LOAN-WORDS IN TIBETAN 595 



historical investigations are necessary in order to establish such an iden- 

 tity ; a merely apparent coincidence of words proves little or nothing. 



170. The Turkish origin of tupak is also maintained by W. GEIGER 

 (Lautlehre des BaluSi, p. 66) : Baluci tupak, tupan, tufan, topak; Yidga 

 tufuk. 



171. The word cdku occurs also in Kurd caku, caxo, etc. (J. DE 

 MORGAN, Mission en Perse, Vol. V, p. 140). 



183. The word se-mo-do occurs in the Tibetan translation of the 

 Amarakosa (p. 166). 



198. pir-t'i (" quick-match") is also connected with Turk! pilta 

 (LE COQ, p. 86 b). 



207. Another Sanskrit term for Panicum miliaceum is cmaka 

 ("Chinese") and cinna. 



279. k'ra-rtse, pronounced t*ar-tse, is perhaps merely a bad spelling 

 of Persian tardzu (No. 128). 



299. t'ai rje is possibly connected with Mongol taiji (cf. O. FRANKS, 

 Jehol, p. 30). 



On p. 421 it is stated that the animal kun-ta is not yet traced to its 

 Sanskrit original. Boehtlingk's Dictionary, however, has Sanskrit 

 kunta with the meaning "a small animal, a worm"; but this entry 

 may be simply based on the Tibetan mDzans-blun. The Chinese tran- 

 scription calls for a prototype *kunda. 



To the Persian loan-words add $o-ra (above, p. 503). 



To the Arabic loan-words add $eg (" chief tain, elder"), from Arabic 

 Saix. 



To the Turkl loan-words add gan-zag (above, p. 577). 



Sir GEORGE A. GRIERSON, editor of the "Linguistic Survey of India/' 

 has done me the honor to look over my Loan- Words in Tibetan, and to 

 favor me with the following observations, which are herewith published 

 with his kind permission: 



The Kashmiri for "egg" (p. 405) is t'ul. 



15. I cannot think that *andanil is a possible Apabhramga (using 

 the word in its technical sense) word. The presence of n seems to 

 point to Kashmiri, in which ni has a tendency to change to ni. The 

 Ksh. equivalent of Skr. mla- is nilu, pronounced nyul, and it is a com- 

 mon-place that ny and n in that language have the same sound. In fact, 

 original medial ny is written n (e.g. dana, from Skr. dhdnya-, "paddy"), 

 in this following Paigaci Prakrit. 



17. 'Arya-pa-lo. This is typical Pigaca, which changes ry to 

 r(i}y and v(b) to p. In all Indian Prakrits, dry a would become ajja- 9 

 with short initial a. 



