506 SlNO-lRANICA 



be a product of K'ari (Sogdiana) and Kuca. 1 The fact that this tran- 

 scription is identical with fiS we recognize from the parallel passage in 

 the Pel Si, 2 where it is thus written. The text of the Sui Annals with 

 reference to Iranian regions offers several such unusual modes of 

 writing, where the Pei Si has the simple types subsequently adopted as 

 the standard. The variation of the Sui Annals, at all events, demon- 

 strates that the question is of reproducing a foreign word; and, since 

 it hails from Sogdiana, there can be no doubt that it was a word of the 

 Sogdian language of the type *navsa or *naf sa (cf . Sanskrit navasara, 

 Armenian navt*, Greek va(j>6a); Persian naSddir, nuSddir, nauSadir, 

 nauSddur, noSddur, being a later development. It resulted also 

 in Russian nuSatyr. In my opinion, the Sogdian word is related 

 to Persia neft ("naphta"), which may belong to Avestan napta 



("moist"). 3 



Tribute-gifts of nao-Sa are not infrequently mentioned in the Chinese 

 Annals. In A.D. 932, Wan Zen-mei 3: iH H, Khan of the Uigur, pre- 

 sented to the Court among other objects ta-p'en Sa (" borax") 4 and sal 

 ammoniac (kan So). 6 In A.D. 938 Li Sen-wen $ H 3C, king of Khotan, 

 offered nao-Sa and ta-p'en Sa ("borax") to the Court; and in A.D. 959 

 jade and nao-Sa were sent by the Uigur. 6 The latter event is recorded 

 also in the Kin Wu Tai Si, 7 where the word is written ffl #J% pho- 

 netically kan-Sa, but apparently intended only as a graphic variant 

 for nao-Sa* The same work ascribes sal ammoniac (written in the same 

 manner) to the T'u-fan (Tibetans) and the Tafi-hian (a Tibetan tribe 

 in the Kukunor region). 9 In the T'ang period the substance was well 



1 According to Masudi (BARBIER DE MEYNARD, Les Prairies d'or, Vol. I, p. 347), 

 sal-ammoniac mines were situated in Soghd, and were passed by the Moham- 

 medan merchants travelling from Khorasan into China. KuSa still yields sal am- 

 moniac (A. N. KUROPATKIN, Kashgaria, pp. 27, 35, 76). This fact is also noted in 

 the Hui k'ian ci (Ch. 2), written about 1772 by two Majichu officials, Fusamb6 

 and Surde, who locate the mine 45 li west of Kuca in the Sartatsi Mountains, and 

 mention a red and white variety of sal ammoniac. Cf. also M. REINAUD, Relation 

 des voyages faits par les Arabes et les Persans dans 1'Inde et a la Chine, Vol. I, 



p. CLXIII. 



2 Ch. 97, p. 12. 



3 Cf. P. HORN, Neupersische Etymologic, No. 1035; H. HUBSCHMANN, Persische 

 Studien, p. 101, and Armen. Gram., p. 100. 



4 As I have shown on a former occasion (T'oung Pao, 1914, p. 88), Chinese 

 p'en (*bun) is a transcription of Tibetan bul. 



5 Ts'efu yuan kwei, Ch. 972, p. 19. 



6 Wu Tai hui yao, Chs. 28, p. 10 b; and Ch. 29, p. 13 b (ed. of Wu yin lien). 



7 Ch. 138, p. 3. 



8 The character kan is not listed in K'an-hi's Dictionary. 



9 Ch. 138, pp. i b, 3 a. 



