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

has to assume that an interversion has occurred in the pilgrim's account, and that he passed the
« great sand mountain » before reaching the Šitā or Yārkānd River.

This is merely a hypothesis, but one which finds a certain support in itineraries of the T'ang
period. Of course Kök-yar is not, like Yārkānd and Qaryaliq, on the main road from Kāšyar to
Khotan so that, if Chê-ch'u-chia or Chu-chü-po were Kök-yar, we should not expect to find it
mentioned among the stages of the route from Khotan to Kāšyar; but if one of the names be Yārkānd
and the other Qaryaliq, or if both be Qaryaliq, the absence of both is more surprising. Such is the
case, however, in Chia Tan's itinerary from Khotan to Kāšyar, dating from the end of the 8th cent.,
which is given in Hsin T'ang shu, 43 B, 15 b-16 a, and has been translated by CHAVANNES, Doc. sur
les Tou-kiue, 123. Not one of the names has been identified, but on some of them certain comments
are possible. In 635 and 648 respectively, « general governments » (tu-tu-fu) had been established
at Kāšyar and at Khotan; they had under their jurisdiction 15 districts (州 chou) for Kāšyar, and 5,
sub-divided into 10 in 675, for Khotan. The list of those districts has not been preserved, but some
of their names are given in Chia Tan's itineraries; others, sometimes the same, are included in a
curious list of « Hu districts » (Iranian and « Tocharian » districts) which are said to belong to Ho-hsi
(Tangut, Kan-su), but which are in fact all situated in the western part of Chinese Turkestan (Hsin
T'ang shu, 43 B, 8 a). In Chia Tan's itinerary, the first stage, 80 li south of Kāšyar, is 濱度州
Yen-tu-chou; this is evidently the same as the 賓度州 Yin-tu-chou of the list of « Hu districts ».
Although the name Yen-tu or Yin-tu has not been identified, one is tempted to locate that « district »
at Yangī-ḥiṣār. Going 160 li farther to the south-east, one reaches 雙渠 Shuang-ch'ü, « Twin
Canals », which was the ancient lodging-house (kuan) 掘俱 Chieh-fan (*Kjāt-b'j*ɐn). Going again
320 li to the south-east, one arrives at the town (ch'êng) of 郅支滿 Chih-chih-man (*T'śjět-t'śje-
muān), which is also called 績南州 Chi-nan-chou. I have no doubt that this is the 鄯及滿州
Ying-chi-man-chou, or « district of Ying-chi-man (*Iāng-g'jop-muān) » of the list of « Hu districts ».
The « River of the lodging-house (kuan) 繫 I (* ·Iei) » is crossed south-east of Chih-chih-man, and
then one goes east; from Chih-chi-hman to Khotan the distance is 670 li. CHAVANNES supposed
that Chih-chih-man, judging from the distances, was the place corresponding to Hsüan-tsang's Chê-
chü-chia.

I think that two of the above names are mentioned in the itinerary preserved by the Ḥudūd al-
'Ālam and by Gardēzī (Mi, 260). The second stage from Kāšyar to Khotan is called كريان Karyān
by the first text, كرمان Karmān by the second. Whether we correct both to كربان *Karbān, or the
first one to the Karmān given by the second, we have in either case a very satisfactory transcription
of Chia Tan's « lodging-house » of Chieh-fan (*Kjāt-b'j*ɐn). The next name is written وجشان
( = وجشان Učḥbān) in the Ḥudūd al-'Ālam, حجمان H.č.mān by Gardēzī. Although Chih-chih-man
(of which the « Hu district » Ying-chi-man must be a clerical corruption) does not come imme-
diately after Chieh-fan in Chia Tan's list, it is very tempting to read ججمان Čičimān, which would
be in exact agreement with the Chinese transcription.

But even if these equations be correct, identifications on the map are very hypothetical. If the
« River of the lodging-house I » were the Yārkānd River, Chih-chih-man might be Yārkānd itself.
The other name Chi-nan-chou would almost favour the identification, since it means « District South
of the Desert ». The « desert » would be the sand dune Bayram-qumī mentioned as being between