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

were of 1/30 at Ch'üan-chou, but of 1/15 everywhere else. Cf. YS, 17, 7 b-8 a; 94, 10 b-11 a;
Yuan-tien chang, ch. 22.

Klaproth was right about Polo's « Gampu » and Chinese Kan-p'u, but he also tried to prove
that Kan-p'u was meant when the Arab travellers of the 9th cent. mentioned in China the great
harbour of خانفو Ḥānfū; I have no doubt he was wrong in this, and that the Ḥānfū of the early Arab
travellers is Canton. I explained the name in 1904 as Kuang-fu, a popular abbreviation of
Kuang-chou-fu which was then the full official name, and I then quoted (BEFEO, IV, 215) three
contemporary examples of Kuang-fu in Chinese texts; I could now add several more. The only
difficulty in this transcription is that we find in it ḥ- transcribing the initial k- of a Chinese
name, and that it might seem to weaken my argument in another case, when one of my objections
to Ḥingsai representing ching-shih (king-si) is that ḥ- is not a regular equivalent of Chin. k- (see
« Quinsai »). But we must not forget that Ḥānfū is a form going back to T'ang times, when the
Western writers knew little about China, and were not so systematic in their renderings as later
Rašīdu-'d-Dīn or Waṣṣāf.

It is true that Ḥānfū also occurs in Arabic texts of the Mongol period, and in such conditions
that those who, in the wake of Klaproth, still hold it to represent the name of Kan-p'u have been
obliged to suppose that the name had been unduly transferred by travellers and geographers from
the advanced port to the main city and was, in fact, a designation of Hang-chou itself. The use
of Arabic texts relating to China is extremely delicate, because they are a hodge-podge of data of
very different ages, with names desperately corrupt. Ferrand had contemplated a critical study
of them, but the volume of notes to his Relations de voyages has never been published. Yule's
and Cordier's indications in Cathay² are antiquated and often contradictory. I would not be
positive on the point, but I am under the impression that all the Ḥānfū of 13th and 14th cents.
Arab geographers are simply taken over from the travellers of the 9th cent. For instance, Abū-'l-
Fidā says that Ḥānfū, in his time, is called Ḥansā, and mentions the lake Siḥū, which can only be
the Hsi-hu (Si-hu) or Western Lake of Hang-chou; hence the conclusion that there are two Ḥānfū,
one being eventually Canton, the other being certainly Hang-chou. I rather think, on the contrary,
that there is only one Ḥānfū, the name of which Abū-'l-Fidā and his contemporaries merely knew
from books, and which he or they wrongly identified with Ḥansā, i. e. Hsing-tsai, = Hang-chou
(see « Quinsai »).

235. GASPAR

gaspar F, L, V, VB, Z guaspar LT, TA¹ jaspar FA, FB

One of the three Magi kings; cf. Y, I, 82-83, and the bibliography added here under « Baltasar ».
This is the most embarrassing of the three names. V. Scheil is not inclined to trace it back to
Gondophares, and the fact is that the two other names are not Iranian, but Semitic. He mentions