192 RESEARCH IN CHINA. 
certain extent, have gathered the flying dust. But the river plain and 
hillsides, so far as they consisted of Huang-t’u, were absolutely bare, and 
every surface exposed to the attack of the wind was scoured, whereas 
every hollow became an eddy in which a drift accumulated. 
The road leading southwest from Tung-yii traverses the flood-plain 
of the Hu-t’o-ho and enters a wind gap which is filled with Huang-t’u. 
At the narrowest point of the gap, 3 miles, 5 kilometers, from the village 
the road passes under a hill which rises 500 feet, 150 meters, above it on 
the northwest. This hill is capped by the Huang-t’u, a remnant of the 
once extensive formation which filled the valley to the level of 3,100 feet, 
945 meters, or more above sea, that is, to a level but little, if any, below 
that of the plain about Wu-t’ai-hién. 
In the hills immediately south of Tung-yii there is a very interesting 
occurrence of the Huang-t’u, at an altitude of 1,000 to 1,400 feet, 300 to 
420 meters, above the Hu-t’o-ho. A spur of the Ki-chéu-shan, projecting 
northward east of Fang-lan-chén (atlas sheet C II) forms a triangular 
embayment which is drained toward the northeast. Within the rock 
rims of this embayment lies a body of loess. It exhibits the usual charac- 
teristics of vertical cleavage and lack of stratification, and appears, so far 
as a casual observation allowed us to judge, to be free from interbedded 
gravels. The position is higher than any body of Huang-t’u which we 
may reasonably attribute to alluvial deposition, except that which lies 
in the adjacent Yau-t’6u district, where the altitude is clearly the effect 
of recent warping. The position is one which is sheltered by ridges on 
the west and north from the winds that blow across the Hin-chéu basin or 
down the valley past Tung-yti, and thus appears favorable to the accumu- 
lation of wind-drifted loess. In the dust storm which we experienced 
near Tung-yti, drifts a foot or more deep accumulated in hollows having 
a like situation, and this body of apparently very homogeneous loess is 
probably in the same manner blown out of the adjacent valleys (see 
Fig. B, Plate XXVI). 
In the Hin-chéu basin, which one enters in journeying southwestward 
from Tung-yti, the Huang-t’u formation covers an area of about 350 square 
miles, forming a wide plain across which flows the shallow Hu-t’o-ho, which 
is neither intrenched nor elevated. For the first time since leaving the Great 
Plain, we came into a region where the Huang-t’u is clearly related to 
the valley of a large river, where the interaction of the strong winds and 
constant stream might be studied. Our route lay nearer the mountains 
than the river, but we were able to note that here, as in the Bay of Peking, 
the water transports silt, sand, and gravel, which high water distributes 
according to the strength of current over the flood-plain, and which the 
