^{page 337 note 1} On the contrary the line of Pūru, etc., is missing in Saurapurāṇa, 31, 8 sqq. There only Yadu and his descent is to be found.

^{page 337 note 2} It is hardly understandable why Kirfel follows here Bḍ-Vā, when he always recognizes the validity of H.B.

^{page 338 note 1} Cf. the author's remarks in Anthropos, 1939, p. 467.

^{page 338 note 2} In H, 34, 29 (Kirfel, 449, 1) Anamitra is called the youngest offshoot of the Vṛṣṇi race. But Nīlakaṇṭha gives another interpretation saying: “He was the eldest son of Kroṣṭṛ” (by Gāndhārī) in accordance with H, 34, 1. But according to H, 38, 12 (Kirfel, 436, 17) Mādrī gave birth to Anamitra. The line is slightly confused.

^{page 339 note 1} P, 276, 5 sqq.; Bh, x, 56–7; Vai, 122.

^{page 340 note 1} The derivation of Vṛṣṇi is different in H, 34, 3 (Kirfel, 449, 2) and H, 37, 2 (Kirfel, 432, 2), and the name Vṛṣṇi as the name of a family is derived in H, 33, 35 (Kirfel, 420, 63) from a Vṛṣaṇa who descended from a Vṛṣa (ibid., 52). But this Vṛṣa has not been mentioned before in H or B. Nīlakaṇṭha explains that this Vṛṣa is the same as Payoda (= cloud; vrṣa = rain) mentioned previously (Kirfel, 410, 2), but this explanation shows only that the problem is unsolved. Bḍ-Vā therefore have changed the name of a man called Krsna into Vṛṣa (Kirfel, 419, 47). B in the same line has the reading Vṛṣaṇa.

^{page 340 note 2} Not noticed by Kirfel; cf. especially his article in Festschrift Jacobi, p. 314 sq.

^{page 341 note 1} As, e.g., in § 4; cf. § 6; it is doubtful in § 8 (twice), § 13a, 1, and improbable in § 11 and § 17 (twice); cf. § 12, Puṣḍra, etc.

^{page 341 note 2} Not noticed by Kirfel. A fuller discussion of the problem of this girl will be given, I hope soon, on another occasion when dealing with the whole biography of Kṛṣṇa.

^{page 343 note 1} See note 2, p. 341.

^{page 345 note 1} H, 12cd has been replaced by B, 21cd, 22ab. H, 15cd–19, are missing in B after B, 24, but H, llcdef = B, 180, 26cd, 27ab (cf. infra about B, 180). H, 18–19, have been replaced by B, 180, 18 sqq. H, 21ab, is missing in B, also H, 28ef, but this line is only a duplicate of H, 36ab. H, 49cd, 55, 59, 83ab (= 84c), 111cd, are missing in B. H, 90ab (= B, 92cd), has been transposed in B after the line 92ab corresponding to H, 91cd. H therefore has some interpolations.

^{page 345 note 2} Cf. the last note.

^{page 345 note 3} B, 213,165–171, is different. Cf. note 2, p. 341. The stanzas of the Kṛṣṇa avatāra contain several problems, to be explained upon another occasion.

^{page 346 note 1} K, 1, 2, and the Nepali MS.: Sukthankar, , Annals Bhandarkar ORI, xix, 201 sqq., 205Google Scholar. They are a very late interpolation, not more than 300 years old (as Sukthankar wrote to me in a letter kindly answering a question of mine).

^{page 347 note 1} This is not a strong proof, because Mt. may have quoted this passage from any MS., not from the archetype of the Ādiparvan.

^{page 348 note 1} But in Rām., vii, 59,14 sq., the curse upon Yadu is that his descendants should become yātudhānas, sorcerers. We may consider that, according to the Bām. Yadu belonged to the descent of Ikṣvāku and was not the ancestor of the Yādava race and Kṛṣṇa (cf. § 15).

^{page 348 note 2} Cf. Sukthankar ad Mbh, i, 71, 1; Kirfel, 409.

^{page 349 note 1} 840* is missing also in the old Nepali MS. (of. note 1, p. 346), according to Sukthankar's letter.

^{page 349 note 2} The line between (H, 23cd) has been taken from the first version (Mbh, i, 70,37cd), cf. supra (b).

^{page 350 note 1} A similar criterion is afforded by the motif of the complaint of the Earthgoddess in H, 42, 4sqq. (B, 181, 8sqq.), and Mbh, i, 58; cf. Ruben, W., Intern. Arch. f. Ethnogr., Lejden, 1939, Suppl. zu xxxvii, Eisensohmiede und Dämonen in Indien, p. 234Google Scholar.

^{page 350 note 2} Cf. § 9(g), and § 7.

^{page 350 note 3} The first book of the Rām. (i, 23 sqq.) is such a collection of local traditions, told by Viśvāmitra.

^{page 351 note 1} Jacobi, , Das Rāmāyaṇa, 72 sq.Google Scholar; Ruben, , Studien zur Textgeschichte des Rāmāyaṇa, 53Google Scholar; Sluskiewicz, , Contribution à l'Histoire du Rāmāyaṇa (Krakow, 1938), 1 sqq., 266 sqq.Google Scholar

^{page 351 note 2} There is such a great difference between the morals of the Rām. and H, that it sometimes looks like a polemic. This point also will be discussed on another occasion.

^{page 354 note 1} The legend of Kandu (Vi, i, 15, 11–58) is taken from B (178, 1–194) or vice versa (translated by Zimmer, , Maya, 72 sqq.Google Scholar):

^{page 354 note 2} The corresponding figures are to be found in Kirfel's table at p. xx.

^{page 355 note 1} Pargiter, , The Dynasties of the Kali Age (Oxford, 1913)Google Scholar; Kirfel, xviii.

^{page 356 note 1} Cf. Ruben, W., A Volume, of Eastern and Indian Studies in honour of F. W. Thomas (1939)Google Scholar, on the original text of the Kṛṣṇa Epic, 188 sqq.

^{page 357 note 1} It is a strange fact that even in so late a text as Bh the line of historical kings does not go beyond that of the other Purāṇas.

^{page 357 note 2} Or we arrange the facts in this order then. B took its vamsa and its Kṛṣṇacarita from Vi, shortening the carita; later on it replaced the vaṁśa by that of H and on that occasion took the chapter on Viṣṇu's avatāras from H. But such an assumption that the original B was composed as late as the thirteenth century with the whole text of Vi, books i–iv, instead of its sargas 1–17, as wo read them to-day, is highly improbable so long as MSS. of such a B are not forthcoming. And if it were proved to be true, it would not touch our theory of the line Mbh-H-Purāṇas.

^{page 358 note 1} The seven chapters could be reduced to five by omitting the fourth and combining the sixth and seventh. The sarga might correspond with ādisarga, pratisarga (prati, distributive, “in detail”) with bhūtasarga, vaṁśa with Dakṣavisṛṣṭi, Manvantara with Manvantara, vaṁśānucarita with the Solar and Lunar kings.