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

which were traced only just before my departure from the site, and could therefore not be surveyed
with accuracy. The wall, only about 2½ feet in thickness, had been reduced by wind-erosion to
a low, in many places almost imperceptible, mound; but its line could be followed north of the fort
for a distance of over 700 yards running east to west, being best noticed where it occupied the top
of rudimentary Yārdangs about 1–2 feet high. It approached the north corner of the fort within
about 100 yards. Of its west and east faces, which were more exposed to wind-erosion, only small
sections survived; the southern one I could not locate within the limited time left.

Former
agricultural
settlement. A wall of this unsubstantial sort could only have served, as it were, police purposes. In this
respect it could be compared with those equally flimsy walls which are to be found at the present
day round several of the Turfan towns and villages, and which never fail to arouse derisive comment
on the part of Chinese accustomed by tradition to seek safety behind far more solid ramparts. The
soil around Chong-hassār is a fertile loess singularly free from that shōr, or salt efflorescence, which
thickly covers the ground near the north shore of the present lake-bed. It only needs water to be
brought under cultivation. As this is carried even now by the intermittent overflow of the Lukchun
canals to within four miles or so of the site, I believe we may safely assume that Chong-hassār in
Uigur times, and probably earlier also, had its agricultural settlement enjoying irrigation, and that
the change which has come over the ground since is due to desiccation.

Ruined
shrines
of Kichik-
hassār. From my camp at Chong-hassār I was able to examine and, with the help of additional labour
easily secured from Bēsh-tam, thoroughly to clear also an interesting group of small shrines known
as Kichik-hassār, the 'Little Castle'. It is situated a little over two miles to the north-east from
Chong-hassār, and reached over ground where vegetation is very scanty and the effect of wind-
erosion distinctly more marked. The trenches cut by it into the loess soil all run from west to east
and attain in places a depth of 5 to 8 feet. On approach to the ruins drift-sand is met with, heaped
up in small 'Barkhāns' to a height of 8–10 feet. It is, perhaps, due to the protection afforded by
the dunes that two of the ruined shrines had preserved their essential features. As the sketch-plan
in Plate 50 and the photograph in Fig. 269 show, the site comprises a number of small Stūpas, with
little domed cellas adjoining them, as well as some much-decayed structures near them which
may have served monastic purposes.

Area of
'Tati' type. The area occupied by the ruins stretches for about 120 yards from north-west to south-east, the
direction in which its extent is greatest. Within this area the ground exhibits a typical 'Tati'
character, being strewn on its eroded surface with small débris of pottery and bone fragments, most
of which seemed human. All structural remains occupy erosion terraces, rising 4–5 feet above the
adjoining ground. On the east or lee side of the better-preserved ruins the drift-sand was found
heaped up to a height of 10 feet or so, as seen in Figs. 269, 270. The bricks used in them are all
sun-dried, but relatively hard, with the fairly uniform size of 18″×8″–9″×4–4½″, the same as found
at the ruins of Chong-hassār. The masonry is very regular, all bricks being laid in horizontal
courses, often with their long and short sides alternately facing outwards. All structural details
seemed to point to an approximately contemporaneous origin of the buildings.

Stūpa and
domed
cella I. The largest complex of ruined structures is that to the west, marked I in the plan. It com-
prises a remarkably well-preserved Stūpa at its north end, with a domed cella and vaulted ante-
chapel adjoining (Fig. 270). Further south lies a hall or court, 47 feet by 23, which may possibly
have had a timber roofing. The purpose of the small arched niche in the west wall could not be
made out. A much-decayed set of rooms adjoining to the south-west may have served for the
quarters of monks. The Stūpa at the north end rises on a base, 24 feet square and 5 feet high,
partly covered by sand. On this is set a circular base, 15 feet in diameter and 4 feet 6 inches high.
This again is surmounted by a tower-like member close on 10 feet in height, representing the third