^{page 363 note 1} Adhyāyas i and ii in DrFriedländer's, edition (Berlin, 1900); iii–viGoogle Scholar in Cowell's, edition (Calcutta, 1901); and vii–xv in an Appendix to my edition of the Ailareya Āra1E47;yaka (in the press).Google Scholar

^{page 363 note 2} The original sources of information are the preface to Cowell's, ed., pp. iv–viiGoogle Scholar; Weber, , Indian Literature, pp. 50, 132Google Scholar; Berlin Catalogue, i, p. 19; ii, pp. 5, 6Google Scholar; Winternitz, & Keith, , Bodleian Catalogue, pp. 59, 60.Google Scholar

^{page 363 note 3} MS. Sansk. e. 2. I have also been enabled by the help of Geheimrath Professor Dr. Pischel to make use of the Berlin MS. Orient, fol. 630 (from Bühler's collection), for the loan of which I am much indebted to the Royal Library.

^{page 363 note 4} ii, 17; iv, 1; 7 (= KauṣsĪtaki Upaniṣad, ii, 1; 7); I cite the Adhyāyas of the Upaniṣad throughout as iii–vi.

^{page 363 note 5} Kauṣītaki Brāhmana, p. ix.Google Scholar Cf. Weber, , Indische Studien, i, p. 393.Google Scholar

^{page 363 note 6} Berlin Catalogue, ii, p. 5;Google ScholarBodleian Catalogue, p. 60.Google Scholar

^{page 363 note 7} Kauṣītaki Brāhmaṇa Upaniṣad, pp. vii, 130.Google Scholar There is similar variation in the title of the Brāhmaṇa.

^{page 364 note 1} For the exact title, see Cowell's, ed., pp. vii, viiiGoogle Scholar; Müller, Max, S.B.E., i, p. xcviii.Google Scholar

^{page 364 note 2} Cowell, , p. vii (MS. F)Google Scholar; Berlin Catalogue, i, p. 19.Google Scholar

^{page 364 note 3} Ibid., p. vii (MS. A).

^{page 364 note 4} Ibid., p. iii (MS. B). This MS. was imperfect, ending abruptly before the conclusion of Adhyāya ix (xi).

^{page 364 note 5} Indische Studien, i, p. 392. It is not at all likely that he had another MS. with this division, and the four books of the Upaniṣad would hardly have been separated.Google Scholar

^{page 365 note 1} See J.R.A.S., 1907, pp. 408 seq.Google Scholar

^{page 365 note 2} S.B.E., xliii, p. 348, n. 1.Google Scholar

^{page 365 note 3} See also Āitareya Araṇyaka, pp. 35 seq., where details are given.Google Scholar

^{page 365 note 4} Ibid., p. 65.

^{page 365 note 5} Friedländer, , p. 7.Google ScholarBrāhmaṇam is not, of course, so used a proper name; cf. Weber, , Indische Studien, xvii, p. 373.Google Scholar

^{page 366 note 1} i, 5; 6. I am not sure how Friedländer takes these passages.

^{page 366 note 2} ii, 17.

^{page 366 note 3} These Adhyāyas seem to be reckoned as making up only one Upaniṣad, unlike books ii and iii of the Aitareya Āraṇyaka, of which Adhyāyas iv–vi of book ii are reckoned as constituting the Upaniṣad par excellence. This double reckoning vindicates Sāyaṇa's accuracy in citing from the Upaniṣ, Aitareya (viz. iii, 2, 2) na ha vā ṛte prāṇād retaḥ sicyateGoogle Scholar, etc., in his commentary on Taittirīya Saṃhitā, ii, 1, 1, 2, 3, against Geldner, , Vedische Studien, ii, p. 306. The Sāṅkhāyana version viii, 2) is slightly different in wording.Google Scholar

^{page 366 note 4} Deussen, : Philosophy of the Upanishads, p. 24.Google Scholar

^{page 366 note 5} Aitareya Āraṇyaka, p. 41.Google Scholar Brahman (m.) is found in iii, 5, and cf. brahmaloka, iii, 3Google Scholar, which word, found also in the Bṛhadāraṇyaka and Chāndogya, almost postulates a personal Brahman: Weber, , Indische Studien, i, p. 396, n.Google ScholarBöhtlingk, , wrongly in my opinion, finds him in Aitareya, ii, 6.Google Scholar None of the passages in Muir, , Texts, v, pp. 320–1;Google ScholarMacdonell, , Vedic Mythology, p. 168Google Scholar, or BR. s.v., are cogent, and I doubt if he is found before the Bṛhadāraṇyaka, and the later part of the Atharva. He is not in the Taittirīya, Aitareya, śatapatha, Kaṣītaki, or Pañcaviṃśa Brāhmaṇas.

^{page 367 note 1} iii, 1.

^{page 367 note 2} a Deussen, I.c.

^{page 367 note 3} ii, 1.

^{page 367 note 4} vi, 1.

^{page 367 note 5} I follow Oldenberg (Buddha, E.T., p. 393Google Scholar, note) rather than BR., Cowell, and Müller, Max (S.B.E., i, p. lxxvii), who read satvanmatsyeṣu.Google Scholar

^{page 367 note 6} Viz., ii, 2, 3, and 4.Google Scholar See Aitareya Āraṇyaka, p. 60.Google Scholar

^{page 367 note 7} It shows no Sāṃkhya traits, Garbe, , Sāṃkhya Philosophe, p. 20.Google Scholar

^{page 368 note 1} Lindner, , Kauṣītaki Brāhmaṇa, p. ixGoogle Scholar; Weber, , Indian Literature, p. 46Google Scholar, Indische Studien, i, p. 404.Google Scholar

^{page 368 note 2} Winternitz, , Geschichte der indischen Litteratur, i, p. 225;Google ScholarMüller, Max, op. cit., xv, p. li.Google Scholar

^{page 368 note 3} See Bühler, , S.B.E., ii ², p. xlii.Google Scholar

^{page 368 note 4} Ibid., pp. xliv seq.

^{page 368 note 5} Supported by no less an authority than DrHoernle, in his admirable Osteology (pp. 106–7)Google Scholar. Cf. also Ludwig, , Ṛyveda, iii, p. 13Google Scholar; Gough, , Philosophy of the Upaniṡads, p. 185.Google Scholar

^{page 368 note 6} Davids, Rhys: Buddhist India, pp. 12–16Google Scholar. Pasenadi held it (Dīgha Nikāya); see Oldenberg, , Buddha, p. 393Google Scholar, n. †. The Buddhist period knows the Kāśikosalas (cf. Weber, , Indische Studien, i, p. 212)Google Scholar; the Brāhmaṇas, the Kāśvidehas, and Kosalavidehas.

^{page 369 note 1} This follows from Bṛhadāraṇyaka Upaniṣad, ii, 1; Kauṣītaki Upaniṣad, vi, 1, where Ajātaśatru refers to Janaka as a contemporary prince. For the contemporaneity of Janaka and Yājñavalkya there is abundant evidence; see Jacob's, Concordance, pp. 369, 771.Google Scholar

^{page 369 note 2} See e.g. Davids, Rhys, op. cit., p. 162Google Scholar; Garbe, , Philosophy of India, p. 69Google Scholar; Macdonell, , Sanskrit Literature, p. 226Google Scholar; Winternitz, , Geschichte der indischen Lilteratur, i, pp. 257–8Google Scholar; Deussen, , Philosophy of the Upanishads, p. 51Google Scholar; Oldenberg, , Buddha, pp. 18, 31.Google Scholar

^{page 369 note 3} See Cowell's, ed., p. vGoogle Scholar; Müller, Max, S.B.E., i, p. xxix.Google Scholar I do not think Cowell, (p. viii)Google Scholar is right in conjecturing that there were two recensions of the Āraṇyaka, and that the different recensions of the Upaniṣad are thence derived. There is no evidence of any such recensions of the Āraṇyaka. What is much more probable is that the Upaniṣad, which was most studied, was handed down in slightly different texts. That preserved in Śaṅkarānanda's commentary has every ppearance of being an attempt at an improved version of the text, and its claim to any great age is not clear.

^{page 371 note 1} Śākalya (vii, 3), Sthaviraḥ Śākalya (vii, 16; viii, 1; 11), Kauṇṭharavya (vii, 14; viii, 2), Pañcālacaṇḍa (vii, 18), Tārkṣya, (sic, vii, 19),Google ScholarVātsya, (Aitareya, Bādhva; viii, 3; 4)Google Scholar, Hārita, Kṛtsna (Aitareya, Kṛṣṇa Hārita; viii, 10)Google Scholar, Kāvaṣeyas, (viii, 11)Google Scholar, āgastya, (vii, 2)Google Scholar, and the Māṇḍdūkeyas, Śauravīra (Sūravīra in Aitareya, ; vii, 2; 8; 9; 10)Google Scholar, Hrasva, (vii, 12; viii, 11)Google Scholar, Dīrgha (not in Aitareya, ; vii, 2)Google Scholar, and Madhyama, Prātibodhīputra Magadhavāsin, (vii, 13)Google Scholar, while Mākṣavya in the Aitareya is replaced by Māṇḍdavya, (vii, 2).Google Scholar

^{page 371 note 2} Indian Literature, p. 50, n. 36.Google Scholar

^{page 371 note 3} For his alphabet, cf. Bühler, , Indian Studies, iii, p. 24.Google Scholar As a teacher he appears in Buddhist tradition, Oldenberg, , Buddha, p. 412.Google Scholar

^{page 372 note 1} iv, 1, 18; Harivaṃśa, 1771. Cf. the Lohicca Sutta of the Dīgha Nikāya.

^{page 372 note 2} Probably it is a Nakṣatra name and need have no connection with the hero of the Epic or with Rādhā and Kṛṣṇa. But Rādhā as a Nakṣatra is late, Whitney, , Atharvaveda, p. 908;Google ScholarLudwig, , Ṛgveda, iii, p. 185.Google Scholar

^{page 372 note 3} See Aitareya Araṇyaka, pp. 51, 52. The Ṛgveda Prātśākhya copies iii, 1, 1, of that work.Google Scholar

^{page 372 note 4} For the construction, cf Āraṇyaka, Śāṅkhāyana, i, 5Google Scholar; Brāhmaṇa, Mantra, ii, 1, 7Google Scholar; Friedländer, , p. 41, n. 2.Google Scholar

^{page 372 note 5} Cf. Müller, Max, Ṛgveda Prātiśākhya, pp. v, vi.Google Scholar

^{page 372 note 6} Cf. Wackernagel, , Altindische Grammatik, ii, i, p. 127.Google Scholar

^{page 373 note 1} Cf. Manu, , i, 64:Google Scholardhvaṃsi seems peculiar to the Śārikhāyanas; cf. Sūtra, Śrauta, xiv, 82, 1.Google Scholar

^{page 373 note 2} Yāny anyāni kṣudrāṇi mahābhūtaiḥ saṃdhīyante is a good deal more intelligible than kṣudramiśrāṇīva.

^{page 374 note 1} v, 3.

^{page 374 note 2} vi, 2, 13 (Mādhyandina) = vi, 1, 13 (Kāṇva). References are made to the former text, when not otherwise specified. Cf. Pischel, , Ved. Stud., i, pp. 10, 234.Google Scholar

^{page 374 note 3} Bloomfield, : Vedic Concordance, pp. 477, 508.Google Scholar

^{page 374 note 4} In the Śāṅkhāyana itself, vii, 10. For l=r, cf. Macdonell, , Vedic Grammar, pp. 43 seq.Google Scholar

^{page 374 note 5} Indian Literature, p. 132, n. ^*Google Scholar

^{page 374 note 6} The gen. with brūyāt, for the dative of Chāndogya and Bṛhadāraṇyaka, is a sign of later date.

^{page 375 note 1} iv, 5.

^{page 375 note 2} For similar sets of metres, ct. āraṇyaka, Aitareya, v, 1, 4Google Scholar; āraṇyaka, Śāṅkhāyana, i, 7Google Scholar; Friedläander, , p. 44, n. 1.Google Scholar

^{page 376 note 1} For similar spells, cf. the references in Bloomfield's, Vedic Concordance, p. 126 (s.v. aśmā, aśmeva).Google Scholar

^{page 376 note 2} Cf. Weber, , Indian Literature, pp. 45, 110;Google ScholarMacdonell, , Vedic Mythology, p. 75.Google Scholar

^{page 376 note 3} See Speyer, , Vedische und Sanskrit-Syntax, p. 73Google Scholar

^{page 377 note 1} For v. 1 see Atharvaveda, , iii, 22, 1Google Scholar; for v. 2, iii, 22, 3 and 4; for vv. 3 and 4, xiv, 1, 35; vi, 19, 1; and for v. 5, vi, 69, 3.

^{page 377 note 2} Cf. Atharvaveda, , vi, 69, 2.Google Scholar

^{page 377 note 3} Cf. Atharvaveda, , v, 28, 14.Google Scholar

^{page 377 note 4} Amulets for medical purposes are common (Bloomfield, , Atharvaveda, p. 59)Google Scholar, and also, as here, for help against foes (ibid., p. 67.) See especially Atharvaveda, , i, 29; ii, 7; iii, 6; vi, 15; x, 3; 6; xix, 28–30; 32; 33, etc.Google Scholar

^{page 378 note 1} Cf. Bloomfield, , Atharvaveda, pp. 41 seq.Google Scholar

^{page 378 note 2} Here pūrvapābhyām might be read for pūrvapābhyām and so make good the metre.

^{page 378 note 3} The exact numbers in these eases depends, of course, on the mode in which the necessary resolutions of Sandhi are made, and on the precise reading adopted in the text, but the general results remain unaffected. In v. 12^a I would read an vṛśca paścāt pra vṛścopariṣṭāt; for an vṛśca, cf. Wackernagel, , Altindische Grammatik, i, p. 59Google Scholar; Maedonell, , Vedic Grammar, p. 11.Google Scholar

^{page 378 note 4} Cf. Oldenberg, , Z.D.M.G., xxxvii, pp. 62 seq.;Google ScholarS.B.E., xxx, pp. xii seq., xxxiv seq.Google Scholar; Prolegomena, pp. 26 seq.Google Scholar; Gurupūjākaumvdī, pp. 9 seq.Google Scholar; Keith, , J.R.A.S., 1906, pp. 1 seq., 486.Google Scholar

^{page 379 note 1} Cf. Hopkins, , Great Epic of India, pp. 264 seq.Google Scholar; Arnold, , Vedic Metre, pp. 183 seq.;Google ScholarLudwig, , Ṛgveda, iii, p. 50.Google Scholar

^{page 379 note 2} Pisohel, , Vedische Studien, i, p. 59;Google ScholarWackernagel, , Altindische Grammatik, i, pp. 317, 321;Google ScholarMacdonell, , Vedic Grammar, p. 65, n. 12;Google ScholarArnold, , p. 78;Google ScholarHopkins, , India Old and New, p. 46, n. 1.Google Scholar

^{page 379 note 3} Cf. Hopkins, , Great Epic, p. 260Google Scholar; Keith, , J.R.A.S., 1908, p. 202.Google Scholar

^{page 380 note 1} See Wackernagel, , op. cit., p. 320.Google Scholar

^{page 381 note 1} So I had emended for vede of the Bodleian MS., and the Berlin MS. confirms the emendation. The omission of a double letter is very frequent n the MS. Yathā katliaṁ cana is comparatively late.

^{page 381 note 2} The Bṛhadāraṇyaka, (iii, 7, 1) has a proper name, Kabandha Ātharvaṇa, where it cannot mean ‘corpse’.Google Scholar

^{page 381 note 3} See Delbrück, , Synt. Forsch., v, p. 52;Google ScholarMacdonell, , Vedic Grammar, p. 91.Google Scholar

^{page 381 note 4} p. 19. The verses are also cited in the Commentary on the Saṃhitopaniṣad Brāhmaṇa, p. 38 (ed. Burnell).

^{page 382 note 1} yad gṛhītam avijñātaṃ nigadenaiva śabdyate | anagnāv iva śuṣkaidho na tu jvalati karhidt ║ Roth's emendation na taj is not necessary.

^{page 382 note 2} See the references in Jacob's, Concordance, pp. 652–3, to which I am much indebted.Google Scholar

^{page 382 note 3} I.e., as regards the exact place occupied in the Āraṇyaka by these books. That they were once a part of the Āraṇyaka is, I think, quite certain. Cf. also Hillebrandt, , Rom. Forsch., v, p. 331.Google Scholar

^{page 383 note 1} Bodleian Catalogue, p. 60.Google Scholar

^{page 384 note 1} Cf. Upaniṣad, Chāndogya, i, 5, 2.Google Scholar The name was perhaps Kahoḷa. Cf. Wackernagel, , Altindische Grammatik, i, p. 221;Google ScholarWeher, , Indische Studien, i, p. 404.Google Scholar

^{page 384 note 2} Oldenberg, : S.B.E., xxix, p. 3.Google Scholar

^{page 384 note 3} Philosophy of the Upanishads, p. 24.Google Scholar If iii–vi are later, then of course a fortiori vii and viii.

^{page 385 note 1} i, 5, 1; Deussen, , op. cit., p. 217.Google Scholar

^{page 385 note 2} ii, 3, 1.

^{page 385 note 3} Compare the solitary reference to tapas in Kauṣītaki, , iii, 2Google Scholar, with the numerous passages cited in Jacob, , Concordance, p. 396;Google ScholarDeussen, , op. cit., p. 69.Google Scholar

^{page 385 note 4} J.R.A.S., 1907, pp. 410–12.Google Scholar

^{page 385 note 5} To judge from the extant specimens of Āraṇyakas, the relation of Āraṇyaka and Upanisad might be regarded as that of whole and part. Each Āraṇyaka contains, inter alia, several Upaniṣads.

^{page 386 note 1} S.B.E., xxix, pp. 4, 5.Google Scholar Cf. also view, Bhandarkar's (Report, 1894, pp. 2 seq.)Google Scholar, accepted by Hillebrandt, (Ritual-Litteratur, p. 28), that Ṛāṅkhāyana is a mere Sūtra caraṇa.Google Scholar

^{page 386 note 2} Ritual-Litteratur, p. 25;Google ScholarṚāṅkhāyana Ṛranta Sūtra, i, p. viii.Google Scholar

^{page 386 note 3} The matter might be further complicated by regarding Guṇākhya Ṛāṅkhāyana as the author of the Śrauta as contrasted with the Gṫhya Sūtra. I do not, however, think this view probable, and Oldenberg, who once was inclined to differentiate the authors (though without aming the elder Guṇākhya), later admitted the insufficiency of the evidence (see Indische Studien, xv, pp. 11, 12Google Scholar; S.B.E., xxix, pp. 4, 5).Google Scholar

^{page 387 note 1} Cf. Macdonell, , Bṫhaddevatā, i, pp. xxii–xxiv.Google Scholar I do not attach any weight to the tradition, even if found in the Bṫhatkathā, which attributes Pāṧini to the reign of the last Nanda (despite Bühler, , Indian Studies, iii, p. 21, n. 1, 27, n. 1),Google Scholar and associates him with Kātyāyana and Āśvalāyana. But the fact that the tradition very possibly existed in the first century A. D. is of interest as tending to show that these writers cannot be dated very near the Christian era, or their chronological relations could not have been confused. Ludwig's date for the Śāṅkhāyana Śrauta Sūtra, 500 B.C. (Ṛgveda, iii, p. 196), rests on no evidence.Google Scholar

^{page 388 note 1} Cf. now Weber, , Sitzungsber. der Berl. Akad., 1895, p. 859, n. 4.Google Scholar

^{page 388 note 2} Mention may here be made of the only important correction supplied by the MS. in the Bodleian to the excellent text of Adhyāyas i and ii published by DrFriedländer, . In ii, 17Google Scholar, the text of the edition reads: tasya vā etasya bṛhatīsahasrasya ṣaṭ triṃśad akṣarāṇām sahasrāṇi bhavanti tāvnti śatasaṃvatsarasyāhāny āpnoti. This is just possible, but the reading of the Bodleian (and also, I now find, of the Berlin MS.), which adds before āpnọti the words bhavanti ta(c) chatasaṃvatsarasyāhāny, is clearly right, the omission being very natural. Smaller corrections are: (1) in ii, 15, it reads divaṃ jaya divaṃ jaya, a Pratīka elsewhere unknown; divaṃ yaya apparently refers to R. V. viii, 34, I^d (repeated in he later verses of the hymn); (2) in ii, 18, it (like the Berlin MS.) inserts the necessary tad in the verse baḷ itthā tad vapuṃe dhāyi darśatam (R. V. i, 161, l^a), as in the Sütra, Śrauta, xviii, 23, 14Google Scholar; (3) in ii, 4, it confirms the reading bhütechaṃdāṃ sāma by reading bhü;techaṃdāṃ sāma; clearly the differences of reading (cf. Friedänder, , p. 18, n. 2; p. 37Google Scholar, n. 1) are all due to the accidental insertion of the superfluous Anusvāra before d; 4) in ii, 8, it has dakṃiṇataḥ and uttarataḥ for dakṃiṇaḥ and uttaraḥ. It has the correct bhavati (p. 21, 1. 7)Google Scholar and pratnaihā (p. 25, 1. 5).Google Scholar

Neither the Berlin MS. nor the Bodleian MS. yields substantial correction for the text of the Upaniṣad, in which they agree very closely with A in Cowell's, ed. In i, 2 (p. 11)Google Scholar, they read dvādaśatrayodaśomāsaḥ; in i, 3 (p. 14)Google Scholar, yayaṣṭihā; in i, 7 (p. 27)Google Scholar, ghrāṇena; in ii, 11 (p. 57)Google Scholar, vedo; in all these cases agreeing with A. In i, 4 (p. 19), the Bodl. has dhunvavāte, the Berl. dhunuvāte, which, in conjunction with the readings of A, B, C, E, shows that a third person dual must be read for Cowell's, dhunute. In i, 5 (p. 23)Google Scholar, the Bodl. has prācīnātānāni, like A, the Berl, . °nātāni-. In ii, 11 (p. 58)Google Scholar, both, with A, have mā bhetthāḥ, then Bodl. has mā vyadhiṣṭliāḥ, Berl, . vyadhiṣṭliāḥGoogle Scholar, A vyatiṣṭhāḥ. In ii, 12 (p. 61)Google Scholar, Bodl. has mṛtvā na mṛchante, Berl. mṛtvānnam ṛchata. In iv, 1, both have kālakhañjān, corrected to °khāñjān in Bodl. as in A. In iv, 19 (p. 120), both have animnyas, A °yās. In iv, 15 (p. 114), Berl. and A have svapnyayā, and in iii, 5, Berl. has several times a correction adūduhat for the strange udūḲham.