Sec. vii] CONCLUSION OF WORK AT TURFAN 713
profess any special interest in such heritage of China's antiquity as paintings of Buddhist idols and rags from heathen tombs, was probably quite as glad as I was that the timely departure of my convoy of cases relieved him of any obligation to furnish the desired information. On the other hand, it seemed but a small concession to hasten by a few days my own start for the Kuruktägh, and by disappearing into the ` Gobi ' to afford him relief from all further trouble about myself and my doings. In deciding upon this course I was prompted quite as much by the wish of avoiding what might serve as an excuse for an attempt to interfere with my convoy while in transit, as by desire to see Muhammad Yäqûb's work satisfactorily concluded. For an inspection of the plane-table sheets brought by him to Yär-khoto showed that his detailed survey of the Turfän depression was then still far from completion.3
These considerations obliged me to bring my stay at Yär-khoto to a close on February 13th and did not allow me to make as close a survey of this impressive site as might otherwise have been possible.3a Our excavation work, too, had to be confined to a portion of the large Buddhist shrine, marked I in the sketch-plan Pl. 35, that did not appear to have been completely searched before. The ruins of Yär-khoto, remarkable both for their unusual character and the conspicuous position occupied by them, have attracted the attention of all European explorers who have visited Turfän since Dr. Regel, and the excavations conducted by the successive expeditions of Professors Grünwedel and Von Lecoq are known to have extended to different parts of them. As, however,
I am unable to trace any special account of the site other than the brief description contained in Dr. Klementz's preliminary report,4 the sketch-plan Pl. 35 prepared on my renewed visit and some rapidly gathered notes on characteristic features of the ruins may prove of use.
The ruined town of Yär-khoto owes its striking appearance and also the survival of many of its structures to its having been built on an island-like plateau rising high above the ravines or ` Yärs ' which surround it. This topographical fact accounts for the natural strength of the position occupied by the town and explains why its buildings, whatever damage they were otherwise exposed to, were absolutely safe from the effects of moisture carried over the ground by streams or irrigation and from direct encroachment by cultivation. The same position renders it also certain beyond all doubt that Yâr-khoto, the name, half Turk' half Mongol, by which the ruined town has been known since medieval times, is identical with the town of Chiao-ho x Xj. ` the interlacing rivers ', which the Chinese Annals from Han to Tang times mention as the capital of Anterior Chü-shih or Turfän.5
Survey of Yär-khoto site.
Topographical character of site.
The plateau, as seen in Pl. 35, stretches for a maximum width near the middle of about two
3 Curiously enough, no objection was taken by the Urumchi authorities to any of the surveys carried out since our return to I-Tsin-chiang in the preceding autumn, even though the permission obtained from the Central Government through the prompt help of H.M.'s Minister at Peking referred only to archaeological surveys ; cf. above, i. p. Sao.
Muhammad Yäqûb's one mile to the inch survey could in no way be brought within this category. But it was continued without hindrance among the villages to the south of Turfän town even after my own departure. That it could not be extended to the westernmost portion of the basin, about the oasis of Toksun, was due only to the fact that the surveyor, a slow worker, did not find sufficient time for the purpose before he had to rejoin me at Korla by the close of March.
3a This applies also to the sketch-plan of the site, which
over a mile from north-west to south-east, with furlongs. Two deep-cut ` Yärs ', each between
was prepared by Muhammad Yäqûb without my being able to afford the requisite direction as to details. For this reason no attempt was made to distinguish the more important ruins within the closely occupied portions of the town. Also the delineation of its main roads is not as accurate as it might have been if Afraz-gul's help had been available.
4 See Klementz, Nachrichten über Turfan, pp. 24 sqq.
5 The identification is clearly indicated by the Ming Shu (see the extract translated by Bretschneider, Med. Researches, ii. p. 191) and was rightly accepted by Dr. Klementz, Nachrichten über Turfan, p. 28. The Later Han Annals distinctly tell us of the town of Chiao-ho : ` A river bifurcates and surrounds the town ; that is why it is called Chiao-ho (interlacing) rivers ; ' cf. Chavannes, T`oung-pao, 1907, p. 211.
Natural strength of position.