The Pomegranate 285 



Iranian colonies outside of Iran proper, those of Sogdiana and Turkis- 

 tan, played a prominent part. We know the Sogdian word for the 

 pomegranate, which is written n'r'kh, and the reading of which has 

 been reconstructed by R. Gauthiot 1 in the form *narak(a), developed 

 from *anar-aka. This we meet again in Persian andr, which was adopted 

 in the same form by the Mongols, while the Uigur had it as nara. At 

 all events, however, it becomes necessary to restore, on the basis of the 

 Chinese transcription, an ancient *riu, *ru, of some Iranian dialect. 

 This lost Iranian word, in my opinion, presents also the foundation of 

 Greek ?>6a or /Joid, — the origin of which has been hitherto unexplained or 

 incorrectly explained, 2 — and the Semitic names, Hebrew ritnmon, 

 Arabic rumman, Amharic riiman, Syriac rumono, Aramaic rummana, 

 from which Egyptian arhmani or anhmdni (Coptic ertnan or hertnan) / 



is derived. 3 



(4) ^B %H lo-liu, *zak (yak, n'iak)-liu (riu). This hybrid compound, 

 formed of elements contained in 2 and 3, is found in the dictionary 

 Kwan ya M 5ft, written by Can Yi 181 M about a.d. 265. 4 It is also 

 employed by the poet P'an Yo of the fourth century, mentioned above. 5 

 Eventually also this transcription might ultimately be traced to an 

 Iranian prototype. Japanese zakuro is based on this Chinese form. 6 



While the direct historical evidence is lacking, the Chinese names of 

 the tree point clearly to Iranian languages. Moreover, the tree itself, 

 is looked upon by the Chinese as a foreign product, and its first intro-j 

 duction into China appears to have taken place in the latter part of; 

 the third century a.d. 



In my opinion, the pomegranate-tree was transplanted to India, 



1 Essai sur le vocalisme du sogdien, p. 49. Cf. also Armenian nrneni for the 

 tree and nurn for the fruit. 



2 The etymologies of the Greek word enumerated by Schrader (in Hehn, 

 Kulturpflanzen, p. 247) are so inane and far-fetched that they do not merit dis- 

 cussion. It is not necessary, of course, to hold that an immediate transmission of 

 the Persian word took place, but we must look to a gradual propagation and to 

 missing links by way of Asia Minor. According to W. Muss-Arnolt (Transactions 

 Am. Phil. Assoc, Vol. XXIII, 1892, p. no), the Cyprian form fivdia forbids all 

 connection with the Hebrew. It is not proved, however, that this dialectic word 

 has any connection with /J6a; it may very well be an independent local development. 



3 V. Loret, Flore pharaonique, p. 76. Portuguese roma, romeira, from the 

 Arabic; Anglo-Saxon read-appel. 



4 This is the date given by Waiters (Essays on the Chinese Language, p. 38). 

 Bretschneider (Bot. Sin., pt. I, p. 164) fixes the date at about 227-240. 



B T'an lei han, Ch. 183, p. 9. 



6 Written also ffi ;R§. E. Kaempfer (Amoenitates exoticae, p. 800) already 

 mentions this term as dsjakurjo, vulgo sakuro, with the remark, "Rara est hoc 

 coelo et fructu ingrato." 



