384 Sino-Iranica 



carui), however, is commonly termed in Persian Sah-zire ("cummin of 

 the Shah") or zlre-i ruml ("Byzantine or Turkish cummin"). 1 



While the philological evidence would speak in favor of a trans- 

 mission of cummin from Persia to China, this point is not clearly brought 

 out by our records. C'en Ts'an-k'i, who wrote in the first half of the 

 eighth century, states that H-lo grows in Fu-si 1$ 1? (Bhoja, Sumatra). 

 Li Sun, in his Hai yao pen ts'ao, says after the Kwan lou ki Ht m nS 

 that the plant grows in the country Po-se; 2 and Su Sufi of the Sung 

 notes that in his time it occurred in Lin-nan (Kwan-tuh) and adjoining 

 regions. Now, the Kwan lou ki is said to have been written under the 

 Tsin dynasty (a.d. 265-420) f and, as will be shown below in detail, the 

 Po-se of Li Sun almost invariably denotes, not Persia, but the Malayan 

 Po-se. Again, it is Li Sun who does not avail himself of the Iranian form 

 s'i-lo=z'tra, but of the Sanskrit form jlraka, possibly conveyed through 

 the medium of the Malayan Po-se. 



Li Si-cen has entered under U-lo another foreign word in the form 

 3& H: Wti ts'e-mou-lo (*dzi-mu-lak), which he derived from the K'ai 

 pao pen ts'ao, and which, in the same manner as U-lo, he stamps as a 

 foreign word. This transcription has hitherto defied identification, 4 

 because it is incorrectly recorded. It is met with correctly in the Cen 

 lei pen ts'ao 5 in the form ^ W) ts'e-lo, *d2i-lak(rak), and this answers 

 to Sanskrit jlraka. This form is handed down in the Hai yao pen ts'ao, 

 written by Li Sun in the eighth century. Thus we have, on the one 

 hand a Sanskrit form jlraka, conveyed by the Malayan Po-se to Kwan- 

 tun in the T'ang period, and on the other hand the Iranian type H- 

 lo=lira, which for phonetic reasons must likewise go back to the era 

 of the T'ang, and which we should suppose had migrated overland to 

 China. The latter point, for the time being, remains an hypothesis, 

 which will perhaps be elucidated by the documents of Turkistan. 



1 Corresponding to Arabic karawya, the source of our word caraway. 



2 The Cen lei pen ts'ao (Ch. 13, p. 27 b) repeats this without citing a source. 

 8 Cf . below, p. 475. 



4 Stuart, Chinese Materia Medica, p. 176. 

 6 Ch. 13, p. 17 b. 



