The evidence, in fact, does not justify us in placing the original home of the Sumerians at Anau, nor indeed in any particular spot in Central Asia or Iran that has yet been examined. But it serves to indicate the region of the world in which we may expect that future excavations will reveal data of a more conclusive character. It may be that the ruined sites of Seistan and the Kirman province will exhibit closer parallels with the civilizations of Elam and Sumer. Meanwhile it is clear that some contact must have taken place between the early peoples of the latter countries and the settlements to the north of the Kopet Dagh. We may thus picture the Sumerians before their arrival in Babylonia as inhabitants of some district to the east of the Euphrates valley, where they evolved the elements of their culture, which is already found in a comparatively advanced stage of development on the earliest of South Babylonian sites.
A further result of the recent explorations in Turkestan is that an adequate explanation is afforded of the unrest in Central Asia, which gave rise to the Sumerian immigration and to similar racial movements westward. It may now be regarded as established that periods of desiccation and extreme aridity have led to the abandonment of extensive tracts of country, with the result that their former inhabitants have, from time to time, been forced to seek sanctuary in more favoured districts. While nomad tribes in their search for fresh pasturage might drift over the broad steppes to the north and west of Turkestan, the agricultural peoples on its southern border would be forced to turn south of the Caspian. The bleak uplands of the Iranian plateau offer small attractions for permanent settlement, and the routes of the migrant tribes would naturally lead in the direction of Asia Minor and the Mesopotamian plain. Such a condition of unrest in Central Asia would naturally react on peoples at a considerable distance, and this fact explains the periodical invasions to which Babylonia has been subjected from the east. It may be added that the immigration of Semitic tribes into Syria and Northern Babylonia should possibly be traced to physical causes of a like nature. Periods of aridity may have occurred in the central portions of the Arabian continent, and may have given rise to the Semitic invasions of prehistoric and historic times.
Thus it is possible that the two races, which we find in possession of Sumer and Akkad during the earliest historical periods, though they arrived from opposite quarters, were forced into the region of the Euphrates by causes of a precisely similar character. As the Semites, on their way northwards from Arabia, colonized the Syrian coast-lands through which they passed, so the Sumerian race may well have left permanent traces of its presence in the valleys and more fertile oases of Iran. There are already indications that work on Syrian and West Mesopotamian sites will throw a flood of light upon the problems of early Semitic history, and it may perhaps fall to the lot of a fortunate excavator, in some region east of the Euphrates valley, to recover the cult-images of primitive Sumerian gods, and to bring to light examples of the picture-writing from which the early cuneiform characters were derived.
[2] Accounts of the first expedition were published under the title "Explorations in Turkestan," as Publication No. 26 of the Carnegie Institution of Washington (1905). The various monographs on the results of the second expedition are published in two volumes, entitled "Explorations in Turkestan; Expedition of 1904," as Publication No. 73 (1908) of the same institution. Both works were edited by Mr. Raphael Pumpelly, who in 1906 had already summarized his conclusions in his Presidential Address before the Geological Society of America (see "Bulletin of the Geol. Soc. of Amer.," Vol. 17, pp. 637 ff.). In a separate volume, entitled "The Pulse of Asia," Mr. Huntington has given an account of his more recent journey.
[3] Cf. Geikie, "The Great Ice Age and its Relation to the Antiquity of Man," 3rd ed., pp. 694, 698. In 1894, Prof. James Geikie had noted the probability that glacial phenomena were more extensively developed in the mountains and tablelands of Asia than he felt justified in representing in his Glacial Map of Asia. In it he incorporated only the results of previous observations, at the same time emphasizing its "necessarily unsatisfactory character" (op. cit., p. 831, PI. xiii.). This lack of evidence has now in great measure been remedied.
[4] Loess was formerly regarded as simply a deposit of glacial or fluvial origin, but Richthofen's theory that its subsequent distribution was largely due to wind-transport (cf. "China," Bd. I., pp. 56 ff.) is now generally accepted. The fact that it is found heaped up against the sides of mountains and contains land, and not water, shells, is unanswerable evidence. For its general character and distribution, see Sir Archibald Geikie's "Text-book of Geology," 4th ed., I., pp. 439 f.; II., p. 1351. It may be noted that the formation of loess-beds and sand-deserts is a continuous process at the present day, under the strong winds which prevail in certain seasons in Central Asia; and even when there is little wind the air is often thick with fine dust. The reverse of the process is visible in the effects of wind-erosion, very striking instances of which have been described by Dr. Stein; cp. e.g. "Ruins of Khotan," p. 189 f., and "Ancient Khotan." I., p. 107.
[5] It should be noted that the substance of the dunes around Khotan is to be distinguished from the true drifting sand of other Central Asian deserts. For Prof. de Lóczy has shown by analysis that there is almost complete uniformity in composition between the recently formed fertile loess of Yotkan (the site of the ancient capital of Khotan) and the moving "sand" now surrounding and covering the ancient sites in the desert; cf. "Ancient Khotan," I., pp. 127 f., 199, 242. The thickness of pure loess above the culture stratum at Yotkan was no less than from nine to eleven feet, a fact which had led earlier European visitors to suppose that some catastrophe, such as a great flood, had overwhelmed the old town. It is merely a striking example of the manner in which vegetation, under irrigation, catches and retains the floating loess-dust.
[6] After his recent journey Dr. Stein writes of the Khotan region that it appears to him certain that "the water-supply at present available in the Yurung-kash could under no system whatever be made to suffice for the irrigation of the whole of the large tracts now abandoned to the desert, and for this broad fact desiccation alone supplies an adequate explanation"; see the "Geographical Journal," vol. xxxiv. (1909), p. 17.