of the masonry, and now leave the top inaccessible. It is certain that there was no place of worship within the structure, though it is possible that the top once contained one. The bricks used throughout measure 14 inches by 9 with a thickness of 4 inches. Traces of red paint, perhaps used as a grounding for fresco-work, could be made out at the back of some of the niches. It is certain that these, as well as the Buddha images, were once painted, and their colour effect must have greatly added to the impressive appearance of this tall and well-proportioned pile.
On November 23rd I returned to Toyuk and halted during the next fifteen days at that most picturesque of all Turf-an localities. During this halt I devoted as much time as my other duties would allow to the work which my previous reconnaissance of the often-searched ruins of Toyuk suggested as still worth undertaking. The lively stream descending from Su-bàshi to Toyuk spreads fertility immediately below the mouth of a strikingly rugged and gloomy gorge. The natural attractions of the site, greatly heightened by the contrast between the luxuriance of the little oasis and the utterly barren and weirdly eroded hill-sides above it, must have made Toyuk in ancient times, as it is now, a much-frequented place of pilgrimage. The prosperity secured to it by the far-famed produce of its vineyards and orchards (Fig. 312) and by its easy accessibility from the chief oases of the basin made it no doubt, at all times, easy to provide for those who looked after the shrines here established. Nor was it difficult, at a site so favoured, to assure sustenance for those pious poor, whether Buddhist pilgrims of old or Muhammadan mendicants in later times, who chose to pay their devotions to the holy spot, and eventually settled down there for good. Since the many Buddhist shrines and monastic retreats in the gorge were finally abandoned as a result of the victorious spread of Islam, local worship has maintained itself with equal tenacity and success by placing the well-known Muhammadan version of the legend of the ` Seven Sleepers ' at the much-frequented Mazâr of Asahâb-Kahaf immediately below the mouth of the gorge (Fig. 312). Like previous visitors with archaeological interests, we found a kindly welcome there in the spacious house of Kare Akhûn, Chiraghchi, the chief attendant of the Mazâr.
The many cave-shrines and temples to be found along both sides of the gorge for a distance of about a mile above its mouth had suffered a great deal of destruction both from vandalism and from treasure-seeking operations, even before Dr. Klementz furnished the first brief description of them.9 Where, owing to the accumulation of heavy debris or for other reasons, manuscript remains and other antiques were likely to have escaped local exploitation of the kind practised at Idikut-shahri, the caves had, for the greater part, been carefully searched, with important results, in the course of Professor von Lecoq's Turfan expedition of 1904-5.10 Subsequently, in 1907, Professor Grünwedel had devoted his expert iconographic knowledge and artistic skill to the study and record of whatever paintings had survived on the walls of the more important caves and temples.1' Since then destructive diggings by natives had proceeded unchecked in spite of the diminishing yield of antiques. They were said to have been particularly stimulated by Mahmûd ` Jisa ', the Kara-khôja headman, who, before he fell a victim to Ahmad Mullah's émeute in the preceding spring, appears to have been collecting the proceeds in the shape of manuscripts, &c., for sale or as presents to Chinese officials at Turfan and Urumchi. The result of these operations was to be seen in the disturbed condition of some smaller ruins which in November, 1907, I had found still apparently untouched.'2
I was therefore obliged to confine my own work to those few spots where heavy accumulation of debris or other difficulties of the kind appeared to have deterred the diggers, and to the rescue