the Yangtsze ; the elevated sandstone plain of northern Anhwei is in places defined by escarpments of the older mountain-chains, rising abruptly from its surface. The traveller in this district is, in fact, forced to the conclusion that he is passing over an ancient sea- bottom, the mountain-masses of the older Devonian and Carboniferous formations forming islands, rising in lines of irregular cliffs, which from their indented outline betoken a long-continued period of depression.
These rocks are succeeded by the Loess. Before its deposition they seem to have been partially, though slightly, denuded, denoting probably some change in the physical conditions of the adjoining land.
What is true of the four provinces adjacent to the Yangtsze is probably true, likewise, with regard to the others over which the Loess extends. There was, at some time subsequent to the upheaval and denudation of the Carboniferous and apparently Triassic rocks, a period of considerable depression in, at least, Central China. The deposition of the Loess probably marked its close.
I have assumed the age of the sandstones and gravels to be Tertiary, principally from their position, overlying all the older rocks, and because comparatively little denudation has occurred in either them or the older rocks since their deposition. As stated above, the evidence of fossils is as yet wanting; but we may hope to see it supplied. The main fact of the depression antecedent to the deposition of the Loess may be regarded, however, as proved.
Structure of the Loess. — The other argument adduced in support of the subaerial theory, as stated above, is founded on the internal evidence of the deposit. Baron von Bichthofen holds that the Loess is unstratified ; this, so far as relates to the absence of apparent layers of deposition, is true, as it is likewise of the Loess of the Rhine. As in the latter, however, the beds of calcareous nodules point to an apparent stratification of materials. In this respect, as in others to be pointed out, it bears a close mechanical resemblance to chalk, though in mineral composition very different from that rock. The occurrence of land-shells and remains of land-animals is looked upon as a strong proof in support of this theory. The fossil origin of these exuviae, however, is more than doubtful. The Loess is perfectly pervious to water ; to use the Baron's own words, " It is perforated throughout its mass by thin tubes, which ramify like the roots of grass, and have evidently their origin in the former existence of roots." As a fact, streams seldom or never flow on the surface of the Loess ; they take by preference underground courses, or work for themselves deep valleys, penetrating into the mass in all directions, like branches from the stem of a tree. This mass has, besides, a tendency to cleave in vertical planes, enabling thereby the smallest streams to penetrate its mass ; so readily, moreover, is it acted on by water, or even by aqueous vapour, that I have in my possession specimens which disintegrated in the moist air of a Shanghai summer, and which have since rearranged themselves at the bottom of the drawer in which they had been placed. These facts render the presence of
