国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

Sec. iv]   VISITS TO RUINED SITES OF TURFAN   1167

(Fig. 271), I started the series of rapid excursions which were to acquaint me with the well-known sites of the Turfân district. In the course of them I visited Kara-khôja, the ancient Kao-ch`ang and the Turfân capital of Tang and Uigur times, with its imposing ruins (Figs. 272, 277); the important series of Buddhist shrines and cave-temples which dot the sides of the narrow valleys descending from Murtuk and Singim, and among them the grottoes of Bezeklik with their fine mural paintings ; and later, from the town of Turfân, the smaller sites along the slopes of the hill range east of Buluyuk. With most of these ruins I was to become more familiar during my stay of 1914-15, and for the reasons already indicated 'at the beginning of this section any observations I may have to offer regarding them must be left for another publication. As regards the local conditions affecting archaeological work there, and the facilities which unfortunately they offer for destructive digging by natives, a reference to my Personal Narrative will suffice here." A few antiques, picked up on occasion of my visits to those sites or acquired by purchase, are described in the List below.18

During my stay at the oasis of Turfân proper, where arrangements for topographical exploration in the Western Kuruk-tagh and for transport detained me for a week, I took occasion to pay repeated visits to the remarkable ruined site of Yâr-khoto (Map No. 54. D. I), which was occupied by the capital of Turfân down to Tang times. Its peculiarly strong position between two deep-cut ravines or ` Yârs', to which the place owes its modern name, half Turki, half Mongol, as well as its ancient Chinese designation Chiao-hot, ` converging streams', is well known and needs no detailed description here. The rough sketch-plan reproduced in Plate 49 shows the site close to its upper end and will help to illustrate the situation of the town, which occupies the southern half of the narrow island-like plateau. Of the striking appearance of its closely packed and in parts very massive ruins, the panoramic view in Fig. 273 and Figs. 275, 276, showing the central portion of the town on either side of its main street, will convey some impression.

The very extent of the area which the remains of dwellings, largely carved out of the live clay, cover in bewildering confusion would have rendered at any time the systematic exploration of the whole site a very protracted and difficult task. Nor could the hope of adequate results have justified such efforts ; for even a cursory inspection sufficed to make it clear how sadly the ruins of the dead town lacked that protection which abandonment to the desert might have assured them. There was practically no drift-sand here to cover up any objects that might have escaped removal after occupation had ceased, and constant digging by the villagers for soil to be used as manure in the adjoining cultivated area had laid bare the natural hard clay in most of the dwellings, big or small. Conditions for archaeological work were obviously more favourable among the ruins of Buddhist shrines, to be found mostly near the northern end of the town and in the open space beyond it (Figs. 278, 279) ; for their walls, being structural, had fallen in their decay and covered the interior with heavier accumulations of débris. This explained why the partial clearings effected by previous European explorers appeared to have been confined mainly to their ruins.

In order to gain some personal knowledge of the conditions in which antiques such as those brought to me for sale by neighbouring villagers is were being obtained at the site, I made experimental excavations at two modest ruins which it was possible to clear within the short time

Visits to Turfân sites.

Ruined town of Yâr-kholo.

Exposed condition of ruined dwellings.

Excavation of dwelling.

17 Cf. Deserl Calhay, ii. pp. 359 sqq.

18 In addition to these a number of Uigur text fragments, brought for sale by natives and acquired at Kara-khôja and elsewhere, still await examination. At the Toyuk site I picked up a considerable number of torn fragments from Chinese Buddhist Sutra rolls in a débris-strewn ravine, where they had been thrown out from shrines above in the course

of some previous exploration ; for specimens of these see Chavannes, Documents, Nos. 990, 991, Pl. XXXVI.

" See for such the small bronze statuettes of Avalokiteivara, Y.K. 005-007, Pl. VI, VII.

A number of fragmentary Uigur documents and Chinese Sutra texts, all badly torn, were also acquired at Yar-khoto.