246 RESEARCH IN CHINA. 
After discussing the residual mantle of decayed material as a source 
of glacial drift, he continues: 
In Northern Asia, north of the 4oth degree of latitude, there are no traces of general 
glacial action such as existed in Northeastern America and Northern Europe. The 
evidence, indeed, is all the other way. And yet, while the rocks of Southern Asia show 
extensive residue of disintegration, the results of a secular decomposition protected from 
erosion by an abundant vegetation, the feldspathic rocks of Central Asia are as free from 
this as are those of Northeastern America. 
The only answer to the question, what has become of them? is that they have been 
blown and sifted and assorted by the winds, the heavier fragments remaining to be reduced 
by weathering and to form the stony steppes, the sand drifting in billowy waves over 
the country, and forming sand deserts, while the fine dust floating in the air, an impalpable 
powder, is deposited far and near, and under the influence and protection of the steppe 
grasses is transformed into the loess. 
Pumpelly’s generalization is sustained by our interpretation of the 
physiographic history. During the Pei-t’ai and T’ang-hién stages the 
region under discussion had long been subjected to erosion at a moderate 
altitude above sea. It had presented a relief not unlike that of the south- 
ern Atlantic states, and under similar climatic conditions may well have 
developed a similar mantle of residual soil or saplite. There is reason to 
believe that the Pei-t’ai and T’ang-hién epochs belong to the Tertiary 
period, and that Asia during that time may have shared in the mild and 
moist climate which is generally assigned to that period. If so, both 
topographic and climatic conditions were favorable to the accumulation 
of the products of rock decay over that part of Asia, within which the 
characteristic features of the Pei-t’ai and T’ang-hién stages had their 
development. The question of the provincial or continental extent of 
these features is discussed in the following chapter on systematic physiog- 
raphy, but I may here anticipate the conclusion that they are extensively 
developed phases of topography. 
As the initial step in the explanation of the Huang-t’u formation, 
we thus assume that prior to the Hin-chéu epoch there existed over a 
considerable area an adequate supply of the products of secular rock 
disintegration, which was held in place by vegetation, flourishing under 
favorable climatic conditions. 
Climatic fluctuations.—The next step in the development of our hypoth- 
esis of the formation of the Huang-t’u is the assumption of climatic change. 
In this we follow both von Richthofen and Pumpelly. Von Richthofen 
discusses the question in his account of the development of the arid basins 
of Asia,* and resting on the brackish character of the great Sarmatic medi- 
terranean of eastern Europe and western Asia,{ and the gradual withdrawal 

* China, vol. I, p. 109 et seq. 
t Neumeyer, Erdgeschichte, vol. 1, p. 523 et seq. 
