INFLUENCE OF CLIMATIC CHANGES ON VEGETATION 359 



deposits on the top of which they have hitherto been flowing. Thus, ap- 

 parently, a terrace is formed ; and a repetition of the process gives rise to 

 a series, provided neither deposition nor erosion proceeds so far as to 

 efface the earlier terraces. In western and central Asia, throughout an 

 area at least 3,000 miles long from east to west and 1,500 miles wide from 

 north to south, one everywhere finds terraces which appear to be of cli- 

 matic origin. Similar terraces are also found in Arizona and Utah, 

 although these have not been critically investigated and usually have been 

 considered to be due to earth-movements. 



Influence op Changes op Climate on Vegetation — Loess 



Perhaps the most important of all the results of changes of climate is 

 the effect on life. Eeference has already been made to large areas of dead 

 vegetation, which are located at the ends of withering rivers and which 

 form a notable feature of the deserts of the Lop basin. The inferred in- 

 fluence of the growth of plants on the accumulation of soil and on the 

 formation of terraces has also been mentioned. An effect kindred to this 

 last is seen in many places on the south side of the Lop basin. The pre- 

 vailing winds of this region are from northerly quarters across the great 

 sandy desert of Takla-Makan. The moving air picks up vast quantities 

 of fine dust, which, during the summer, fills the air with a dense haze for 

 weeks at a time. Fine yellow dust is constantly falling at such a rate 

 that a visible layer accumulates in a day in a pan of water set in the mid- 

 dle of a field of grass. Time and again I tried this experiment with the 

 same result, even though the region within ten or twenty miles had been 

 moistened by a shower within a few hours. The great desert to the north 

 had not been moistened, and its loose, dry dust was continually being car- 

 ried into the upper air, not only by ordinary winds, but by great dust- 

 whirls a thousand or more feet high. From a mountain-top I several 

 times watched a cloud of dust approach from the desert after a storm. 

 It filled the air to a height of thousands of feet, and advanced steadily 

 before a gentle breeze scarcely perceptible to the observer. In this way 

 the northern slope of the Kwen Lun mountains south of the Lop basin 

 is being showered with yellow dust, which is nothing more nor less than 

 loess of the most typical sort. Wherever there is sufficient vegetation, 

 which is only close to the higher mountains — where the rainfall is compar- 

 atively large — the dust is held in place, and heavy deposits of loess are in 

 process of accumulation. Lower down the mountains, where the rainfall 

 and vegetation are less abimdant, there is also much loess, but it is being 

 dissected and carried away. Apparently it w,as deposited at some former 

 time, when conditions were different from those of today. The difference 



