216 THE NIYA SITE [Chap. VI
of vertically placed and closely packed rushes fixed to horizontal sticks which posts, less carefully carved, supported. These rush walls had almost everywhere lost their original facing of mud plaster ; yet here as elsewhere the rushes had kept their position remarkably well, probably because they offered less hold to the corrosive action of driven sand.
The clearing had scarcely commenced in the southernmost end room (i) when there emerged, almost on the surface, parts of a wooden chair bearing decorative carving which closely resembles in design and technique that of the ancient chair discovered in 1901 at the ruin N. 111.2 In the two legs recovered, N. XIII. i. 004, 005, as well as in the panel N. xIII. i. ooi which in all probability formed part of the same chair and is reproduced with them in Plate XIX, the most striking feature is a four-petalled and four-sepalled flower, frequent in Gandhâra ornament and directly derived thence.
As soon as the floor was reached, I had the great satisfaction of discovering a complete wooden document in Kharosthi. It was an excellently preserved double-wedge tablet, N. xIII. i. 1, retaining its clay seal impression intact. The same seal appears on two other documents, N. XIII. ii. 6, Io (Plates Xx, xxvii), found in this house, as well as on the tablets N. xxiii. ii. 8, 9, and was taken from an oval intaglio of undoubtedly classical workmanship. It shows a nude male figure, winged, which at first I was inclined to take for an Eros but which Mr. Woolley compares to the type of the Genius populi Romani appearing on coins of Diocletian.3 Under its outstretched left arm a much smaller figure, of which only head and arm are visible, seems to face it. The impression left by the square setting of the seal shows faint traces of characters, probably in Kharosthi and, perhaps, intended for the name of the owner.4
Among the other finds here the small wooden box, N. mil. i. 003, seen in the photograph, Fig. 52, may be noticed. The two string-holes found on either long side, and the rabbet running round inside the top edge, suggest that it was once closed by a lid which could be secured by strings sealed down on its top, after the fashion illustrated by such lids as N. xv. 345,6 or the lid of the ancient Chinese medicine case, T. VIII. 5, found at the Tun-huang Limes.° To judge from the size of the box it might well have been intended to hold wooden documents such as the usual Chinese ` slips ' found at N. xv, as well as at the ` Lou-lan ' site, and along the ancient Chinese Limes.
An ample harvest of ancient records in Kharothi script rewarded the clearing of the adjoining central room ii. It had evidently served as an office for the last dweller, probably a petty official, who had left behind his ` waste paper ' when the site was abandoned towards the end of the third century A.D. Here no less than a dozen wooden tablets came to light, almost all in perfect preservation. It was interesting to note that among the five double-wedge tablets, usually serving for brief, quasi-official, orders, as many as three were unopened and still retained their string fastenings intact. The method of fastening, already fully discussed in my former Detailed Report,7 is well illustrated by the reproduction of N. xIII. ii. 6 in Plate xxiv. Two of these tablets, N. xIII. ii. 6, to (Plates XX, XXVII), showed the impression in clay of the same classical seal which has already been mentioned. Also among the five rectangular double tablets here found one, N. xIIi. ii. 7, was still unopened. I shall have occasion below, in connexion with the ` archive ' discovered in N. xxiv, to discuss the reasons which explain why so many double tablets, especially of the rectangular type, had never been opened.8 A curious piece is N. mu. ii. i i (Plate XXVii), which provides an
2 Cf. Ancient Khotan, i. p. 397 (N. vii. 4) with Pl. Lxviii. • See below ; also Chavannes, Documents, p. r 26 ; Pl. XVII.
' Cf. Hill, Greek and Roman Coins, Pl. xV. 9. 7 Cf. Ancient Khotan, i. pp. 348 sqq.
4 Cf. Ancient Khotan, i. p. 355. ' Cf. below chap. vr, sec. iii; also Ancient Kholan, i.
° See ibid., i. pp. 361, 409 ; Pl. Cv, cxiv. pp. 353 sq.