PHYSIOGRAPHY OF NORTHWESTERN CHINA. 225 
which prevailed when we were there. The fact that houses are built on 
the bank, 9 to 12 feet above the water, shows that in this broad channel 
ordinary floods do not rise to that level. On the south side, the bluff 
of the Huang-t’u formation is 60 to 80 feet, 18 to 24 meters, high. At the 
base it is composed of sand or very soft sandstone for 20 feet, 6 meters, 
followed by gravel, 8 to 12 feet, then loess to the summit. The loess is not 
distinctly stratified and has the usual columnar structure. Ascending to 
the top of this bluff, one sees the south side of the valley formed by hills 
of the Huang-t’u formation, which reach an altitude of 300 feet, 90 meters, 
above the river and extend southward to the foothills of the Ta-hua-shan. 
A cross-section of the valley of the Huang-ho from north to south 
is given in Fig. 54, which is drawn without reference to any exact relation 
between horizontal and vertical scales, since we have no survey of the valley 
on which to base a precise representation. It will be seen that at this 
point the Huang-ho flows parallel to the normal fault, of which the Ta-hua- 
shan is the upthrown block, and that the valley occupies a position on the 
downthrown block, which is covered by the Huang-t’u formation to an 
unknown depth. 
To explain the aggraded condition of the valley, we may suppose 
that the district once stood somewhat lower with reference to sea-level 
than now, permitting the Huang-t’u to accumulate as an alluvial deposit 
up to the level which is 300 feet, 90 meters, above the present river bed; 
and that, in consequence of elevation, the river has now sunk its channel 
to a position 300 feet, 90 meters, below that surface which it occupied. 
Or we may, as an alternative, infer that in consequence of greater throw 
of the fault in the vicinity of Tung-kuan, as compared with some point 
further east, the bed-rock lies deeper at Tung-kuan than it does to the 
eastward. ‘Then the river, in developing its course on the downthrown 
block, would encounter the rock ridge somewhere east of Tung-kuan, but, 
being retarded, would deposit in its upper course, while cutting a rocky 
canyon below. As the channel in the canyon was sunk, that in the allu- 
vium would also be lowered. Somecolor is lent to this hypothesis by the 
fact that rapids dangerous to navigation are reported east of Tung-kuan. 
The lower course of the river has never been examined by any scientific 
explorer, as the road which follows the valley diverges from the stream, 
presumably in the section which is through the canyon. The exact condi- 
tion which determines the heavy deposit of the Huang-tu formation at the 
great elbow of the Huang-ho remains, therefore, indeterminate. It may 
be a fact bearing upon general movements of the district with reference 
to the great plains of China, or one determined by purely local warping in 
the lower valley of the river. 
