254 RESEARCH IN CHINA. 
of the material, we find them large and the surfaces of cleavage irregular. 
The irregularities increase as the proportion of loess decreases, and pass 
into the characteristic fracture of a bank of residual loam when the pro- 
portion of loess becomes insignificant. It is probable that the structure 
is perfected through a certain amount of drying and wetting, and it is pos- 
sible that it may be impaired or lost by too long-continued or too frequent 
wetting. Should the flow of water through a body of Huang-t’u suffice 
to remove most of the loess particles of capillary or subcapillary diameter 
and the salts, the remaining mass would approach a loam of relatively coarse 
texture. The cementing substance being taken away, the residuum would 
be more or less incoherent and would fail to maintain a vertical structure. 
The cementing, which is occasioned by the evaporation of moisture 
in the Huang-t’u, is the occasion of its endurance in steep walls. Not 
that the cement itself is very firm (itis but a film of loess paint and salts), 
but it is so arranged in the mass as to give great strength, the weakness 
of the substances being considered. Each tubelet may be regarded as a 
self-supporting element, or each rudely vertical and cemented column of 
particles may be considered as such. If the mass is permeated with mois- 
ture and the cement is dissolved, the structure remains firm because the 
interspaces are filled; if the moisture retreats, the cement is redeposited 
and the mass continues to stand fast. The structure is maintained because 
ever renewed, unless the amount of water is sufficient to supersaturate the 
body; then it flows like a liquid. 
Geographic features.—The altitude of the land during the Hin-chéu 
epoch may be inferred from the relations of the Huang-t’u formation. 
During the preceding T’ang-hién epoch it had been low, yet such that 
degradation had not ceased. The district in Shan-si had also, like other 
areas of northern central Asia, presumably been subject to a moist climate, 
and had presented a surface of decayed rock, saplite. At the present 
time the surface on which the Huang-t’u rests is not deeply decayed; on 
the contrary, the rocks are fresher than one would expect, until one reflects 
on the climatic conditions which have reigned throughout the Huang-t’u 
epoch. It follows that any mantle of saplite which may have been (and 
probably was) present when the T’ang-hién epoch of moist climate passed 
into the Hin-chéu epoch of aridity, was, to a greater or less extent, eroded 
before the Huang-t’u was laid down in Shan-si. The plains where the 
Huang-t’u is now being deposited by rivers are situated near a local or 
general base-level, and it may be assumed that the same condition governed 
during the Hin-chéu epoch. Combining these considerations, we are led 
to infer that in Shan-si and Chi-li altitude and climate favored erosion 
of saplite during the early Hin-chéu time, and in the same region depres- 
sion favored the accumulation of the Huang-t’u during later Hin-chéu 
