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

road lines crossing this area, particularly in its eastern portion where a main street leading up from
the east gate to a ruined temple raised high on a clay mound could be followed with ease also on
the ground. But even along this it was easier to make sure of flanking refuse heaps than of the
position of buildings which had crumbled away almost completely. Most of the structures within
the walls, and probably all private dwellings, had been built of stamped clay and timber walls of no
great thickness, and these walls were bound to decay rapidly once occupation had ceased. Wilful
destruction and long-continued burrowing for 'treasure' at a site so easily accessible from ground
still inhabited have probably accelerated the process, and at certain points traces of fire were unmis-
takable. This levelling of the interior had doubtless already proceeded far by the time that the
accumulation of drift-sand outside overtopped the walls. Hence the protection which dunes might
have afforded to ruined structures had here been of no avail.

Limitation It was easy to realize from the first that a complete clearing of the interior could not be attempted
of search. with the small band of inefficient Mongol diggers. It would have implied a disproportionate sacrifice
of time on our part, even if it had been possible to persevere with it in spite of the difficulties due to
refractory workers, the heat, the dust-storms, and the scarcity of water. So I decided to confine
our search to those few among the temple ruins recognizable in the western portion of the town
where layers of debris still remained likely to hide objects of interest, and to the extensive refuse
heaps elsewhere which held out some promise of fragments of writing, &c., capable of affording
chronological evidence. In this hope I was not disappointed, and as the finds of written records
in these places, together with the coins picked up within the walls and immediately outside them,
removed from the start all doubt as to the period down to which occupation must have extended,
we may review them at once.

Written Among the deposits of rubbish, composed mainly of stable refuse, chippings of wood, broken
remains pottery, &c., those found along the sides of what appeared to have been the chief thoroughfares
amidst
refuse. were the largest. They lay mostly in places where no remains of substantial structures adjoined,
and the conclusion seems justified that they had chiefly accumulated close to modest habitations
which may have served as booths, inns and the like, and which, in consequence of their poor con-
struction, have suffered complete effacement. Apart from a few fragments of wooden objects
bearing stray Chinese characters, all the inscribed remains are of paper, and their general condition
leaves no doubt that they had found their way into the rubbish-heaps as 'waste papers'. The great
majority of the documents are torn pieces, many of quite small size, while others, though apparently
complete, were found twisted into rolls or otherwise crumpled up. In some cases a number of strips
bearing Chinese characters had been tied together in knots. Among the records thus recovered,
those in Chinese were by far the most numerous, and so far as appeared from a hasty examination
at the time, all, with the exception of some printed pieces, were hand-written. A rough inventory
prepared before the submission of the Khara-khoto materials to different collaborators shows a
total of some 230 Chinese documents and fragments from this source, as against 57 pieces in the
Hsi-hsia or Tangut script, close on half of these being printed. Of fragments showing Tibetan
writing only three were found, in addition to a Chinese document with a Tibetan endorsement and
two pieces containing writing both in Hsi-hsia and Tibetan. Finally the yield of the refuse heaps
includes 9 pieces inscribed in Uighur script and Turkish language [see Prof. von Lecoq's App. K].

Chrono- As no data have yet reached me from the two scholars who have kindly undertaken the examina-
logical
bearing of tion of the Chinese and the Hsi-hsia and Tibetan remains respectively, I am unable to surmise what
documents. indications of definite chronological bearing, if any, these 'waste papers' from the rubbish-heaps
of Khara-khoto may furnish. The presence of Hsi-hsia pieces, both written and printed, suffices,
however, to prove that the town must have been inhabited during the period of the Hsi-hsia dynasty