Vol. 51.] THE LOESS AND OTHEK DEPOSITS OF SHANTUNG. 247 



watersheds, depositing such vast thicknesses of debris ; and secondly, 

 the continuity of the mass over plain and valley alike, independent 

 of any streams whatever. A very good example of such a laud- 

 locked valley is seen 1| mile south-west of Chefu. A glance at 

 the map (fig. 2, p. 245) l shows that the catchment-basin of this 

 little valley covers less than 2 square miles, yet the deposit in it is 

 quite 100 feet thick. 



The deposit bears evidence of vast superficial waste. It is eroded 

 down to the bed-rock by the mountain-torrents, here as elsewhere 

 in Shantung very intermittent, and the bare vertical sides of the 

 resultant ravines which score the mass in every direction give a 

 weird and unfinished aspect to the landscape. The pebbles occupy 

 the lines of deposition, and, though in practically horizontal lines, 

 are only subangular, and the first idea on seeing the mass was that 

 it was an ordinary case of subaerial waste, the beds being deposited 

 by mountain-torrents from the wreck of the older higher rocks. 



Several considerations rendered this idea untenable. The first 

 was the unmistakable level-line in which the debris terminated 

 at its summit, some 300 feet above sea-level. Next was the in- 

 sufficiency of the gathering-ground, the land never rising above 

 1400 feet, and the passes through which any streams have flowed 

 being much lower, and in most cases quite close to, if not below, the 

 level of the upper terrace, as in the case above cited. The valleys 

 behind the hills south of Chefu also bore evident traces of having 

 been more or less continuously filled with the deposit up to the 

 same level, and no place existed for a large river flowing from 

 higher levels. 



Other reasons equally cogent pointed to a subaqueous origin. 

 These were the level nature of the surfaces over which the parting 

 gravels had been strewn, and the peculiar terraced character of the 

 slopes to seaward, which (though partly artificial) must to some 

 extent at least be referred to natural conditions. In this part of 

 the province the sea has generally penetrated to the foot of the 

 mountain-masses, and the terraced plains are reduced to modest 

 dimensions, not more than a few square miles as a general rule.. 



On the north side of Chefu harbour rises a rocky mass known as 

 the Bluff (see map, fig. 2, p. 245) of formation similar to that of the 

 mainland, but remarkable for the fine development of its quartzites. 

 It is slightly less than 2 miles distant from the port, and is now 

 connected to the mainland by a narrow strip of shingle, only a few 

 feet above high-tide mark. We found the superficial deposit 

 exactly reproduced on the Bluff, and rising to approximative^ 

 the same level. Here the deposit could not be accounted for 

 locally, as only a few points rose over the summit-level, and 

 evidently it had once been connected with that of the mainland, 

 forming a continuous whole. 



Leaving Chefu, we started westward for Tsi-nan and, crossing a 

 spur of the mountains, entered the plain through which the river 

 Taku (long. E. 121° 17') flows. The deposit formed a well-marked 

 1 The valley is marked A. 



