国立情報学研究所 - ディジタル・シルクロード・プロジェクト
『東洋文庫所蔵』貴重書デジタルアーカイブ

accumulated within the enclosed area, and it had so far succeeded in protecting the west wall from
being breached, though its top showed incipient cuttings. At the north-west corner of the walled
enclosure a massive watch-tower, built of clay and containing at its foot a passage vaulted with sun-
dried bricks, still rose to a height of about 30 feet. No doubt, when in the course of time the Prospect of
cuttings are carried down deeper, the wind will regain full play over the sand at present filling the progressive
interior and drive it out through the then breached wall westwards. Then erosion will set to work wind-
within the walls as thoroughly as it has outside, and convert the enclosed area, at present half- erosion.
smothered under dunes, into an eroded 'Tati' with remnants of the north and south walls only to
show that a walled town once stood here.

It was interesting to note the incipient stages of this process. In the area between the outer Interior of
and inner east walls the sand lay quite low, leaving the small mounds from dwellings, etc., as seen ruined town.
in the foreground on the right of Fig. 185, exposed to erosion. Consequently fragments of old
pottery, charcoal, and similar débris were to be found here on the surface, though not in such plenty
as on the completely eroded soil outside. Most of the tamarisks in the outer enclosure were dead,
and the sand-cones that they held together were being disintegrated and levelled. Behind the
inner east wall the enclosed area contained many tamarisk-cones with bushes still green, and near
the west wall it showed even some flourishing wild poplar and Eleagnus (Jigda) trees. In most
places the accumulated drift-sand lay to a height of from 6 to 10 feet. Only at relatively few points,
where larger refuse-heaps or mounds of clay, evidently formed by the débris of completely decayed
buildings, rose above the drift-sand level, could fragments of pottery, including porcelain and glazed
stoneware, and bronze coins of the T'ang period be picked up on the surface. To the west of the
town, both within and beyond an outer enclosure marked by traces of less massive walls, there
stretched a strip of ground covered with plenty of scrub and reeds suggesting that subsoil water
reached there. It probably is derived from the drainage which the dry stream bed, mentioned above
as debouching to the south-west of the site, carries down from the third hill range to the south.⁸

The porcelain and glazed pottery fragments, together with the coins picked up on my first Occupation
inspection of the site, made it at once clear that regular occupation of the little town and the proved
adjoining ground had continued down to Sung times, if not somewhat later. This fact, and the down to
obvious decay which remains of a perishable nature must necessarily have suffered on ground still Sung times.
supporting vegetation, made it inadvisable to sacrifice time to systematic excavations. The heavy
layer of drift-sand within the largest portion of the walled enclosure, as well as the difficulty about
raising an adequate number of labourers at that season, would have protracted them too much.
Experimental digging, carried out on a somewhat larger mound rising above the drift-sand towards
the south-west corner of the walled area (marked in plan, Plate 46), disclosed what seemed remains of
a later dwelling that had been built on the top of a great débris heap made up mainly of completely
charred wood, decomposed sun-dried bricks, and stable refuse. There were indications elsewhere
also that the structures occupying the town site at the time of its virtual abandonment had suffered
from a conflagration.

Complete desolation, however, does not seem to have come over the little town all at once, but Later visits
as a slow lingering death. This is suggested by the fact that, while the finds of coins stop short with to site.
a piece belonging to a regnal period of the Chin dynasty (A.D. 1156–61), and the great majority of
the specimens of glazed pottery detailed in the List below belong to T'ang or Sung times,⁹ there are
also a few fragments of porcelain to which Mr. Hobson is inclined to assign a seventeenth-century
origin.¹⁰ Considering the proximity of the site there is nothing unlikely in the statement, heard from