LOESS OF CHINA—BARBOUR 291 
ferent, perhaps slightly older horizon. Several fragmentary mammal 
bones were recovered. 
Fresh-water beds with fossils of the San-Men type have been 
found by members of the Geological Survey near Pao-ting Fu; Doctor 
Andersson has mentioned finding them south of Ta-tung Fu, and 
they will certainly be reported from many other spots. They imply 
fairly widespread conditions of greater moisture before the onset 
of the aridity. 
B. YOUNGER THAN THE LOESS 
1. Redeposited loess—In many places, especially on the lower 
slopes of the hills, the loess is overlaid by series of sand and gravel 
beds with layers of loess. For the most part this formation is poor 
in fossils, but bones of bighorn sheep, oxen, and deer are occasionally 
found in the gravels. I have found remains of both the latter two 
animals in gully banks in and near the Hanoorpa Pass from Kalgan 
to Mongolia. The definitely stratified nature of this formation 
shows that torrent action was the chief determining factor in its 
deposition, though wind may have played a minor part in the case 
of the layers of loess. The individual lenses of loess are never 
more than a few feet in thickness but the aggregate thickness of 
the sediments may reach at least 50 feet. The material occurs 
typically as a valley or torrential deposit, and, where seen in contact 
with the true loess, is invariably found to overlie it, or to occupy 
gullies cut through it. Doctor Andersson has applied the name 
“ Redeposited loess” to this formation. The animal remains and 
the character of the material making the beds point to a date dis- 
tinctly more recent than that of true loess, and indicate also a great 
change of climate from that ruling during the earlier steppe epoch. 
2. Alluviwm.—sStill more recently the rivers of the present cycle 
of erosion have deposited gravel and alluvium on the broad flood 
plains or in narrow valleys. 
AGH OF THE LOESS 
The age of the loess may be determined in two ways. Firstly, by 
comparison of those few types of animals whose remains have been 
found embedded in it with fauna of other regions that have been 
studied, taken in conjunction with our knowledge of the climatic 
conditions ruling in eastern Asia and other parts of the globe at 
different stages during the last half million years. Secondly, by 
bracketing its age between the dates of older and younger beds which 
are found respectively to underlie and overlie the genuine loess. 
With regard to the latter line of reasoning, the loess must be more 
recent than the Hipparion beds which mark the close of Miocene 
