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

a narrow passage built along the south foot of the tower. On either side of the doorway might be
seen the sockets meant for the insertion of heavy bolts, about 4 by 5 inches in section, which
once barred the door. In the narrow passage along the east foot of the tower, which seems to have
been intended for stairs but did not contain any, the original whitewashed plaster of the tower still
showed behind the rubble of hard clay with which the end of the passage had been walled up at
a later time.¹ The innermost room i is likely to have been used for the commandant's accom-
modation, and here half a dozen wooden records, all fragmentary, were found, besides the neatly
ornamented fragment of a lacquered bowl.

Miscel- The room iii, which may have been used for the men's quarters, yielded two wooden brackets,
laneous T. vi. c. iii. 002, 003 (Plate LIV), of the type already described which served for hanging up clothes,
relics from
T. vi. c. equipment, &c. There, too, was found the curious wedge-shaped wooden block, T. vi. c. iii. 001,
inscribed with two Chinese characters no longer legible, to which reference has already been made in
connexion with a similar find from T. vi. b.¹ᵃ It was strange to find a quantity of perfectly fresh-
looking horse dung, with remains of green reeds cut into straw, under the débris of the little
ante-room iv. This measured less than 7 feet across, and could have barely allowed the horse to
turn round. These tight quarters recalled life on board ship, and served to illustrate the need
of shelter which in such an inclement climate was felt here even by animals. Elsewhere, too,
I have made similar observations about stabling. In the entrance passage ii I came upon a neatly
decorated leather tongue, T. vi. c. ii. 002 (Plate CX), which probably formed the end of a saddle
strap, and what may have been part of a wooden lock, T. vi. c. ii. 001.

Wooden But the find which is of most interest was made just within the doorway or a few inches
tablet with
Early beyond. It was the remarkably well-preserved right-hand portion of a thin wooden tablet, T. vi. c.
Sogdian ii. 1 (Plate CLVII), 10½ inches in length and in its present state about 1½ inches across. It showed
script. at the top four short lines in an Aramaic-looking, then as yet unknown, script, which I had first
come across in a paper fragment of the Lou-lan site, L.A. vi. ii. 0104, and subsequently in
the important find of paper documents made at the Limes station T. xii. a. As I shall have
occasion to explain when discussing this find, the script, first partially deciphered by Dr. A. Cowley,
has since been identified as Early Sogdian by M. Gauthiot.² The untimely death of this gifted
collaborator has removed for the present all hope of securing some indication of the possible
meaning of the short inscription. But the very place of discovery and the external appearance
of the record may help us to realize its antiquarian import.

Script In the first place, I think, due stress must be laid on the fact that the tablet was found at an
points to
Iranian outlying watch-station of the Limes, far away from the line which the ancient trade route had
auxiliaries. followed. This speaks strongly against any assumption which might attribute the document, written
in a script of Western Asiatic origin, to traders or other mere passing visitors from that side.
Together with the wooden material, which is obviously local, this consideration made me wonder
at the time whether the discovery of the relic at the far-off post T. vi. c might be due to the
presence among its garrison of men drawn from that Iranian portion of Central Asia, Sogdiana,
and the adjoining regions, with which I was already inclined to connect both the script and the
language of the record.³

This conjecture has since found distinct support in certain documents of T. vi. b, briefly