation should, through the chemical 
action of infiltrating waters carrying free oxygen and organic acids, become dis- 
integrated and decomposed, either in mass or broken up into polygonal blocks 
surrounded by clay detritus. Sucha state exists often to depths of 100 to 300 feet 
in regions that have not been glaciated. When in consequence of a climatic change 
toward aridity such a region should lose its protecting vegetation, it would offer 
to the action of the winds an immense store of loosely aggregated material ready 
for easy and rapid removal. On the other hand, in the oncoming of a glacial 
epoch, such a condition would explain the origin of the greater part of the glacial 
till and large and small bowlders transported in the slow movement of the great 
ice-cap. ‘This modification was accepted by Richthofen and reproduced in the 
second volume of “China.”’ 
The expedition of 1903 offered opportunities for studying the loess of Turke- 
stan and observing the way in which it was formed, and certain facts were noted 
that explain some points that have been considered by many to be inconsistent 
with Richthofen’s theory. ‘These observations showed clearly that all loess is 
primarily brought by wind and arrested by grasses under the protection of con- 
tinued generations of which it slowly accumulates. But there are two sources 
from which the wind derives the dust. One is from the bare surface of deeply 
disintegrated mountains, as on the Pamirs, and the uninterrupted deflation by the 
winds of the rocks of the desert, which are perpetually disintegrating under the 
influence of the great daily changes of temperature. The other is from the flood- 
plains and dry deltas of rivers and from the shores of retreating lakes, as I had 
suggested in the paper referred to. ‘This brings in water as an intermediate factor 
in transportation, in connection with the origin of loess in certain regions. In every 
case, however, the process must be in an arid or more than semi-arid climate, where 
extensive rock areas or widespreading flood-plains are in a state of desiccation. 
On the Pamirs we found the mountains buried in the unmoved products of their 
own disintegration. In this worst of desert regions, rocks at the surface crumble 

* The Relation of Secular Rock Disintegration to Loess, Rock Basins, and Glacial Deposits. Amer- 
ican Journal of Science, February, 1879. 
