We are not here concerned with details of the earlier geological evidence, except in so far as they illustrate or explain the physical changes in the character of the country during more recent times. It has long been recognized that the deserts of Central Asia owe their existence to a process of desiccation that has taken place since the Glacial epoch,[3] and recent investigations have shown that the contrast to present conditions was even more marked than was previously supposed. The members of the first Pumpelly Expedition have noted that glaciers existed on a greatly extended scale throughout the mountains bordering the great basins of Central Asia on the south and east, and they have proved the existence of several great glacial expansions, each of which naturally reacted on the climate of the central region. During the sub-glacial period there was a general trend towards desolation, and the dried silts of seas and rivers were carried by the wind across the surface of the ground. The lightest material was carried farthest, and, wherever the scanty vegetation could hold it, it was deposited in beds of "loess," the extraordinarily fine and fertile soil which covers a great part of Northern China and Turkestan, and extends in a continuous zone from North of the Caspian to Central Europe.[4] The heavier silts in the shape of sands moved more slowly under the pressure of the wind, and they formed great deserts of sand-dunes, heaped in places more than a hundred feet high. It is to the shifting or formation of such sand-deserts in historic times that we owe the burial of the cities in the Khotan region, which have been so successfully excavated by Dr. Stein for the Indian Government.[5]

Although it is clear that since Glacial times there has been a general trend towards the present arid condition of Central Asia, there is reason to believe that, as in the Glacial epoch, the subsequent climatic changes have not been uniform. Periods of extreme aridity have occurred in which the condition of certain regions may have been more desolate than it is to-day. But these appear to have alternated with more humid periods, when the tracts which were deserted may again have been rendered capable of sustaining life. Already in the prehistoric period, however, the sea of sand-dunes had encroached upon the fertile plains of loess, and it is mainly in the delta-oases, formed by streams emerging from the mountains, or at points where large rivers lose themselves in the plain, as at Merv, that traces of man's handiwork have been discovered.

Throughout the region of the oases in Southern Turkestan, to the north of the Kopet Dagh, the Pumpelly Expedition constantly noted the sites of former habitations in regions which are now desolate. Not only are there traces of occupation where villages exist to-day, but there are also large areas which must once have been densely peopled, although they are now deserted. The present supply of water in the region could support but a small proportion of its former inhabitants, and it is necessary to suppose either that there was a greater rainfall, or that evaporation was less rapid owing to a lower temperature. Similar evidence has been collected with regard to the former condition of Chinese Turkestan,[6] and it is clear that extensive tracts in Central Asia, which are now abandoned to the desert, at one time supported a considerable population. The evidence points to a change in climatic conditions, which has reacted on the character of the country in such a way as to cause racial migrations.[7]

In the hope of throwing light on the character of the former dwellers in the deserted regions of Russian Turkestan, the second Pumpelly Expedition undertook excavations at selected sites. At Ghiaur Kala in the Merv Oasis it was ascertained that the earliest period of occupation was not older than a few centuries B.C., though it is probable that among the great number of mounds in the oasis some are of a considerably earlier date. Far more important were the results obtained by excavations in the region below the northern slopes of the Kopet Dagh. It was at one of the delta-oases, at Anau, near Askhabad, some three hundred miles east of the Caspian, that the Pumpelly Expedition found traces of prehistoric cultures, and obtained its principal material for archaeological study.

Near the middle of the Anau oasis, and about a mile apart, are two hills with rounded contours, rising some forty and fifty feet above the plain, and marking the sites of long-forgotten cities. The structure of the North Kurgan, or tumulus, had already been exposed by a trench cut in it some twenty-five years ago by General Komorof, which showed stratified remains, including bones of animals and potsherds of plain and painted wares. It was this trench that first directed Mr. Pumpelly's attention to the mound during his first expedition, and his subsequent excavations, both here and in the South Kurgan, exposed the same stratified structure.


Fig. 68. Designs on painted potsherds of the Neolithic period (Culture I.) from the North Kurgan at Anau.—From Pumpelly, Expl. in Turk., I., p. 128, Nos. 67-73.