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

Earliest Limes record, 98 B.C.

Patchwork of figured silk fabrics.

Interest of textile design.

Tower T. xxli. d across Khara-nôr.

Signal-tower T. XXIII.

720   THE JADE GATE BARRIER   [Chap. XIX

of colours and makes. Among the documents T. XXII. c. 22, Doc., No. 271 (Plate Ix), claims special interest, as it is the earliest of all records from the Limes, bearing a clearly written date corresponding to g8 B.C. In it, too, reference is made to Ping-wang, the section of the Limes previously discussed. In Nos. 279-88 we have a series of daily returns specifying the number of men out of a squad of ten horsemen employed on various duties, such as mounting guard, cooking, and making bricks.

Among the very numerous fragments of fabrics silks prevail, but other materials, such as wool (T. xxii. c. 002), cotton ( ? , c. 009), and perhaps the paper mulberry fibre (c. 004), are also represented. The most interesting piece, however, is the patchwork, c. ooio a (see Plates LV and CXVIII for drawing of design), made up of small strips of a figured silk in indigo and yellow-green. The elaborate pattern has been fully analysed by Mr. F. H. Andrews in the Descriptive List. It contains figures of dragons, birds, and phoenixes set in highly conventional scroll-work, and other ornamental designs. The general style and treatment are unmistakably Chinese, as may reasonably be expected in a decorated silk dating from Han times, when China still claimed the monopoly of producing textiles of that material. The extremely fine weaving attests the high technical level reached in this essentially Chinese manufacture. The piece claims importance as one of the oldest specimens of textile art undoubtedly Chinese, and the indications it furnishes as to style and technique have since received striking confirmation from the abundant finds of brocades, dating from Han times, which I discovered in 1914 at cemeteries of the Lou-lan area.3 In view of the early date of the specimen and the certainty of its origin, it is all the more interesting to find in it certain elements of design, like the four birds grouped round what suggests a conventional tree, foreshadowing motifs which we are apt to consider as characteristic of textile art in the Persia of Sassanian times and in the Near-Eastern regions influenced by it.

Looking in the evening from T. xxii. c across the green expanse of the Khara-nor eastwards, I could sight no other ruined tower beyond except T. xxii. d, some five miles away by the southern shore of the lake. It was the last day, May 13, 1907, which I could spend over work by the Limes, and the distance from camp precluded a visit then. So the clearing of this ruin, and the search for other remains which might link the westernmost portions of the Limes now explored with those I had first surveyed to the north-east of Tun-huang, had reluctantly to be left for the future. The occasion for filling this gap duly came when I returned once more to Tun-huang by the ancient Lou-lan route nearly seven years later. The work was then successfully achieved, but its description must be left for the report on my third journey.

There was one more ruin, however, which, as it lay actually by the caravan route, I could visit when on my way back to Tun-huang. It comprised the two towers T. xxiii and T. XXIII. a, perched in a conspicuous position on the extreme northern end of a narrow plateau jutting out towards the Khara-nor. There the route to Tun-huang emerges from broken ground on an open marshy plain skirting the south-eastern shores of the lake, and itself turns to the south-east. The upper tower, T. xxiii, occupies a very narrow and steep crest, which affords no space whatever for quarters but, rising about r 10 feet above the plain, commands a distant view to the north and east. T. xxIl. d was visible from the foot of this tower, but could not be sighted from the lower one, T. xxiii. a. So the idea suggests itself that the former was built merely for signalling, or possibly was added subsequently when the inadequacy of outlook from T. XXIII. a had been realized. T. xxiii appears to have measured originally about 16 feet square at the base, but much of the coarse masonry had fallen down the steep slope, and the remaining portion stood to a height of only to feet. The material consisted of salt-impregnated lumps of clay with fascines of brush-

3 For a very curious figured fabric from Chien-fo-tung, Ch. oo i 18 (Pl. CXI), recalling this style, see below, chap. xxiv. sec. ii.