PHYSIOGRAPHY OF NORTHWESTERN CHINA. 249 
material of the loess. In the course of its transportation down the valleys, 
from the inner basins, from which it is supposed to have been sorted, to 
the lowlands where it is now found, it was spread again and again upon 
the flood-plains of the streams, and in the alternation of the seasons was 
continually subjected to the process of sorting and re-sorting, which is 
still going on under similar conditions. In so far as the valleys in which 
it was deposited were environed by hills and mountains, the loess proper 
was liable to interstratification with the detritus of alluvial cones washed 
from the hillsides. These two constituents of the Huang-t’u formation, 
the loess, which is the eolian product, and the gravels and sands, which 
are the fluviatile products, were brought together by the joint and alter- 
nating activities of the winds and streams. 
Physical characters of the Huang-t’u.—The manner of accumulation of 
the loess formation has been frequently discussed on the basis of its peculiar 
characters: constitution, texture, and structure. In the preceding para- 
graphs the sorting and reassembling of the mechanical constituents has 
been sufficiently described. The salts which, through their abundance and 
constant renewal, render the Huang-t’u unusually fertile, next demand 
consideration. They are not accounted for by the activities of winds and 
rivers, for we would suppose that fine material such as the loess must 
become thoroughly decomposed and leached in the alternating processes 
of wetting and drying, sorting and transporting, to which it has been 
subjected during the Quaternary. In fact loess is composed of chemically 
obdurate substances: quartz, silicate of alumina, and oxide of iron; yet 
these are associated with soluble carbonates, sulphates, and chlorides of the 
alkaline earths and alkalies. These salts may be regarded as secondary 
constituents derived from ground water. 
In a deposit of Huang-t’u during alternate seasons of moisture and 
drought, the movement of ground water is probably intermittent. If we 
suppose a mass to be deposited dry, or to have been dried out during a 
period of general dessication, it must absorb a large volume of water from 
rains of a succeeding season or episode. The pore space in fine alluvial 
clays of this sort is at least from 40 to 60'per cent of their volume.* Rains 
are promptly taken up by the Huang-t’u, from which there is almost 
no run-off, but they can moisten the mass to a considerable depth only 
if long continued, since large pore space and evaporation limit their 
penetration. Thus the lower part of the mass may long remain dry, so 
far as wetting from the surface is concerned. But beneath the Huang-t’u, 
at the contact with the underlying rocks, is a surface on which waters 
flow from the adjacent mountain slopes. 
*C. S. Slichter. 1gth Ann. Rep., U. S. Geol. Surv., Part II, 1899. Van Hise, Treatise on Metamor- 
phism, p. 125. 
