Sec. Hi] ON HSÜAN-TSANG'S ROUTE TO KASHGAR 79
of Tar-bâshi, and for over two miles beyond lies in the stream bed itself between high frowning rock walls, which in places overhang (Fig. 29). Owing to the flood from the melting snows, which completely fills the gorge, the route through it becomes quite impracticable during the summer months when the passage from the Chichiklik is diverted to the Yangi-dawan or Yam-bulak Passes. An exceptionally late spring allowed me to pass by the Tangi-tar route on June 5 ; but even then the deep pools of tossing water and big slippery boulders to be constantly crossed between almost perpendicular cliffs of limestone made the passage very trying and in places dangerous for the baggage.''
The conditions must have been much the same when Goëz' hard-tried caravan made its way down here in September or October of 1603. I could well understand on the spot the seriousness of the accident which here befell his faithful companion Isaac the Armenian. But there was for me the memory of an adventure far older and of a more famous traveller haunting this forbidding passage. Hsüan-tsang's biography relates how the ` Master of the Law ', while journeying from Chieh-p`an-t`o, or Tash-kurghân, towards the north-east, on the fifth day ` encountered a troop of robbers.18 The traders accompanying him were seized with fear and clambered up the sides of the mountains. Several elephants, obstinately pursued, fell into the water and perished. After the robbers had been passed, Hsüan-tsang slowly advanced with the traders, descended the heights to the east and, braving a rigorous cold, continued his journey amidst a thousand dangers. After having thus covered eight hundred li, he passed out of the Tsung-ling Mountains and arrived in the kingdom of Wu-sha.' Now the time occupied by the journey from Tash-kurghan to the point where the attack was encountered and the general description of the spot clearly point to some defile east of the Chichiklik, and there is certainly none offering the same natural facilities for such an exploit as the Tangi-tar gorge. As a competent observer has noted, ' a few determined men might in places defend it against an army '.1° The reference to the rigorous cold experienced on the onward journey is also significant. We know that the pilgrim crossed the Pamirs during the short summer, and spent fully twenty days in Sarikol. Hence he probably made his way over the Chichiklik and on towards Wu-sha and Kâshgar in the autumn. At that season none of the streams encountered on the route would be likely to hold sufficient water to prove dangerous to elephants excepting the Tangi-tar stream which, owing to the extremely confined nature of its rock-cut bed, retains deep pools of water even in the winter.
That the Tangi-tar gorge must have always been considered a portion of the route specially exposed to attacks is shown by the ruined watch-tower which rises at the lower end of the gorge where the latter joins the valley coming from the Yam-bulak and Yangi-dawan Passes further north. Its construction was attributed by my local informants to an ancestor of Ibrahim Beg, the headman of the Kirghiz grazing in the adjacent valleys. But of greater archaeological interest is the evidence I discovered of the early use of the Tangi-tar route at a very confined point of the gorge, about half a mile from its upper end (Fig. 29). There the rock walls on either side show a line of seven well-cut holes, about six inches deep and eight inches across, either square or circular, which were undoubtedly meant for the insertion of beams. A bridge or platform laid over these must have saved travellers and their animals the crossing of slippery and half-submerged boulders at a particularly awkward place. The work in these holes, chiselled out with much care and neatness, was, apparently, ancient.
17 See Desert Cathay, i. pp. 99 sq. ; for a description of the route cf. also Yarkand Mission Report, p. 267.
"a See Julien, Vie, p. 274 sq.; Beal, Life, p. 200. The former translation has north-west, a manifest error, as shown
by Of/moires, ii. p. 214.
" See Captain (now Colonel) H. Trotter's description of the defile, Yarkand Mission Report, p. 267.