^{page 111 note 1} Avadāna-015B;ataka xxxvi; Divyāvadāna xxxviii; Avadāna-kalpalatā xcii (not 24, as quoted by Dines Andersen and Sylvain Lévi; the mistake goes back to Feer, JA, 1878, 161), Bibl. Ind. ed., II, 841; Bhadrakalpāvadāna, xxviii, Cambridge University Library, Add. MS 1411, f. 225b. The last of these is as yet unpublished, but adds nothing of significance for our present purpose. The version of the Avadāna-śataka is the oldest of this group, and is either the source (or the principal source) of the others, or is very closely related to that source. For detailed studies of the subject-matter of the various versions, see Feer, L., ‘ Maitrakanyaka-Mittavindaka, la piété filiale’, JA, 1878, 360–443;Google Scholar Speyer, J.S., ‘De koopman, die tegen zijne moeder misdreef. Een op den Boro Boedoer afgebeeld Jataka‘, Bijdragen tot de Taal-, Land-, en Volkenkunde, LIX, 1906, 181–206.CrossRefGoogle Scholar Since the story is well known, I give only the points relevant to the problems dealt with in the present article.

^{page 112 note 1} Jātaka 439, with summary versions in 82, 104, and 369, and a divergent version of parts of the story in 41.

^{page 112 note 2} JA, 1878, 370.

^{page 112 note 3} If the manuscript tradition is to be trusted, the version in the Bhadrakalpāvadāna gives the name a feminine form in some places; but this would obviously be unworkable in everyday life.

^{page 112 note 4} JA, 1878, 441. This envisages vinda as equivalent to Sanskrit vrnda. The PTS dictionary, however, does not know the word.

^{page 112 note 5} Tassa nibbattito patthāya tarn kulam paramaduggatam eva jātam. nābhito uddham -udakakanjikamattam pi na labhi. tassa pana Mittavindako ti nāmam ahosi (Jātaka i. 238).

^{page 113 note 1} XXXII, pp. 50–7. In the same publication, in addition to the relevant Tibetan, Pali, and Chinese texts (of which, however, only the Tibetan contains the Maitrāyajna story) Lévi was able to include fragments of a Kuchean version in which the name appears as Maitrajña-, ibid., 245.

^{page 113 note 2} ibid., 51 n.

^{page 113 note 3} See in particular iv.1.76 ff.; iv.1.162 ff.; and my article ‘The early history of the gotras’, JRAS,1946,41 ff.

^{page 115 note 1} Such a tendency would naturally be encouraged by the fact that in the Middle Indian period a number of gotra-namea came to coincide in form with their eponyms. Thus for example the form Gotama had to do duty for both the Sanskrit forms Gotama and Gautama.

^{page 115 note 2} Even Maitri may have occurred: cf. below, p. 117. The remaining theoretical possibility, Maitreya, would naturally have been avoided for doctrinal reasons

^{page 116 note 1} For example vii.94.9 gómad dhíranyavad vásu yád vām áśvāvad mahe.

^{page 116 note 2} In the Brahmanical texts the name Mitravinda occurs in the Vamśa-brāhmana of the of the Sāmaveda, Weber, , Indische Studien, IV, 1858, 372;Google Scholar Harivamśa 9186 (a son of Krsna); and cf. also Mahābhārata III.210.19. We may remark in passing that Krsna's epithet, Govinda, is a normal formation, and the often quoted suggestion that the word is a Prakritic form for gopendra is quite superfluou.

^{page 117 note 1} The ‘Pali Mettakaññaka’ which appears in Chizen Akanuma, Indo Bukkyō koyū meishi jiten, s.v. Maitrakanyaka, is a pure invention.

^{page 117 note 2} Quoted by E. Chavannes, Cinq cents contes et apologues extraits du Tripitaka chinois, I, 132, from the bas-reliefs of Pagan.

^{page 117 note 3} Taishō Tripitaka, Vol. 3, No. 152, p. 21; Vol. 53, No. 2121, p. 223; Vol. 4, No. 203, pp. 450–1; Vol. 3, No. 190, pp. 884–7. Of these, the first and third have been translated by Chavannes, Cinq cents contes, I, 132, II, 10, and the fourth by Beal, The romantic legend of Sâkya Buddha, 342 (also in IA, Ix, 1880, 224–6). For these references, and for assistance with the Chinese material in general, I am indebted to Professor W. Simon.

^{page 118 note 1} Once he had become the Bodhisattva, there was naturally a temptation to minimize the sin. In the version in Chavannes, III, 10, when the mother clasps her son's feet, he tries to disengage her hands, and accidentally breaks a few hairs; and in the Mi-lien version (Lévi, Mahākarmavibhanga, 51 n.) he accidentally strikes her on the head with his foot while emerging from his bath. In the version in Chavannes, I, 132, the incident of the kick has disappeared altogether, and the wheel is the punishment only for his dalliance with the divine damsels.

^{page 118 note 2} Bijdragen tot de Taal-, Land-, en Volkenkunde, LIX, 1906,Google Scholar 199 ff.

^{page 119 note 1} Cinq cents contes, I, 132: ‘On peut done se demander si Mi-lan n'etait pas à l'origine Mithra, e génie de la lumière céleste’. It is curious that he should have quoted the Iranian form of the name rather than the Indian.

^{page 119 note 2} Nirnaya Sagar edition, 1951, 270.

^{page 120 note 1} xxvi-xxix; also the fragments of the Kalpanāmanditikā of Kumāralāta, ed. H. Lüders, Kleinere Sanskrit-Texte, II.

^{page 120 note 2} Ed. J. Nobel, chap, xviii, pp. 202–25.

^{page 120 note 3} Cowell, E.B. and Neil, R.A. (ed.), The Divyâvadâna, Cambridge, 1886, 586–609.Google Scholar

^{page 120 note 4} pp. 204–6. The two other principal contributions to the textual criticism of the Divyāvadāna, Speyer, J.S., ‘Critical remarks on the text of the Divyāvadāna’, Vienna Oriental Journal, XVI, 1902, 103–30,Google Scholar 340–61, and Bailey, D.R. Shackleton, ‘Notes on the Divyāvadāna’, JRA8, 1950, 166–84Google Scholar, 1951, 82–102, do not deal with this chapter.

^{page 121 note 1} Kālidāsa has the elegant simile, KumS i.46 pravātanīlotpalanirciśesam adhīravipreksitam āyatāksyāh}, but this is very different from the verse under consideration.

^{page 122 note 1} Strictly, in classical Sanskrit, prariutta; but a spelling with the dental n would not be unexpected in a Buddhist manuscript.

^{page 122 note 2} The interchange of I and r being common in Nepalese manuscripts, we could readily assume jvarana as a probable intermediate stage in the corruption.

^{page 122 note 3} This implies only a small dittography, e.g. vaivabhava-, and a subsequent minor scribal emendation.

^{page 123 note 1} Perhaps a pun on vamśa is also intended, in the sense of the stem of the lotus.

^{page 123 note 2} Edgerton, Buddhist Hybrid Sanskrit dictionary, notes māla- only as the prior member of a compound.

^{page 124 note 1} Sep below, p. 129.

^{page 124 note 2} Cambridge University Library, Add. MS 1411, f. 228a.

^{page 125 note 1} Narthang Kanjur, Mdo, vol. ha, f. 146a.

^{page 125 note 2} JA, 1878, 370.

^{page 125 note 3} Translation of Avadāna-śataka, in Annales du Musée Guimet, XVIII, 1891, 133.Google Scholar

^{page 125 note 4} JRAS, 1955, 16.

^{page 126 note 1} Professor Simon has drawn my attention to Nobel, J., Udrāyana, König von Roruka, eine buddhistische Erzählung, Wiesbaden, 1955, II, 71,Google Scholar Glossary, s.v. yul: ‘~-gyi mi (-dag, -mams), janapadah (insbesondere die Leute in den Provinzen im Gegensatz zu denen in der Residenz oder Stadt)’. I think that this goes rather too far. In all the passages cited by Nobel at this point, the sense is certainly not ‘the people of the countryside’, but rather ‘the inhabitants of the kingdom’. In the compound also noted by Nobel, paura-janapadāh (classical -jāna-padāh): pho-bran-bkhor-gyi mi dan yul-gyi mi, the opposition of town and country is admittedly present, at least by implication; but it seems very doubtful whether yul in isolation would really convey the implication of ‘rural’.

^{page 126 note 2} Ed. mtsho.

^{page 127 note 1} Narthang Kanjur, Hdul-ba, vol. ja, f. 121a.

^{page 127 note 2} The compound is missing in Jäschke and Desgodins s.v. bzah, but given s.v. bcah.

^{page 127 note 3} Taishō Tripitaka, Vol. 23, p. 800b.

^{page 127 note 4} Divyāvadāna XXXV, p. 500.

^{page 127 note 5} So read by Böhtlingk, Sanskrit-Wörterbuch in kürzerer Fassunq, s.v. utkarikā, probably correctly. It is noted there, however, that it is printed as -siddhas tatkariketi, and this is in fact the reading in all the editions I have been able to consult. From the Dravidian languages, J. R. Marr has given me Tamil ukkāri, Telugu ukkěra, ‘a variety of sweet cake’.

^{page 128 note 1} I am indebted to Sir Ralph Turner for this suggestion, and also for the modern forms in the previous paragraph.

^{page 128 note 2} I. Gershevitch, Grammar of Manicheean Sogdian, §§ 63, 392, 969.

^{page 130 note 1} Dvāvimśaty-avadāna iv.

^{page 131 note 1} It is indeed possible that kutha is a ghost-word, arising merely from a misoopying of kuśa.

^{page 131 note 2} Ed. Nobel, p. 202.

^{page 131 note 3} See Edgerton, Dictionary, s.v., and Lévi, S.,‘Māla vihāra’, BSOS, VIII, 2–3, 1936,Google Scholar 619–22.