CHINA, MONGOLIA, AND JAPAN. 73 



which is mentioned in a previous chapter as occurring along the southern edge of 

 the plateau, in the erosion of the lava region. In all these instances the depressions 

 are entirely m the solid rock, and vary in size from a few yards to several thousand 

 feet across. They have the appearance of being produced by erosion and not by 

 sinking. In the instance before us this conformation is often assisted by cross 

 dykes of greenstone. But the occurrence generally would seem to arise from ine- 

 qualities in the texture of the rock. Whatever the cause of these depressions may 

 be, their manner of formation is probably closely connected with the origin of a 

 large class of desert lake beds. 



For many miles the surface of the rock was entirely bare of soil, excepting in the 

 bottoms of the depressions just mentioned, where ponds are probably formed ua wet 

 years. 



From this hilly region we came gradually into another of those broad plains, 

 which form, in the aggregate, the true plateau. These plains, the steppes of the 

 Russians, and tola of the Mongols, are like those of our own deserts in the Rocky 

 mountains. They are great vaUeys, often from twenty to sixty miles broad, filled 

 with marine deposits that have retained their horizontal position and remained often 

 intact from erosion. Their surface is not, strictly speaking, horizontal, but slopes 

 from both sides to the centre. 



The deposit forming the substructure of this plain, seems to be the same sand- 

 stone and conglomerate that we have seen on the Tamchintala, judging from some 

 blocks of these rocks seen near a Mongol dwelling. 



Crossing this plain we came, near its northern edge, to a line of basaltic cones 

 from 100 to 150 feet high, isolated from the low flat hills to the north, and appa- 

 rently resting on clay slate. They seemed thus to belong to a bed or stream rather 

 than to a dyke. Whether the flat hills near by are a continuation of the same 

 volcanic rock I could not determine. 



The rock is a brownish-black, minutely crystalline basalt. On the surface of the 

 plain, near these hills, I found large numbers of fragments of black and red cellular 

 lava, and abundant angular pieces of chalcedony, and red and green jasper, etc. 



Dec. 2. During this day we crossed two broad valley depressions, the same cal- 

 careous sandstone and conglomerate already mentioned, forming apparently the sub- 

 structure both of the long vaUey slopes and of the higher land intervening between 

 these. A few fragments of blue limestone and white quartz, derived probably from 

 the formation we crossed yesterday, were found in the surface gravel ; but a large 

 percentage of this gravel consisted of chalcedony, cornelian, and agate. 



From the highest ground the flat outline of the plateau was visible in every 

 direction, excepting to the south, where we could see the hills of the past two or 

 three days rising to the height of perhaps 1000 feet above the neighboring plateau. 



Dec. 3. We travelled the past night and this day on the continuation of the steppes 

 of the last two days. During the afternoon the plain descended gradually to the 

 north till it ceased abruptly against a granite ridge from 50 to 100 feet high. Beyond 

 this ridge, for a few miles, the country though somewhat lower than the plain of 

 the morning, is bare of the steppe deposit, and presents a rough, granite surface. 



Dec. 4. Detained one day by a bouran or snow-storm of great violence. ■ 



10 May. 1866. 



