part of Saghalien, an island as large as Ireland; the mountain-chain which rises between the Strait of Tartary and the valleys of the Amoor and Ousouri; the vast spaces of Manchooria; the mountains dividing Corea from China, covered with great forests, and containing so considerable mineral wealth as to afford a profit to the Chinese vagabonds who work for it with the most primitive processes; and the forbidden land of Corea. On the classic ground of the Celestial Empire are spaces larger than Great Britain that may be ranked among unknown lands. Eastern and northern Thibet, the least accessible part of all the empire, presents in particular many interesting problems, the most important of which is that of the river systems. What is the true relation between the rivers which we see hypothetically represented on the map of Thibet and those of Indo-China and India? Is the chain of the Kuen-Lung, which appears on the maps as one of the principal ranges of the continent, really worthy to be ranked with the Himalaya and the Thian-Shan systems? The latter question is now complicated with some apparently contradictory circumstances. The southern part of eastern Turkistan deserves the attention of explorers equally with Thibet. It is the most inaccessible desert of the continent, a land of jade-stone and gold, of camels and the wild horses that are not known anywhere else in the world.
Prejévalsky, during his last expedition, touched a country of quite exceptional geographical interest, the sources of the Hoang-Ho. He was not able to penetrate to the "Sea of Stars" itself, but he saw a considerable part of the narrow valley by which the upper Yellow River flows toward the east. Access to the sources themselves of this great stream has thus become one of the geographical desiderata; but no doubt exists concerning the absence of the subterranean connection between the Hoang-Ho and the Tarim, of which the Chinese geographers have often spoken.
The great desert of Gobi has lately been tolerably well explored, but the question is still to be answered whether it is crossed by a chain of mountains connecting the eastern end of the Thian-Shan with the In-shan. The mountains, if they exist, can not be very high, for no large rivers flow from the region, and some streams flow toward it; but they are marked on several maps of China, including that of the Russian staff; and the existence of a direct route between Koukou-Khota and Barkoul indicates that there are springs along the line, and they must have hills to maintain them.
Northern and northeastern Mongolia have been topographically delineated with some exactness, but no naturalist has visited them; and only three or four European travelers have crossed the Kingan range between Mongolia and Manchooria, whose geological structure and mineral, zoölogical, and botanical riches have still to be found out.
Returning along this range into China proper, we enter a country the superficial character of which has been often described, but con-
