from what remained of the projecting mass of rock that once formed the core of the statue ; it pointed to a colossal standing figure.
Considering the only too obvious effects of moisture on these cellas, the hope of recovering recognizable remains of relievos or more interesting relics under the heavy masses of débris that filled the interior in parts seemed too small to justify the heavy expenditure of time and labour which a complete clearing would have cost. But on the top of the ridge, which at its eastern end has evidently been occupied by a number of small structures, either shrines or monastic quarters, I had the two cellas vi and vii cleared. From a few fragments of painted plaster brought to light in vii it appeared that these, too, were places of worship. In addition to the walls shown in the plan, which probably served mainly to secure level building space, there were remains of terraces built up against the north-east end of the ridge, at a height of 20-3o feet above the level ground. These, too, may once have borne small structures.
The only ruins at Ara-tam which still remain to be mentioned are six small cellas, some only a few feet square, found perched in a line, as Fig. 192 and Plate 48 show, on little spurs jutting out from the foot of the hill directly to the north of A. III. They were found completely empty and were declared to have beet, searched two years earlier by Dr. von Lecoq, who had paid a flying visit to Ara-tam when on his reconnaissance tour eastwards on behalf of Professor Grünwedel's Mission. With what result I have not beeit able to ascertain.
My own search of the ruins had yielded no definite chronological evidence. But, in view of the close resemblance which the remains of the fresco decoration in the rock-cut cellas showed to designs familiar to me from the later cave-temples of Chien-fo-tang, it appears to me highly probable that the Ara-tam shrines dated from the period of Uigur dominion (ninth-twelfth centuries), during which Hami is likely to have enjoyed protracted spells of peace and prosperity. It is more difficult to guess the date at which Buddhist worship had finally ceased at the site. We have seen above that at the time of Shah Rukh's embassy (142o) Buddhism still continued to be professed at Hâmi by the side'of Islam. Clear chronological evidence, such as a site definitely abandoned to the desert might have easily yielded, was not to be looked for at a place which, favoured by its abundant water-supply and fertile soil, must have ever invited continued occupation. That the advantages and facilities for archaeological work which are offered by ruins within the cultivated areas, along the south foot of the Tien-shan and far away from the desert, have their antiquarian drawbacks also is a lesson I first learned amidst the pleasant surroundings of Ara-tam.
On November 2 I left Hami for Turfan, after having completed the safe packing in twelve wooden cases of the additional manuscript acquisitions from Ch`ien-fo-tung, which so far had to travel in huge bags and without adequate protection. Regard for the available transport and our survey-work induced me to follow the rather circuitous high road which, for the sake of wells and some chances of grazing, keeps close to the foot of the Tien-shan.7 By doubling marches where
The first portion of the line followed by the high road, section of the more direct route from Hami to Turfân, which
as far as Ch'i-ku-ching (Map No. 66), lies on what must always leaves the high road at Toghucha and strikes across abso-
have been the most direct route from Hami to Guchen lutely barren hills and plateaus to Chik-tam ; see Map n in
(` Posterior Chu-shin', the later Pee-tine) and the other fertile the Russian publication of Captain Roborovsky's expedition
territories along the north foot of the Tien-shan. Whenever which first surveyed it. Donkey caravans to Turfân frequently
these were safely held by the Chinese, traffic from Hami to use this route in the winter. It is this route which the
wards the West is always likely to have followed this northern Chinese envoy Wang Yen-tê followed in A.D. 981 ;
route during the summer months in preference to that leading cf. Chavannes, T'oung-pao, 1905, p. 53o, note.
via Turfân. The further portion of the high road to the Further south lies the track, quite waterless, which leads
latter leads over very barren stony slopes as far as Chik-tam, from the Shona-nbr depression south-west of Hirai to Chik-
but offers at least halting-places with water. tam. It was first followed by Col. Kozloff in 1895 and
Of this there is practically none now on the corresponding surveyed again in 1914 by M. Muhammad Yâqùb under my