40 GEOLOGICAL RESEARCHES IN 



is found at the head of the water system of this northern branch of the Yang Ho, 

 it must be continuous, unless washed away, in all the valleys of this basin between 

 the plateau and the Barrier range. Thus the deposit in the valley of the Kir 

 Noor probably continues, through the break in the plateau to the southeast, into the 

 vaUey of the Si Ho, and through this to the Yang Ho. Indeed, judging from the 

 appearance of the region lying between the plateau and the Barrier range, as seen 

 from the tower at Ha Noor, this deposit seems to occupy here a large area. 



We can trace some of the more important islands that were isolated by the lake 

 in which this deposit originated. One of tlfese seems to have been that part of the 

 plateau lying between the Si Ho and the Kir Noor. Another instance is the low 

 ridge that separates the Yang Ho from the Hwaingan creek, while a much larger 

 one is the hilly country between the Yang Ho and Sankang Ho. 



Thus the body of water in which this deposit was formed consisted of a series of 

 lakes several hundred feet deep, occupying the valleys of the Sankang Ho, Yang 

 Ho, and Si Ho, and standing at a level sufficiently high to cover the lower water- 

 sheds between these streams. 



This deposit is everywhere a calcareous loam formed of an almost impalpable 

 powder, easily crushed between the fingers, and yet so firm that vertical clifi^s of it 

 remain unbroken for many years, which is sufficiently proved by the fact, before 

 stated, that the inhabitants of the country excavate entire villages in the base of 

 perpendicular cliffs that rise more than 100 feet above their dwellings. When 

 breaks occur, the loam falls in immense plates, or tabular masses, leaving a new 

 vertical face. Near the mountain sides and in the narrow gorges the loam is more 

 sandy, and contains the gravel and fragments of rocks coming from the immediate 

 neighborhood, but everywhere else it consists uniformly of an almost impalpable 

 powder. 



A characteristic feature of this loam deposit is its tendency to cleave according 

 to two vertical planes at right angles to each other, causing it to assume the form of 

 needles under certain conditions of erosion. 



The effects of erosion in this deposit are often very interesting, illustrating in a 

 marked manner the retrograde formation of ravines. The country is often cut up 

 by gullies 30 to 70 feet deep, and from 10 to 20 feet wide, with vertical walls. In 

 these channels wagon roads run for many miles without rising to the plain. In the 

 vaUey, between Kwantung (pu) and the Yangkau defile, I crossed a gully 40 or 50 

 feet deep, and not more than four feet wide, having the same breadth all the way 

 down, and which, with these dimensions, foUows a tortuous course for more than a 

 mile. In the same valley another ravine of this kind, only eight or nine feet wide, 

 and not less than 100 feet deep, compelled us to make a detour of over a mile. 



Wherever a cliff of this deposit presents itself the beginning of this action is 

 visible. The surface drainage of a small neighboring area of the plain being con- 

 centrated toward one point on the edge of the cliff, cuts, in its fall, a channel from 

 top to bottom, and this, with each succeeding rain, works its way backward toward 

 the mountains. As the erosion progresses the sides of the gullies offer new starting 

 points for tributary ravines. 



We have here, in the softest material that can support such action, a repetition 



