CHINA, MONGOLIA, AND JAPAN 71 



be the remnants of a deposit the horizontal beds of which were continuous over the 

 area in question. I examined one of these hillocks, about 50 feet high, near lake 

 Bilika Noor, and found it made up of the following beds, from younger to older : — 



Compact, yellowish-gray limestone, with a tendency to oolitic structure. 



Thin bed of dark clay, or earth, with concretions of manganese. 



Bed of finely crystalline, white, saccharoid gypsum. 



Gypsum in massive, transparent crystals associated with more or less red clay. 



The stratification is horizontal throughout, and the same structure seemed to be 

 contmuous as far as the Mingan hills. What the character of the plateau is I could 

 not determine ; as seen in the distance it limits the depression with a cliff and long 

 talus. 



An alluvial deposit of red loam is present in many of the valleys, and is, perhaps, 

 nearly contemporaneous with the erosion of the water-courses. 



Nov. 27. In the mornmg we found ourselves in the Mingan hills, apparently an 

 isolated protuberance rising only a few hundred feet above the plateau. The rocks 

 of these hills, where first observed near the southern edge, were chiefly quartzite, 

 compact sandstone, and a talco-argillaceous schist, in highly inclined strata trending 

 N. W. and dipping to N. E. Several miles further to the northwest we came to 

 ridges of limestone, in beds also highly inclined, with a strike W. N. W. and dip to 

 S. S. W. This rock resembles the limestone of the hills west of Peking. It is tra- 

 versed by dykes of greenstone. In the Mingan hUls I found a few roUed fragments 

 of basaltic lava similar to that of the southern edge of the plateau. 



To the west of these hills lies the broad deep valley of Olannoor, which seems to 

 connect the depression south of these hUls with the great plain of Tamchintala, 

 to which we now descend. As we enter upon this steppe we see before us nothing 

 but an unbroken sandy and gravelly plain with a little scattered grass. A con- 

 siderable percentage of the pebbles on the surface consists of agate, cornelian, and 

 chalcedony. 



Nov. 28. The morning found us still travelling on the Tamchintala, but we 

 soon descended into a large vaUey-like depression. The plateau is here cut into 

 to the depth of perhaps 150 feet, the vertical wall giving an insight into its local 

 structure. The whole exposed thickness consists of horizontal strata of white cal- 

 careous sandstone with thin beds of arenaceous limestone interstratified. At the 

 bottom of the section a bed of red arenaceous clay crops out. The sandstone varies 

 in grain from a fine grit to a fine conglomerate, the ingredients of both being ap- 

 parently identical with those of the gravel on the surface, between which and the 

 iinderlying rock there is no line of demarcation. If the pebbles of agate, cornelian 

 and chalcedony are derived from the amygdaloidal lava, so common farther south, 

 their occurrence in this deposit throws light on the relative ages of the two forma- 

 tions. 



After crossing this vaUey depression, which is several miles broad, we ascended 

 to the plain at about the same level, apparently, as on the other side. 



Nov. 29. During the previous night we left the plain and entered a rough and 

 very undulating country. Here a belt of older rocks, about seventy miles broad, 



