the Chien-ch'ang Valley]). Contrary to LAUFER'S opinion, the turquoise mines Polo speaks of were known there from Han times; on the Tibetan border turquoise was then used in particular for making arrow-heads. The problem of the name is more difficult. It has often been supposed
to be , Chien-tu, a name often met with in the pen-chi section of YS, but never mentioned
in the geographical chapters. In 1904 (BEFEO, Iv, 771), I remarked that Chien-tu, according to some texts of the pên-chi (YS, 13, 1 a, etc.), must have been near Burma since it was prevented by Burma from entering into regular relations with China, and I proposed for « Gaindu » an
etymological origin from $1 Chiung-tu, a name which goes back to the Han and, at least since
the T'ang, has been applied to the region of the present Ning-yüan (YS, 61, 9). CHAVANNES agreed in TP, 1905, 17, and CORDIER quoted my opinion (Y, III, 83; « Kiang-tou » in BEFEO, iv, 771, is a misprint, so that the would-be alternative form « Chiang-tu » must be suppressed in Y, III, 83, and RR, 421). I now think that the problem is not so simple.
The only Chien-tu mentioned in Chinese geographical repertories is an unidentified district somewhere in south-western Yün-nan, for which in any case a localization in the Chien-ch'ang Valley is excluded. We can even go further; in the text I referred to in 1904, which narrates events
of 1284, it is said that the Burmese king had taken refuge at « the city of T'ai-kung of Chien-tu»;
and that city of T'ai-kung is certainly Tagaung on the Irawaddy (cf. BEFEO, ix, 669, and the map after p. 680) ; so « Chien-tu » in that case is not even in Yün-nan, but beyond Yün-nan in Upper Burma. But, at the same time, most of the mentions of Chien-tu in the pên-chi cannot apply to Burma or even to south-western Yün-nan. The pen-chi says ( YS, 6, 6 a) that, in 1268, Ch'ieh-mien (*Kämän, *Kämäl) was ordered to transmit the Imperial orders to Chien-tu, but it is only in 1272, according to the same pên-chi (YS, 7, 6 b), that an important military force was sent to conquer Chien-tu ; and the geographical section ( YS, 61, 9 a), when speaking of the district (lu) of Hui-ch'uan, which was the official name of the Chien-ch'ang Valley when the materials used in that section were compiled, says that it submitted to the Mongols in 1272. In the same way, the chapters on official postal transport (jamci) reproduced in the Yung-lo ta-tien speak in succession (ch. 19418, 5 a and b) of the postal station, of Ssü-ch'uan, Chien-tu and Yün-nan, and of the more rapid land transport by way of Chien-tu than by river. In 1274, 11,500 men were sent to garrison Chien-tu, and a military command (tu-hu-fu) of Ning-yüan of Chien-tu was created ( YS, 8, 3 b) ; Ning-yüan-fu is in modern times the name of the main administrative centre of the Chien-ch'ang Valley. In all such cases, Chien-tu can only apply to the Chien-ch'ang Valley. My conclusion is that, for the name of Chien-tu unknown to the geographical sections of YS, there are two different values in the pên-chi of the same YS, one being the Chien-ch'ang Valley, the other the region of Tagaung and north-east of Tagaung. Chien-tu, as applied to the Chien-ch'ang Valley, was only a popular form, probably derived from the ancient regular Chiung-tu. The pen-chi of YS, as well as the yamci sections of the Yung-lo ta-tien and some other parts of YS, are full of popular names, sometimes Chinese, sometimes Mongol or borrowed by the Mongols from non-Chinese tribes. It is only too natural that foreigners like Polo and Persian writers should have known and used only these popular forms, not sanctioned (and often not even quoted) in the later official nomenclature. Under such conditions, I do not hesitate to say that Polo's « Gaindu » transcribes the popular form phonetically and arbitrarily noted as Chien-tu in YS and in Yung-lo ta-tien. The transcription Chien-tu
IIIF Color HighRes
