^{page 1192 note 1} Cf. Z.D.M.G., lxii, 113–18.Google Scholar

^{page 1192 note 2} i, 3.

^{page 1192 note 3} lxxii.

^{page 1192 note 4} Z.D.M.G., lxii, 119.Google Scholar

^{page 1192 note 5} Z.D.M.G., lxii, 358–60.Google Scholar

^{page 1192 note 6} Cf. also Z.D.M.G., lxii, 362.Google Scholar

^{page 1192 note 7} In any case, it is obvious that no stress can be laid on a distinction of e and ai in Sanskrit MSS.

^{page 1193 note 1} For Ṛgvedic examples of the use, wrongly called ‘late’ by Scheftelowitz (Die Apokryphen des Ṛgveda, p. 79),Google Scholar cf. Oldenberg, , Z.D.M.G., lxi, 823.Google Scholar The use is very common in the Śatapatha Brāhmaṇa, and the Aitareya Āraṇyaka, iii, 2, 6,Google Scholar has vāgbrāhmaṇam ivopodāharati, while the parallel Sāṇkhāyana Āraṇyaka, viii, 10, has brāhmaṇam evodāharati. Both the Pet. Lexx. and Monier-Williams recognise this use.Google Scholar

^{page 1193 note 2} See Wackernagel, , Altind. Gramm., i, 320.Google Scholar