________________
SEPTEMBER, 1917] A THIRD JOURNEY OF EXPLORATION IN CENTRAL ASIA
201
The first week of November 1914 found the four parties into which my expedition had divided, since September safely reunited at Kara-khoja, an important ancient oasis in the centre of the Turfan depression. A combination of geographical and archæological reasons had made me fix upon Turfan as the base and chief ground for our labours of the ensuing winter. It was certainly the natural and most convenient starting-place for the series of tours I was anxious to organize for the exploration of unknown or as yet inadequately surveyed portions of the Kuruk-tagh and Lop deserts to the south. I myself, ever since my brief visit of 1907, had felt drawn back to Turfan by the hope that its abundant ruins of Buddhist times were not yet completely exhausted, even though, easily accessible as they are, within or quite close to oases, they had received much attention from successive archæological expeditions, Russian, German, and Japanese. Finally, geographical and antiquarian interests united in prompting me to make an accurate large-scale survey of the Turfan Basin; for, apart from its containing in its terminal salt lake what probably is one of the deepest depressions below sea-level of our globe, there is the important fact that, within close topographical limits, and hence in a concentrated form, as it were, it exhibits all those characteristic physical features, which make its great neighbour and counterpart, the Tarim Basin, so instructive both to the geographer and historical student.
This detailed survey of the Turfan depression, on the large scale of one mile to an inch and with clinometrically observed contours, was taken in hand by Surveyor Muhammad Yakub, almost as soon as he had joined me after a difficult desert crossing from the terminal drainage basin of Hami or Kumul. A few days later I could send off R. B. Lal Singh, pining as always for fresh hard work, to the Kuruk-tagh. The rapidly increasing cold, felt even here close to sea-level, gave hope by then that he would be able to overcome the difficulties arising in those truly " Dry Mountains" from the want of drinkable water, by the use of ice formed on salt springs-or of snow if such happened to fall.
With my remaining two Indian assistants I had already started the archæological labours that were to keep us busy for the next three and a half months. The ruined town, known as Idikut-shahri, which was their first scene and adjoins Kara-khoja, has long ago been identified as the site of Kao-chang (or Khocho in early Turki), the Turfan capital during T'ang rule (seventh to eighth century A.D.) and the subsequent Uigur period. Massive walls o stamped clay enclose here an area, nearly a mile square, containing the ruins of very numerous structures, built of sun-dried bricks or clay. Most of them were Buddhist shrines and several of imposing dimensions. For generations past these débris-filled ruins have been quarried by the cultivators of the adjoining villages in search of manuring earth for their fields, and many of the smaller structures had been levelled to gain more ground for cultivation. Since the excavations made here between 1902-06 by Professors Grünwedel and Von Lecoq, of the Berlin Ethnographic Museum, the villagers had extended their destructive operations in the hope of securing manuscript remains and antiques as valuable byproducts for sale to Europeans. Of such finds I was able to acquire a fair number. But it was more satisfactory to find that in some ruins deeper débris strata had escaped exploita tion. Their systematic clearing was rewarded by a variety of small but interesting remains, such as fresco pieces, fragments of paintings on paper and cloth, stucco relievos, illustrating Buddhist art at Turfan. Manuscript fragments in the Uigur, Tibetan, Chinese, and Manichæan scripts were also recovered. The discovery of a hoard of well-preserved metal objects, including decorated bronze mirrors, ornaments, etc., offered special interest, as the large number of coins found with it permits the date of its deposit in Sung tires to be fixed with approximate accuracy. Simultaneously with these clearings I had an exact plan of the whole site prepared.
