^{page 26 note 1} In the Glossar to E the vara of this passage is explained as a form of kara- “circuit”.

^{page 26 note 2} Ed. C. Régamey, §§ 11, 12, 13, 16.

^{page 26 note 3} I take pajsadā, to be from *pajsandaa-, beside the noun pajsama- “honouring”. The word ṣakäle “dry, dried” can be seen in P 2782.20 vira ṣakala brrimji khu mī ūtci udiśä and ibid. 23 tti khu ṣakala brrirnjvam kara utci ni byide. Here there is reference to brriṃja- “pools (?)”, of which the water has been dried up. In E 24.44the henei candanä skälä is dried red sandalwood, to be used for an image. The word may be a derivative from hiṥku- “dry” by a suffix -ila-, with reduction and loss of the hi- as in ṣṭa- ‘stand’. The same suffix occurs in P 2891.30 utcäla kauysä “a water jar” from ūtca “water”; kauysa- occurs beside kūysa-, NPers. kūzah. In gubā' I see a form of gumbaka-, Bud.Sansk. gulmaka-, on which see BSOAS. 13.923, and add gumika- “captain” from the inscription in H. Lüders' List, no. 1200.

^{page 27 note 1} Quoted BSOAS. 10.922. Khot. Texts ii, 104.

^{page 27 note 2} BSOAS. 13.921–2.

^{page 27 note 3} The variants are varai and varāe, see the texts now in Khot. Bud. Texts 21, 29, 37.

^{page 27 note 4} See BSOS. 7.765; Dātastān ī dēnīk 36.81 quoted in Zoroast. Problems 222.

^{page 27 note 5} BSOS. 7.295–6 (vairi loc. sing, should be kept). J. Duchesne-Guillemin, Jour. As. 1936.1.246.

^{page 27 note 6} A recognized usage, see for example, K. Brugmann, Grundriss ii 2. 562. There is no need to replace anrāi by anrahe.

^{page 27 note 7} I have become doubtful of the reading of this epithet. In Zoroast. Problems 21 I still, without full conviction, held to ganāk as a verbal derivative with the suffix āk of the agent. It could have been from a denominative gan- “to strike”. Support could be offered to that from Ossetic ϒänä, Iron qän wound, fracture, fractured” from *gana-, see G. Morgenstierne, NTS. 12.268. It has recently occurred to me that if spēnāk were derived from the comparative spanyah- in the nom. sing. spanyā^h > spēnā- to which was suffixed the -k after long vowel in final position (on which -k see Zoroast. Problems 183), then in a similar comparative could be concealed, from the adj. anra-. That would be of the form *anhyā^h from which *anyā and *ēnā- could have arisen. With the same -k the word would be *ēnāk. This could be read in the Pahlavi spelling if it is recognized that, though rarely, initial alif is sometimes absent before y-. Thus frequently in ēstēt written ystyt, and in ywk'n, discussed in BSOS. 7.761 and in which I now see ēvakān “one by one ” on its way to become *yukān before the stage yagānah in NPers.

^{page 28 note 1} Asia Major, n.s., 2.32, BSOAS. 14.421 ff.

^{page 28 note 2} Trans. Philol. Soc. 1945, 28. For “hole ” in Sanskrit vi-vara- was used.

^{page 28 note 3} G. Morgenstierne, IIFL. 2.549; Bloch, J., La langue marathe, 303. For words outside Indo-Iranian from uer- see Walde-Pokorny i 280Google Scholar.

^{page 28 note 4} P 3513, 43 v 3 in Khot. Texts i 222.

^{page 28 note 5} A similar etymology of the name Manjuśrī follows immediately in the same text; quoted BSOAS. 10. 910.

^{page 29 note 1} Asia Major, n.s., ii 32 following the explanation of T. Burrow in BSOS. 7.787.

^{page 29 note 2} Pers. Studien p. 33.

^{page 29 note 3} Khot. Bud. Texts 29 and 38. See BSOAS.14.422.

^{page 29 note 4} Ed. Leumann, E., Maitreya-samiti 186Google Scholar. For the Chinese parallel, see ibid. 259.

^{page 29 note 5} BSOAS. 10.910.

^{page 29 note 6} Khot. Bud. Texts 48.

^{page 30 note 1} The explanation by the base varz- “work ”, offered earlier by Sten Konow, Saka Studies 196, and by myself in BSOS. 9.77, is excluded by the Iranian replacement of -ĝt- by -št- (Av. varšta-), not by -žd-. In Khotanese we find even rrīs- in P 2022.18 rrīśtä “he licks ” and rräṣta- “licked” from - ĝht-, as NPers. lišt. For a trace of -ghs- becoming -gž- see Trans. Philol. soc. 1952, 57. [Not var-, JRAS. 1942, 26.]

^{page 30 note 2} G. Morgenstierne, Metathesis of Liquids in Dardic, Festslcrift til Prof. O. Brock, 150; Turner, R. L., Position of fiomani p. 9Google Scholar.

^{page 30 note 3} Other forms in R. L. Turner, Nepali Dictionary, s.v. buro “old”. There is no reason to assume (as in the Nepali Dictionary and M. Mayrhofer, Handbuch des Pali i 48) influence from the base vardh-. From the base brah- “remove” (Old Ind. vrah- and brah-) similarly the long syllable of the partc. brḍha- is shown by Pali which has in compounds both -būha- and -bālha-, see abbūḽha- and pabāḽha-.

^{page 30 note 4} In Paṣto ozai, Waziri wözai “bone of arm ” has also lost the Old Iran. b- before ö. It is traced by G. Morgenstierne to bāzu-ka- NTS. 12.262. Also ūžd “long ”.

^{page 30 note 5} Mahāvyutpatti 3100.

^{page 30 note 6} Khot. Texts i p. 142, 51 r 5. To be added to the examples in BSOAS. 14.422.

^{page 30 note 7} See Asia Major, n.s., ii 32.

^{page 31 note 1} N. 163.25; 164.11, in Saha Studies.

^{page 31 note 2} ZDMG. 91.108.

^{page 31 note 3} V. I. Abaev, Oset. Jazyh i Fol'klor i 314; Jazyki Severnogo Kavkaza 2.16; 6–7; Pamiati Marra 245, 252.

^{page 31 note 4} Abaev, loc. cit. 312 ff.; Pamiati Marra 245; N. Marr, O jazyke Abxazov, 413.

^{page 32 note 1} The similar Parth. urdywn is explained as a compound by E. Benveniste, Jour. As. 1936.1.201.

^{page 32 note 2} Siddhasāra 20 v 5.

^{page 32 note 3} Jātaka-stava 9 v 2 ‘turning’.

^{page 32 note 4} Suvarṇabhāsa 69 v 1 in Khot. Texts i 246, translating Sansk. arṇava-.

^{page 32 note 5} E 18 13 and Jātaka-stava 15 r 2.

^{page 32 note 6} Khot. Texts ii 5.70.

^{page 32 note 7} These words have been compared by V. I. Abaev, Oset. Jazyh i 314.

^{page 32 note 8} Dumezil, G., Etudes comparatives sur les langues caucasiennes du nord-ouest, p. 75, from Sbornik Mater. Kavhaza 14, ii, p. 3.3Google Scholar.

^{page 32 note 9} Pamiatniki 2.170. For Ossetic mythological names, see V. I. Abaev, “O sobstvennyx imenax nartovskogo ëposa” in Jazyk i Myṣlenie 5. Earlier Trubetzkoy, N., Mém. Soc. Ling. 22.247 ffGoogle Scholar.; Dumezil, G., Loki 169 ffGoogle Scholar

^{page 32 note 10} dŭar “gate”, Miller, Vs., Digorskie Skazanie 38Google Scholar; Dumezil, G., Les Légendes surles Nartes, p. 113Google Scholar.

^{page 33 note 1} One may note that in Digor there are the two forms of irād and ärŭäd. Iron iräd “bride-price”, and Digor är&ŭágŭäs and ärágäs “convincing”. Whether the ŭ-here were original or not, an analogical form could have been developed.

^{page 33 note 2} Dumezil, G., Mélanges Grégoire, 223–6, and JRAS. 1953, p. 112Google Scholar.

^{page 33 note 3} Artāy Vīrāz nāmak18.5. On Av. ərəϒant-, Pahl. argand, see Trans, Philol. Soc. 1945, 4.

^{page 33 note 4} Sbornik Mater. Kavkaza 17, ii 157; Bleichsteiner, R., Rossweihe 467Google Scholar.

^{page 33 note 5} See Trans. Philol. Soc. 1945, 28.

^{page 34 note 1} See Dumézil, G., Études comparatives, pp. 124–5Google Scholar.