Sec. ü] EXCAVATION OF ANCIENT RESIDENCES, N. II, N. III, N. iv. 335
stances. The relatively large size of the house similarly helps to suggest that this was the residence of a man of means, and possibly in authority—an impression that led my labourers to christen it the ` Yamên.' In spite of the expectations thus raised or, perhaps, just because such hopes had attracted searchers to the abandoned dwelling long before the sand had completely invaded it, no objects of any kind were found in the other rooms cleared. In view of this it appeared improbable that the large area to the east of the hall, which looked like an open courtyard, should retain antiquarian relics ; and as its excavation, owing to the depth of the sand, would have cost even more time than that of the hall, I decided to leave it unopened. Of the small detached structures to the west (see plan, Plate XXX), built with walls of rushes and plaster, it was easy to ascertain from the accumulations of animal refuse that they had served merely as cattle-sheds or stables.
The clearing of the ruined dwelling-house (N. iv) situated to the west of N. ni at a distance Excavation of a little over a furlong, proved far easier, since its broken walls retained less of drift-sand, of ruin N. iv. while at the same time it was attended by results equally curious. The plan of this house and its immediate surroundings, reproduced on Plate XXXI, is a good illustration of the typical arrangement of these ancient residences. The disposition of the rooms, of the adjoining orchard and arbour, &c., recalled what I had observed in modern Khotan dwelling-places of some pretensions. The effects of erosion of the ground had not caused so much damage here as, e.g., at N. i and N. in ; on the other hand, the thinness of the walls, the plaster of which was put against horizontal layers of reeds, had caused much of the superstructure to decay and the protecting cover of sand to be reduced in consequence. Among the rooms ranged on both sides of a central passage the one marked N. x., though covered only by about one foot of sand, proved particularly rich in interesting objects. There could be no doubt that we were clearing the miscellaneous waste left behind in an ancient scriptorium or ` Daftar,' to use the comprehensive term of the modern East, when there emerged from close to the floor, besides a number of Kharosthi documents on rectangular, oblong, and ` Takhti '-shaped tablets (see N. x. 2. C in Plate XCVII) a variety of writing implements and unused wooden `stationery.'
The four roughly-cut wooden styles consisting of tamarisk twigs sharpened at one end, of Writing
which two specimens are reproduced in Plate CV, put me at once in mind of the remark of the implements
Tang Annalist that the people of Khotan ` used pieces of wood instead of brushes [for writing]' b. N. x. The subsequent discovery of the far more finished wooden pen from N. xv. (see same plate) disposed of any possible doubt on the point. Two oblong tablets (N. x. 9. a, b), which together with the styles were found on the low plaster platform running round two sides of the room, are blank ; their curious shape, with a handle at each end, suggests their having been intended for labels. Another blank tablet (N. x. 9. c, see Plate CV), io in. long and only half an inch wide, presents points of special interest. Its very narrow shape and general resemblance to the slip-like pieces of wood on which the subsequently-discovered Chinese documents of N. xv. are written 6 make it appear distinctly probable that it, too, was meant to receive a single row of Chinese characters. Now N. x. 9. c differs from the inscribed pieces of N. xv. by showing a hole at each end, and these holes (one of them pierced through a small handle) could manifestly have served only for stringing purposes. We shall see later that the Chinese documents on wood found in N. xv., plentiful as they were, do not solve the question how records extending over more than one row of characters were conveniently ranged and kept in order. Is it possible that in this piece (N. x. 9. c) we have the solitary example of that kind of wooden stationery
6 See Rémusat, Ville de Khotan, p. 33 ; now Chavannes, ° See below, pp. 358 sqq. ; for illustrations, Plates CXII-
Tures occid., p. 125. CXIV.