^{page 920 note 1} If budām(di) were read, we should have ‘they brought’.

^{page 920 note 2} lyiba' is Tibetan lib ‘all’.

^{page 920 note 3} For tsīya.

^{page 920 note 4} An edition of the British Museum documents with signature Or has been in the hands of the printer since 1942. An edition of the Hedin documents is almost ready and will be shortly in the hands of the printer.

^{page 921 note 1} With kuvaniya- cf. Khotanese kvana- in the passage quoted Asia Major, n.s. I, 45, from p 5538 a 81.

^{page 922 note 1} Probably dvīpha was changed to dvīpīja and pha not marked for deletion.

^{page 922 note 2} Read varai as in Ch 00266.

^{page 922 note 3} This scene is illustrated on the Borobudur, see a reproduction in Asiatic Mythology by Hackin, J. and others, London, 1932, p. 237.Google Scholar

^{page 922 note 4} Below 457. 1 nadī(ḥ sa)matikramya. The plural is supported by the Tibetan chu-bo rnams.

^{page 922 note 5} MSS vidrārryam, vidrāyya. The correct vidrāvya ‘put to flight’ occurs, e.g. in Bodhīsattva avadāna-kalpalatā 13. 57.

^{page 924 note 1} Wrongly read hrrūva, as re-examination has shown, although the prr is somewhat unusual.

^{page 924 note 2} ysīmā might have connections with *zarmya- or *ysamya-. The other words remain still unexplained.

^{page 924 note 3} Documents containing these lists from the British Museum and Hedin collections are in the hands of the printer.

^{page 924 note 4} pirova- may have replaced pirovala- ‘overseer of the pirova’ as in 154 B 2, then the instr. -ena refers to the agent. I take it as instr. -ena, not -e loe. sing, followed by na ‘not’, since else-where pirovami is the locative in use.

^{page 925 note 1} pu or ke; ri, vi or di.

^{page 925 note 2} The pūhī of Ch 00266. 150 may have lost its r through scribal carelessness, of which this MS shows an unusual amount.

^{page 926 note 1} Words for ‘bridge’ in this region exist in the Khotan. hī <*haitu, and in modern times in Turkī küprük (see G. Raquette, English-Turkī Dictionary, p. 13 b).

^{page 927 note 1} He assumed that Iran, -ān was from the gen. plur. in place-names, but it should not be forgotten that Old Iranian had both a gen. plur. -ā/nām and an adj. suffix -āna-, which in Western Middle Iranian in both Persian and Parthian became -ān. In an inflected language like Khotanese the gen. plur. -ānu can be distinguished from the adj. -āna-.

^{page 927 note 2} ca written for va: the scribe of Ch 00266 is often careless.

^{page 927 note 3} An aksara is lost at the end of lines in P 2025.

^{page 929 note 1} An edition has been in the hands of the printer since 1942.

^{page 930 note 1} The controversy is registered by L. Kenou, JA 1936. 1. 42 ff.

^{page 930 note 2} Bruchstücke der Kalpanāmanditikā des Kumāralata, p. 67.

^{page 934 note 1} hambeca ‘summary’, older hambirsta, hambista, which renders Skt. samāsa-.

^{page 934 note 2} For this short text Ed. Conze has recently given a full bibliography in JRAS 1948, Text, Sources and Bibliography of the Prajñāpāramitāhrdaya.

^{page 936 note 1} Benveniste, E., Monumenta linguarum Asiae majoris, 3 (facsimile); Textes sogdiens 1940Google Scholar (transcription).

^{page 936 note 2} Probably dh, not s, was intended by the scribe.

^{page 937 note 1} The Bud. Skt. text with the parallel Pali is in A.F.R. Hoernle, Manuscript Remains of Budḏhist Literature, p. 39.