Gates are traceable near the middle of the longer sides facing north and south. The eastern side is adjoined by what appears to have been an outer enclosure, with a tower about 18 feet square at the north-east corner. The walls of the whole, varying from two to three feet in thickness, appear to have been built partly in large-sized bricks and partly in unhewn stone. Within the inner enclosure, which seems to have been utilized as a shelter by graziers within comparatively recent times, is .a jumbled mass of ruined walls enclosing courts and rooms, of which the surveyor was unable to disentangle the exact disposition and character. Neither from his plan nor from my rapid inspection of the ruin am I able to form any definite opinion as to its origin and date. But considering that this ground below Toghlan-shahr has certainly been abandoned for a very long period, I am inclined to believe that the enclosure is old, but that it was probably repaired and adapted for temporary occupation in Muhammadan times. I may add in conclusion that some 30o yards farther down the plateau edge there is the ruin of another but much smaller walled enclosure, measuring outside some 53 by 26 feet, which from the bricks used appears to be of similar origin.
My single day's halt at Tash-kurghan on September 12th was more than fully taken up with manifold arrangements for transport, & c., in view of the intended division of my party a few marches farther ahead. What I was able to see once more of the old capital of Sarikol or to learn about its present conditions did not modify the views concerning the antiquity of the site and the character of its extant remains 1ß that I have recorded in connexion with my previous visits. I may, however, briefly mention that not more than a few hundred yards below the north-east foot of the ruined town site examined in 1900, the identity of which with the Sarikol capital seen and described by Hsiian-tsang is not subject to doubt, I now noticed the presence of a Muhammadan shrine marked by a large ruined ` Gumbaz ' and an ancient cemetery adjoining. The former is supposed to shelter the remains of Shah Auliya, a renowned saint, whose grave attracts pilgrims from all over Sarikol. Is it possible that this pilgrimage place is connected somehow by continuity of local worship with the ` convent remarkable for the height and largeness of its towers and pavilions ' as well as for its majestic statue of Buddha, which, as Hsiian-tsang relates, an early king of Chiehp`an-t`o had built at the site of the former royal residence in honour of the venerated teacher Kumara-buddha i' 17
SECTION V.—BY THE KARA-TASH RIVER TO KASHGAR
On September 13th I set out from Tash-kurghan for Kashgar. I was anxious to use the chance offered by the season for exploring en route the Kara-tash valley, which lies on the most direct line from Sarikol to Kashgar but owing to special difficulties had never been surveyed. In order to reach it, it was necessary for the first few marches to follow again the main caravan route connecting Sarikol with Kashgar and Yarkand across the great southern buttresses of Murtagh-ata. As I proceeded by it to the high plateau of Chichiklik via the Dershat valley I felt sure of being once more on the track of Hsiian-tsang.
Already on my passage of June, 1906, I had convinced myself that Hsiian-tsang's journey had led him across the Chichiklik.1 But the route up the narrow Shindi valley,2 which the early season had then obliged me to take, was not likely to have been that of the great pilgrim when he
is See Ancient Khotan, i. pp. 35 sqq.
17 Cf. Julien, Mémoires, ii. p. 213 ; Ancient Khotan,
i. p. 37. [I may correct here an error made in the last quoted passage when discussing Hsiian-tsang's notice about the
royal residence at the capital of Chieti-p`an-t`o. The
mention of ` an enclosure of some 30o paces ' refers not to this place but to the ruined fastness of Kiz-kurghan.]
1 Cf. Serindia, i. pp. 76 sqq.
2 See Desert Cathay, i. pp. 97 sq.