LOESS OF CHINA—-BARBOUR 289 
As Andersson points out, however (1-107), there exists in many 
places no sharp line of demarcation between the loess and the red 
clay. Instead, locally a gradual transition occurs through layers 
showing characteristics intermediate between them. It is, however, 
a distinct and older formation. 
In other places Zdansky describes the loess as making sharp con- 
tact with the Hipparion beds below. In the district Kast of Wan 
Ch’uan, near the Mongolian border, I have observed shallow deposits 
of typical yellow loess occupying gullies cut in an underlying red- 
dish clay-loam formation, that has all the appearance of such a re- 
sorted residual deposit. Im this are large carbonate “ loess- 
piippchen ” often over a foot in length. 
Pére Teilhard de Chardin has mentioned in conversation that a 
sunilar concretion-bearing red clay underlies the loess in many parts 
of the Eastern Gobi Desert recently visited by Pére Licent and 
himself. At the same time he pointed out that, as far as his observa- 
tions went, gravel beds were frequently found at the base of the 
loess, and he expressed the opinion that in certain restricted areas 
fluviatile conditions may have existed after the greater part of the 
land surface was already ruled by aridity. In any case it is not sur- 
prising that on a continental area sharp contrasts of conditions 
should exist locally even during arid times. 
2. Kansu continental deposits —Recent observations made inde- 
pendently by Dr. J. G. Andersson and by Dr. George B. Cressey of 
Shanghai College, point to the existence in many parts of Kansu of 
heavy beds of red sands and gravels overlying the Hipparion clay, 
and, therefore, presumably of Pliocene age. They look like the 
deposits of a great delta or alluvial system, and from their descrip- 
tion may well represent the products of the period of greater mois- 
ture that came between the time of the red clay and the days of the 
loess. 
3. San-Men beds.—In a number of localities sand and gravel beds 
may be seen at the base of the loess. The moisture conditions sug- 
gested by these seem to have been widespread over North China. 
Dr. V. K. Ting first described a series of such beds from the San-Men 
rapids of the Yellow River (1-118). These underlie the loess and 
carry large fresh-water bivalves (Quadrula), which Doctor Dall 
of the Smithsonian Institution of Washington regards as probably 
early Pleistocene in age. 
When the caissons for the new bridge opposite the Governor’s 
yamen in Tientsin were being lowered, shells of similar large fresh- 
water mollusks were found at a depth of 81 feet below ground level. 
Through the courtesy of Mr. P. L. Yang of the Chihli River Con- 
servancy Board, I was given facilities for examining on the spot 
the material brought up from the various levels while the caisson was 
