i86 ROUTE UP THE HOANG-HO VALLEY. 



and the former missionaries (Martini and Gerbillon), 

 we determined to keep in the valley of the River. 

 This route afforded greater interest for zoological 

 and botanical researches than could be found in the 

 sandy wastes in the interior of Ordos ; besides 

 which we wished to settle the question of the bifur- 

 cation of the Hoang-ho in its northern bend. 



We marched along the River for 290 miles 

 from the ferry opposite Bautu to the town of Ding- 

 hu,^ and the conclusion we arrived at was that the 

 river does not divide Into separate channels in its 

 northern bend, as is usually represented on the maps, 

 and that this part of its course has undergone a 

 change. For the sake of clearness, I will first give 

 a general sketch of the character of that part of the 

 River and its valley which we surveyed, and then 

 continue the narrative of our journey in Ordos. 



Winding considerably for so large a river, the 

 Hoang-ho flows at the rate of 300 feet a minute '^ 

 through a valley bounded on the north by the In- 

 shan and its western continuations, and on the south 

 by a belt of drift sands known to the Mongols under 

 the name of Kuzupchi. The banks of the river and 

 its bed are composed of slimy clay ; the water is 

 very thick, and when standing for a short time 

 deposits I '3 per cent, of mud. However, the thick 



'^ Marked CJiagan-subar-kan on Klaproth's and Kiepcrt's maps. 



^ [Nearly 3^ miles an hour.] We observed this velocity of current 

 near the shore during our crossing near the town of Bautu ; in mid- 

 stream the river was more rapid, but the velocity of the current de- 

 pends on the high or low state of the water ; when we crossed it was 

 about the average level. 



