National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

164   ON THE WAY TO LOP-NOR   [Chap. V

ch`êng or ` Stone Town ' of the Tang Annals.2 Similarly as regards an earlier period, I have shown it to be very probable that the town of I-hsiin -is, which in the Han Annals and in Li Tao-yüan's commentary on the Shui-thing is mentioned as a chief place of the kingdom of Shan-shan or Lou-lan and the site of a Chinese military colony established in 77 B. C., must be looked for within the present oasis of Charkhlik.3

Though the length of my renewed stay at Charkhlik was directly due to the practical difficulties above referred to, I was fortunately able to apply it also to profitable archaeological work. Within the oasis itself, indeed, the only ancient structure traceable, beyond the remains already described in Serindia, was a small mound of solid brickwork, about 15 feet in diameter and situated near the centre of the ruined circumvallation or sipil. By its shape and the large size of its sun-dried bricks, about 18" x 9" x 31", it suggested the base of a Stûpa. As it was adjoined on two sides by modern dwellings no closer examination was possible. But to the south of the cultivated area I was able to explore two small sites that had before escaped my attention, and these proved of some interest. Both had been visited in 1910-I I by Mr. Tachibana, the Japanese traveller, and the remains of both showed evidence of having been superficially searched in places.

Following the right bank of the main bed of the Charkhlik river for about a mile beyond the southern edge of the present cultivated area, I reached a small ruined enclosure known as Koyumal, standing on the bare, gravel-covered, alluvial plain. Its much-decayed walls, built of sun-dried bricks about 8 feet thick, appear to have formed a somewhat irregular quadrangle of which the eastern face measures approximately 218 yards (see plan, Pl. 8). As the west wall had been carried off completely by an encroachment of the river, the lengths of the other sides could not be exactly determined. It therefore remains doubtful whether the enclosed area was intended for a square or an oblong. Near its centre rise the remains of what undoubtedly was a Stûpa base, 28 feet 3 inches square, to a height of about 14 feet. The sun-dried bricks used for its masonry measured as elsewhere 17 x 9 x 4 inches. A passage, a little over 4 feet in width, was traceable on the north, south, and east between the Stûpa base and a much-decayed enclosing wall (see inset, Pl. 8).

On the west there stood, at a distance of about 9 feet from the base, what appear to have been two small Vihâra chapels, each about 20 feet long and 9 feet wide ; between them stairs about 8 feet broad seem to have led up to the foot of the Stûpa proper. But as all the masonry to the west of the base had decayed to within a foot or so of the floor level, the constructive details could not be determined with certainty. The remains of five small niches, separated by stucco pilasters and each retaining the feet of a standing stucco figure, could, however, be traced along the east wall of each chapel (Fig. I II). In the niche nearest to the north side of the stairs the legs of a robed figure in stucco were preserved up to the knees, i. e. to a height of 14 inches. In each of the chapels there survived near the foot of the stairs an oblong pillar base in wood, measuring about 16 inches by io, with a raised circular socket.

From the debris covering the floor of these chapels numerous fragments of painted plaster emerged (Koy. 1. 05-46), which had undoubtedly once formed part of a large floral background in the mural decoration. Most of the fragments show lotus-petal or acanthus-like imbrications. The fragment of painted wood marked Koy. i. 03, showing plant motifs, may possibly have belonged to one of the wooden pillars of which the bases were found in situ. Among several fragments of carved wood may be mentioned one from an open-work detail, Koy. of (Pl. XVI), and part of a gilded right hand, of life-size and well modelled, Koy. 1. 02 (Pl. XVI), which, by the clearly shown web between the fingers, is proved to have belonged to a Buddha figure. Among the stucco relief fragments, some of which may have belonged to the figures already mentioned,

2 See Serindia, i. pp. 318 sqq.   3 Cf. ibid., i. pp. 325 sqq•, 342 sq.

Ancient remains at Charkhlik.

Site of Koyumal.

Remains recovered at Stiupa.