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

in my opinion, by the fact that both Hsüan-tsang and Hui-ch'ao speak of it not in their notices on
Kashmir and on the regions in the north-west of it, but when they are in the upper basin of the
Ganges. This does not mean, however, that I am tempted, with Francke and Herrmann, to
locate the Suvarṇagotra at Rudok: although Rudok may have formed part of it, the extent and
the seat of the kingdom remain undetermined.

Either the Suvarṇagotra or the Su-p'i (Sum-pa) may have been the Strīrājya of Indian history
and legend, but, as we have seen, the Suvarṇagotra (gSer-rigs) and the Sum-pa are clearly differentiated
in Tibetan prophecies. But these prophecies are retrospective and reflect conditions which pre-
vailed in T'ang times. The Shih-chia fang-chih says that the Suvarṇagotra or 'Eastern Kingdom
of Women' was the same as Great Yang-t'ung and that an authoritative historical text locates
Great Yang-t'ung near the sources of the Huang-ho. We may suppose that, at an early stage
of the Tibetan advance, the old 'Kingdom of Women' split into several parts, and that, while
some of its tribes remained south of the Kökö-nör, the central ones formed the restricted Su-p'i
kingdom of Central Tibet, and the western ones Great Yang-t'ung.

On the location of the Su-p'i kingdom of the 7th-8th cents., the itineraries translated above afford
fairly precise information, at least for its north-eastern and south-western frontiers. In the notice
of the Hsin T'ang shu on the To-mi, we are expressly told that the To-mi lived on the banks of the
Yak River, i. e. the Murus usu; but the itineraries show that they remained on the northern bank of
the river. The southern bank at least was in the territory of the Su-p'i, which extented west (read
« south-west ») to the Hu-mang Gorge. The original form of Hu-mang is not ascertained. The same
characters hu-mang (*kuɑt-mâng or *ɣuɑt-mâng) are used in the Hsin T'ang shu (221 B, 8 a;
cf. Laufer, Sino-Iranica, 385) to render the Middle Persian name of the fruit, Pers. ḥurmā,
but the transcription, somewhat abnormal, may have been influenced by the name of the Hu-mang
Gorge; in any case, there can be no question of « dates » in Central Tibet. The main itinerary gives
1 390 li from the Yak River to the stage (i) north of the Hu-mang Gorge, and 1 140 li from it to Lhasa.
The Hu-mang Gorge, to which the Tibetans came to meet a Chinese princess, can only be either
the Dañ-la or the « Ta-tsang-la ». With due allowance for the relative value of the estimates of dis-
tance in the text, it seems almost certain that it is the « Ta-tsang-la » which was known under the T'ang
dynasty as the Hu-mang Gorge; consequently, the Su-p'i kingdom of the 8th cent. must have
extended from the Murus usu in the north-east to the « Ta-tsang-la » in the south-west. This at least
is a fairly safe solution, perhaps the only one in a most complicated problem, rendered more obscure
by the vagueness and the contradictory statements of the various sources. I hardly need say that
the above discussion cannot be considered as final; it is primarily meant as a repertory of information
and an incentive for further research.

A few more words remain to be said about the Jo-shui, or « Weak Water », « Weak River », which
has been mentioned above in connection both with the K'un-lun and with the 'Eastern Kingdom
of Women'. The name already occurs in the Tribute of Yü of the Shu ching, then in the romanced
Mu t'ien-tzŭ chuan, and later in cosmogonic works like Huai-nan-tzŭ's chapter Ti-hsing hsün
(cf. Erkes, in Ostasiat. Zeitschr., v, 46, 72) and the Shan-hai ching; the Shuo wên writes 溺水 Ni-
shui, « Drowning water », which may represent the original word intended, but has not prevailed. As
far as I know, the earliest datable text in which the traditional explanation of the name occurs is Kuo