Roth has beautifully illustrated the matter by shuffling together verses from Schiller's Riddles ("Von Perlen baut sich eine Brücke" and "Es steht ein gross' geräumig Haus"): see ZDMG. xxxvii. 109.—Cf. notes to xviii. 2. 35; 3. 47.⌋
3. Unto Yama the king offer ye an oblation, milk rich in ghee; he furnishes (ā-yam) to us among the living a long life-time, for living on.
The verse is RV. x. 14. 14 (found also in TA. vi. 5. 1), except that b is 15 b: ⌊see note to the preceding verse⌋. ⌊TA. agrees with RV.⌋ For páyas in a, RV. reads havís; for havís in b, havyám; and in c, d, it has devéṣv ā́ yamad dīr-. SPP. reads in c-d yamed dīr- ⌊so also Caland, Todtengebräuche, note 243⌋, which is certainly better than yame dīr-: half his authorities give the former; but of our mss. only Op. has yamet, while D. has yamat, like RV. ⌊TA. and the comm.⌋. ⌊The case is strikingly like that of mānaye tathā, mānayet tathā, at xv. 10. 2.⌋—The Prāt. (ii. 76) distinctly requires ā́yuṣ prá to be read in d, but of our mss. only O.s.m. (in margin) gives it; nearly half of SPP's saṁhitā-mss., however, have it, and it ought to be received as the true AV. text, though both editions read ā́yuḥ. These three verses make no appearance in Kāuç. ⌊This last statement now appears to be wrong: see note to vs. 1.⌋
⌊Considering the exaggerated nicety of the theory of the Hindus respecting consonant groups (cf. Whitney, AV. Prāt., p. 584-90), and in particular their doctrine of the varṇakrama ("At the end of a word, a consonant is pronounced double," padānte vyañjanaṁ dviḥ, Prāt. iii. 26), it is strange that the mss. sometimes fail to come up even to the simple requirements of orthography as set by grammar and sense. On the other hand, it can hardly be said that the mss. in the cases of these shortcomings are a less truthful representation of the actual connected utterance of the text than would be for instance the graphical representation of the English some more by the words some ore.—I have thought it worth while to assemble a few notable cases where the one of two needed double letters is omitted. Thus besides yame[d] dīrghám and mānaye[t] táthā, just mentioned, we have: at xviii. 3. 3, jīvā́m [m]ṛtébhyas (a most striking example: there is abundant ms. authority for the false ṛtébhyas, which is yet shown beyond all peradventure to be a blunder by the mṛtā́ya jīvā́m of TA.); at xviii. 4. 40, juṣantām | ā́sīnām ū́rjam úpa yé sácante, shown to be a gross corruption by HGS., which reads juṣantām | māsī ’mām ūrjam etc.; at xix. 31. 2, sá[m] mā sṛjatu puṣṭyā́, where Ppp. has in fact sam mā, and where sám is supported by the parallelism of our vi. 5. 2; and, at xix. 7. 3, áriṣṭa[m] mū́lam, where all authorities agree in omitting the -m.—On the other hand, a superfluous double is sometimes written. Thus we find: at xix. 42. 3, sumatím [m]āvṛṇānáḥ, where the pada-text reveals its modernity and lack of insight by reading unaccented mā and vṛṇānáḥ instead of ā॰vṛṇānáḥ; and at xix. 58. 4, púraḥ kṛṇudhvam [m]ā́yasīḥ. At xix. 46. 6, the ghṛtā́d dúrluptas or úrluptas of some authorities, instead of -ā́d úllup-, is a blunder of similar origin. Cf. sā́danā[t] te, xviii. 3. 52, note. The well-known sám [m]ahema of RV. i. 94. 1, as taken by BR. at vii. 1609, i. 567, would belong in this category; but Grassmann manages to refer it to root mah.⌋
4. Do not, O Agni, burn him up; do not be hot upon (abhi-çuc) him; do not warp (kṣip) his skin, nor his body; when thou shalt make him done, O Jātavedas, then send him forward unto the Fathers.
