CHINA, MONGOLIA, AND JAPAN. 39 



This lava seems to belong to the trachydoleritic series. Of its varieties consist 

 nearly the whole of that portion of the volcanic formation that w^as traversed by 

 my route. That it obtained its great development on the surface by successive 

 flows, is evident from the stratiform structure of this part of the plateau. 



The only locality in which I observed an exposed section of comparatively fresh 

 rock, was in a quarry at Kwantung (pu), on the lower plateau. Here a bed of lava, 

 crystalline at the bottom of the section, becomes porous toward the top, and, finally, 

 highly vesicular and highly scoriaceous, this structure marking the top of the flow. 

 Above this is a bed of more compact lava than the lower. Crevices extending 

 through both of these beds are filled with a calcareous segregation product. 



I am unable to accaunt for the occurrence of this immense lava formation, excepu- 

 ing by the supposition that the successive flows took place from an immense crack, 

 the position of which is perhaps indicated by the great fault line along which the 

 dislocation took place between the higher and lower plateau. 



Terrace Deposit} — The loam of this formation has been frequently mentioned in 

 the previous pages. It occurs in the valley of every tributary of the Yang Ho and 

 probably also of the Sankang Ho. It exists in the form of terraces between Chatau 

 and Kiming, and these undoubtedly occur in the valley of the Sankang Ho from 

 Paungan (chau) to Tatung (fu). Between the Kiming hill and the Papau moun- 

 tain, a terrace of coarse detritus overlooks the valley of Hweilei (hien), its surface 

 being several hundred feet above the Yang Ho. 



In the valley of Siuenhwa (fu) this deposit seems to have suifercd less from 

 erosion, and rises, generally without terraces, at first gently then rapidly toward the 

 bordering mountains, filling ravines high up their sides. Our road to the north lay 

 over this deposit, as we skirted the hills between Siuenhwa and Kalgan, and we 

 saw it fringing the Kalgan gorge with isolated terraces high above the river. 

 Leaving this gorge, and ascending the valley of the Siwan creek, we found it in 

 continuous tprraces, which even at the Roman mission of Siwan, rise 200 or 300 

 feet above the creek. 



Going southwest from Kalgan, we find this deposit continuous from the valley 

 of Siuenhwa into that of Hwaingan, and we have already seen how it forms a 

 plateau capping the ridge between this valley and the Yang Ho at Tienching. It 

 is also undoubtedly represented along the Yang Ho from this place to Kalgan. 



We have seen it, between Tienching and Yangkau, rising unbroken from the 

 plain to high up the sides of the Barrier range, and continuous from here, in 

 terraces, through the defile west of Yangkau into the valley of Kwantung (pu), and 

 thence around the southern spur of the lower plateau through the vaUey of Fung- 

 ching and the deep break in the higher plateau, west of Maanmiau, into the valley 

 of the Te Hai, where its lofty terraces occupy the eastern part of this great 

 depression. 



The plain of the Kir Noor is formed by this deposit, which also extends through 

 the valley on the east to the Si Ho tributary of the Yang Ho. As this formation 



* For results, mostly negative, of a microscopical examiaation of the loam of this deposit from dif- 

 ferent localities, see Nos. 1, 2, 3, 5, 7, 8, 12, in Mr. Arthur Mead Edwards' Letter, Appendix 3. 



