Sec. Hi] RECORDS FROM A HIDDEN ARCHIVE, N. XXIV 225
extreme western edge of what had then appeared to be the central portion of the ancient oasis. Only one of them, N. xi', the westernmost of the group, could be cleared then, and that not quite completely, no ` finds ' rewarding the labour.
Our work on the morning of October 24 began in a very promising fashion at the badly Clearing
eroded remains of a house, situated about one-third of a mile to the south-east of N. xII (see plan, dwelling,
Plate I3). The three rooms still traceable with their timber and plaster walls were covered nowhere N. XXIII.
by more than a foot of drift sand. But this had sufficed to save over thirty wooden documents, though on some the Kharosthi writing had suffered badly by exposure. Rectangular and wedge-shaped tablets prevailed, and on one of the latter, N. xxiii. ii. 8, there survived a clay impression from the same classical seal which I had first encountered on tablets excavated at N. xiII.1 Among the miscellaneous antiques found there may be mentioned a curious piece of superior basket work, N. xxiii. 003, apparently of bamboo, and of Chinese origin ; a band of green' leather, N. XXIII. ooi, once lacquered and probably belonging to scale armour ; an incised wooden stamp, N. XXIII. 002 (Plate xix) ; and an oval, trough-shaped piece of horn, N. XXIII. ii. oo5, which may well have served as an inkstand. The close resemblance in the make and decoration of the lacquered bowl fragment N. XXIII. i. ooi to a piece found at a station on the ancient Limes west of Tun-huang suggests Chinese manufacture for all such lacquered ware. The fence enclosing the courtyard had withstood erosion far better than the ruined dwelling itself, and could be traced for upwards of 130 feet in one stretch from south-east to north-west with adjoining extensions. The big poplars once lining it lay as dead trunks stretched out in a row, while to the south dead fruit-trees strewed the ground marking an orchard.
From here I turned my diggers to the large ruined residence, N. xxiv, situated about a quarter Excavation
of a mile to the north of N. xxiii and nearest to N. xi', with which my excavations of 1901 had of residence
concluded. Its remains, shown by the photograph (Fig. 59) in the course of excavation, occupied
a small plateau which ground eroded to a depth of about sixteen feet surrounded on all sides except
on the south. The plan reproduced in Plate 14 shows the arrangement of the numerous
apartments, of which the walls, mostly built in timber and plaster, still stood some height above
ground or were otherwise traceable. But that this residence was once even larger was shown by
the débris of timber strewing the slopes of the extant plateau to the north and east. Here on the
east were found also the dead trunks of poplars, some still upright as in Fig. 59, which must have
belonged to rows adjoining a courtyard or garden. It will be seen from the plan that, while the
living-rooms occupied the centre and east side, the outhouse and stables lay westwards.
On commencing systematic clearing from the north in what remained of room i we came first Finds in
upon some badly bleached and warped tablets which had lost all their writing. The room adjoining rooms i-iv.
it in the north-east corner had already been cleared in the course of my hurried examination of igoi,
when the only find was a large jar, nearly three feet in diameter and let into the floor.2 From
rooms ii and iii came some tablets in better condition, two still showing traces of their Kharosthi
writing. The inner room, iv, was provided with plastered sitting platforms on three sides, after
a fashion which, as the plan of a modern residence at Bâgh jigda, reproduced in Plate 12, shows,
still survives to this day in timber- and plaster-built houses of well-to-do people in this region. Some
of the posts of the wall framework which retained their original height proved that the ceiling had
been about eight feet from the floor. Two large beams, twenty-two feet long, had once supported
the ceiling and now lay stretched out on the sand which covered this room to a height of four
to five feet. To judge from the massive nature of these beams it seems possible that they were