CHINA, MONGOLIA, AND, JAPAN. 43 



Indeed the only direction from which a river of any importance could have come, 

 Avas from the west, in which case it could only have been the Hwang Ho (Yellow 

 river). Let us examine into the possibility of the existence of a communication 

 between the valley of the Yellow river and the lake basins. When I was in the 

 valley of the Te Hai, I saw distinctly that the break in the plateau continued to 

 the W. S. W. as far as the eye could reach. A low, hilly country, much below the 

 level of the plateau, appeared to shut in the valley at the distance of about twenty 

 miles from the lake. Now on Klaproth's large map of Central Asia, on which, so 

 far as my experience goes, the streams of this region are laid down with a remark- 

 able approximation to accuracy, a branch of the Tourgen GoF is given as rising in 

 the very region occupied by the low hills observed by me. A native map of the 

 province of Shansi, not always correct in its details, represents this stream as rising 

 in the Te Hai. 



Thus, I think, there is little doubt that a communication exists between the val- 

 ley of the Te Hai and that of the Tourgen Gol, sufficiently depressed to be below 

 the surface level of the terrace deposits. The Tourgen Gol is a tributary of the 

 Yellow river, and if the watershed between the Te Hai and this river was below 

 the level of the ancient lakes, these must have occupied part of the valley system 

 of the north bend of the Yellow river, and must have left a corresponding deposit. 



Now, although we have no information concerning the occurrence of the terrace 

 deposit in the valley of the Tourgen Gol, we have direct testimony with regard to 

 its existence over a large area in the land of the Ortous — the desert region inclosed 

 by the northern bend of the Yellow river. Abbe Hue passed through this country 

 on his way to Tibet, and describes it as -a flat, sandy desert, frequently cut up by 

 deep ravines, in the sides of which he observed, in one place, dwellings excavated 

 in the same manner as those at Siwan.^ 



Indeed, all the information we possess concerning this region goes to show that 

 it has been the basin of a great lake, which once extended from the northern bank 

 of the Yellow river southwards to the mountains crowned by the Great Wall.^ 



Thus I think there can be little doubt that the terrace deposits, so common in 

 the system of the Yang Ho, were precipitated in a chain of connected lakes, extend- 

 ing from Yenkingchau, N. N. W. of Peking, to near Ninghia (fu) in Kansuh, a 



' Haishui of the Chinese. The valley of Tourgen Gol is probably also connected with the valley 

 of the Kir Noor; see p. 29. 



' "When the Chinese establish themselves in Tartary, if they find mountains the earth of which is 

 hard and solid, they excavate caverns in their sides. These habitations are cheaj)er than houses, and 

 less exposed to the irregularities of the seasons. They are generally well laid out ; on each side of 

 the door there are windows giving sufficient light to the interior ; the walls, the ceiling, the furnaces, 

 the kang, everything inside is coated with plaster so firm and shining that it has the appearance 



of stucco. These caves have the advantage of being warm in winter and cool in summer 



These dwellings were no novelty to us, for they abound in our mission of Siwan. However, we had 

 never seen any so well constructed as these of the Ortous." — Abbe Hue, Travels in Tartary, etc., 

 Vol. I, p. 180. 



^ Compare Ritter's Erdkunde. Asien, especially "Vol. I, p. 153 — 160; also Hue, Yol. I, p. 285; 

 and Travels of Gerbillon, in Du Halde. 



