PHYSIOGRAPHY OF NORTHWESTERN CHINA. 253 
body of Huang-t’u, whether water-laid or wind-drifted, is very porous. 
The porosity may be partially due to removal by water of subcapillary 
grains of loess after consolidation, but is in greater degree an original 
condition. In alluvial Huang-t’u the interstices were originally occupied 
by water. In eolian Huang-t’u air filled them; and so long as the air 
continues dry, the deposit remains incoherent and structureless, as one 
may see in newly formed drifts during cold weather. A footprint in such 
drifts resembles one in dry, powdery snow. 
Whatever the original condition, whether wet or dry, the Huang-t’u 
settles, at least during an early stage of consolidation. Under the influence 
of gravity all interstitial spaces lessen, and interstices which approach a 
horizontal position must be narrowed more than any which approximate 
verticality. Theoretically the resulting structure should exhibit linear 
openings at angles of 45° to the horizontal, if the grains were spherical and 
of uniform size,* but since they do not fulfil this condition that structure 
is likely to be very imperfect. The effect of gravity is to bring the irregular 
particles in closer contact so far as they rest one above another, and to 
leave them in more open order so far as they lie side by side. Interstitial 
air or water is forced into the more open spaces, and in any movement 
follows the upward-downward lines of least resistance. 
The Huang-t’u contains two classes of material which tend to cement 
surfaces on which they may be left: the fine loess particles and the chemi- 
cally dissolved salts of the ground water. Let us suppose that a mass 
which was either deposited wet, or which has been moistened by rain and 
capillary elevation of ground water, becomes dry. Evaporation from the 
surface may cause moisture to rise, leaving a dry mass below if the under- 
flow is checked; or lowering of ground water may leave a dry layer above. 
In either case the moisture in the capillary spaces will coat the walls of 
those spaces with a paint of the finest loess particles and with a deposit 
of salts left by evaporation. There will be a tendency to close the smaller 
interstices, those approximating a horizontal attitude, and the larger up- 
and-down ones will be left comparatively open. If the mass be moistened 
or wet again, they will become the channels through which moist air or 
water will move; they may to some extent be opened by softening of the 
loess paint or solution of the salts; and they will be coated again when 
the mass again dries out. It is conceived that repetitions of this process 
give rise to the rudely vertical tubes described by von Richthofen, and 
thus produce the tendency to vertical cleavage. 
It is to be noted that cleavage of the Huang-t’u is imperfect. If we 
regard the inequalities of the structure in comparison with the fine grain 
* Slichter, loc. cvt. 
