INTRODUCTION 199 



In the following example there is no positive evidence as to the 

 significance of hu. Hu wan Si ce W 3i &> ^ (" envoy of the king of the 

 Hu") is a synonyme of tu hwo M ? (Peucedanum decursivum) . l As 

 the same plant is also styled k'ian ts'in $> W, k'ian hwo, and hu k'ian 

 $i & H ^6 ffi 31 , the term K'ian (*Gian) alluding to Tibetan tribes, it 

 may be inferred that the king of the Hu likewise hints at Tibetans. 

 In general, however, the term Hu does not include Tibetans, and the 

 present case is not conclusive in showing that it does. In the chapter 

 on the walnut it will be seen that there are two introduced varieties, 

 an Iranian (hu t'ao) and a Tibetan one (k'ian t'ao). 



In hu ts'ai (Brassica rapa) the element hu, according to Chinese 

 tradition, relates to Mongolia, while it is very likely that the vegetable 

 itself was merely introduced there from Iran. 2 



In other instances, plants have some relation to the Hu; but what 

 this relation is, or what group of tribes should be understood by Hu, 

 is not revealed. 



There is a plant, termed hu hwan lien S8 3t 31, the hwan-lien (Coptis 

 teeta) of the Hu, because, as Li Si-Sen says, its physical characteristics, 

 taste, virtue, and employment are similar to those of hwan-lien. It 

 has been identified with Barkhausia re pens. As evidenced by the 



acter fan, on the authority of K'an-hi, could never have had the pronunciation po 

 nor a final consonant, and that, accordingly, in the tribal name T'u-fan (Tibet) the 

 character fan, as had previously been assumed, could not transcribe the Tibetan 

 word bod. True it is that under the character in question K'an-hi has nothing to 

 say about po, but ^ is merely a graphic variant of ^, with which it is phonetically 

 identical. Now under this character, K'an-hi indicates plainly that, according to the 

 Tsi yun and Cen yun, fan in geographical names is to be read p'o (anciently *bwa) 

 | (fan-ts'ie Jjjf $fe), and that, according to the dictionary Si wen, the same char- 

 acter was pronounced p'o (*bwa) ij&, p'u Jf , an d p'an^(cf. also SCHLEGEL, Secret of 

 the Chinese Method, pp. 21-22). In the ancient transcription | or^ JE fan-ton, 

 *par-tav, reproduction of Old Persian Par0ava (see above, p. 1 87) Jan corresponds very 

 well to par or bar; and if it could interchange with the phonetic ^ pa, *bwat, bwar, it is 

 perfectly clear that, contrary to Pelliot's theory, there were at least dialectic cases, 

 where ^ was possessed of a final consonant, being sounded bwat or bwar. Con- 

 sequently it could have very well served for the reproduction of Tibetan bod. From 

 another phonetic viewpoint the above case is of interest: we have *bak-xa and 

 *bwat-xa as ancient names for the mint, which goes to show that the final con- 

 sonants of the first element were vacillating or varied in different dialects (cf . T'oung 

 Pao, 1916, pp. 110-114). 



1 T'un ci (above, p. 196), Ch. 75, p. 12 b. 



2 See below, p. 381. In the term hu yen ("swallow of theHu"), hu appears to 

 refer to Mongolia, as shown by the Manchu translation monggo cibin and the Turkl 

 equivalent qalmaq qarlogac (Mongol xatun xariyatsai, Tibetan gyi-gyi k'ug-rta; cf. 

 Ross, Polyglot List of Birds, No. 267). The bird occurs not only in Mongolia, but 

 also in Ce-kian Province, China (see Kwei ki sanfu lu ^ H H SK ft, Ch. 2, p. 8; 

 ed. of Si yin huan ts'un $u). 



