Sec. iii] RELIEVOS AND FRESCOES FROM N.W. PORTION OF MING-OI' 1199
decorated circular base, about 16 inches in diameter, which is seen complete in Fig. 290 and of which one half was safely brought away, Mi. xviii. ooi (Plate cxxxviii).° The base was decorated with a floral design in relief, exactly alike on the front and back semicircle ; its brilliant and harmonious colours had remained in excellent preservation. The main motif consists of a vase in the centre from which spread horizontally flowers resembling carnations and acanthus leaves. The same rich colouring survives on the large fragments from a seated and a standing Bodhisattva, Mi. xviii. 002, 009, both life-size (Plate CxxXVIIi).
In much better preservation was the well-modelled relief statue seen in Fig. 29o, showing a Bodhisattva in richly embroidered garments seated on a lotus base. The floral designs on the short upper coat and the robe enveloping the legs were very naturalistic in style, and recalled the workmanship of Chinese figured silks. This large image was found thrust into the entrance of the north passage, and may have been thrown down from the platform on the corresponding side of the cella. Its weight was much too great for removal. So, when the time came to have this like the other excavated- shrines of the site reburied, I had it placed once more in the vaulted passage for safety. Two large torsos (see Fig. 29o) of standing figures, which were found prostrate in front of the cella and had suffered far more damage, may have belonged to images of Dvarapalas. The life-size stucco arm, Mi. xviii. 003 (Plate CXXXVIII), showing a fine monster's head as shoulder-piece, must have belonged to a Lokapala. Among other relief pieces, all still retaining their painting, may be mentioned the life-size Bodhisattva head, Mi. xviii. ooto (Plate CXXxI), the curious figure of a child, Mi. xviii. oo6 (Plate Cxxxv), and the large hand holding a Buddha relief medallion, Mi. xviii. oo5 (Plate CXXXVII).
The plaster surface of the walls outside the vaulted passage had disappeared everywhere, except in the west corner of the antechapel. There, at the foot of the wall, survived the curious fresco fragment, Mi. xviii. 0014, reproduced in Plate CXXVI. For a description of the scenes presented in its two panels I must refer to the List below. Here it will suffice to point out that the work in the upper one, though hasty, shows considerable skill in the spirited drawing of the dragon rising from the waves to attack a bare-legged man. In the lower panel nine short lines in Uigur script, by the side of the figures of the donors, remain to be read. The scene, or what survives of it, in the upper panel also still awaits interpretation. I may add here that a fragmentary paper leaf with Uigur writing was found on a low pedestal to the left of the cella entrance.
Before leaving the ` Ming-oi ' site I must briefly mention the conspicuous watch-tower solidly built of bricks which rises, as shown in plan Plate 51, on the top of the narrow ridge overlooking the main group of caves to north of the site. Owing to its commanding position the tower is clearly visible from the latter, as seen in Fig. 281. It measures about 24 feet square at its base, and still rises to a height of over 25 feet. When examining the tower on a hurried visit to the caves, I noticed the thin layers of reeds inserted at regular intervals between the courses of sun-dried bricks. They could not fail to recall the method of construction familiar to me from the ancient watch-towers of the Tun-huang Limes. But the true significance of the observation did not strike me until, when proceeding on my way from the Khôra site towards Korla, I passed another tower equally solid and of exactly similar construction perched on a high foot spur of the range, about 31 miles south of the former tower (see Map No. 49. B. I). I shall have to recur further on to the question of the probable origin of these towers, and need add here only that the ' Ming-oi' tower must in fair weather command a very distant outlook over the Kara-shahr valley and be easily visible from Baghdad-shahri, only about 5â miles distant in a straight line. Adjoining its north face are
' The weight of the whole was too great for transport had to be hollowed out to permit of safe packing as half
across the Kara-koram passes, and even the half removed a camel-load.