National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

in Arabic writing (cf. Mi, 209). The confusion was the easier as the Phison itself, one of the four
rivers of Paradise, was believed to have an extraordinary course across almost the whole of Asia
and, according to Marignolli for instance, to flow through China, India and the countries north of the
Caucasus (Hallberg, 405-407).

The name « Quian » used by Polo was known to Mussulman geographers at least three and
perhaps four centuries before the date of our traveller, and it has been preserved in two later compila-
tions, the Ḥudūd al-ʿĀlam of 982-983, and Gardīzī of the 11th cent. The true form قيان Qiyān
or Qiān appears only in Gardīzī's text, but the قنان ʿInān of the unique Ms. of the Ḥudūd al-ʿĀlam
is certainly a misreading for قيان yiyān (or perhaps قيان Qiyān), as has been shown by Minorsky
(Mi, 70, 206, 229-231). Unfortunately, these second-hand compilations are so full of misreadings
and of contradictions that most identifications must remain extremely doubtful. Among the
eastern rivers, the Ḥudūd al-ʿĀlam mentions the Ḥumdān river, the كسو Kisau river which,
after it enters the boundaries of Buyšūr, is called ʿInān (read yiyān), and the river وجاب Wajāb;
these three rivers empty themselves into the Eastern Ocean. Ḥumdān is well known as the ancient
name of Hsi-an-fu among Westerners (see « Quengianfu », and cf. Mi, 229). Naturally enough,
Minorsky says that the river of Ḥumdān must be the Huang-ho, and that the yiyān must be the
Yang-tzŭ. In the name « Kisau » given to the upper course of the yiyān, Minorsky proposes to
see a corrupt reading of كنسو K.nsw, which he thinks would represent 金沙 Chin-sha (Kin-ša)
of Chin-sha-chiang, the ordinary Chinese name of the Upper Yang-tzŭ (see « Brius »). But this is
hardly possible. Down to the 15th cent., 金 chin was pronounced kim, and we ought to have
*K.mša, not *K.nsw. « Kisau » remains unexplained.

The case of the other names is also more intricate than appears at first sight. The Ḥudūd
al-ʿĀlam and Gardīzī have drawn from more than one source, and although most of these sources
are lost, the texts have partly survived in quotations. For instance, such a late author as Dimašqi
(c. 1325), writing at a time when the name of Ḥumdān had sunk into oblivion, has two paragraphs
on the two rivers of Ḥumdān, the greater and the smaller; one of the two must be the Huang-ho,
but the other may be the Yang-tzŭ (cf. Fe, 367-368; see also Ibn Saʿid in Fe, 332-333; Abū-ʾl-Fidāʾ,
in Reinaud, Géographie, II, ii, 123). A similar confusion has perhaps taken place in the case of
the Qiyān. Gardīzī gives an itinerary from Čīnānčkāt (= Qočo, see « Carachoço ») to Ḥumdān,
which is certainly Hsi-an-fu. This itinerary, clearly corrupt in some of its indications, has been
studied by Minorsky (Mi, 229-232), who divides it into two parts, one leading to the Yang-tzŭ,
the other to Ḥumdān. I incline to take another view, and, provisionally, I would suggest the
following solution : Gardīzī divides the itinerary into two sections : (1) From Čīnānčkāt to B.γ-šūrā
(spelt « Buryšūr » and « B.γsūz » in the Ḥudūd al-ʿĀlam), where a river is crossed in a boat; (2) From
B.γ-šūrā to Ḥumdān. In the first section, the first place named is Qomūl (see « Camul »), then Ša-čū
(see « Saciou »), then Sang-lāḥ (?), then S.b-čū (Sub-čū; see « Succiu »), then Ḥam-čū (see « Campçio »),
then K.jā, then the crossing of the river Qiyān. For the second part, Gardīzī is content with saying
that it takes one month to go from B.γ-šūrā to Ḥumdān by a road provided with fortified stations
and postal stages. It is well known that the crossing of the Huang-ho took place at Lan-chou, and
I think that the Huang-ho may be the river which is crossed in a boat at B.γ-šūrā. In such a case,
B.γ-šūrā would be a name of Lan-chou, and K.jā would be Liang-chou (see « Ergiuul »). But we