^{page 886 note 1} The following abbreviations are used in the notes: Vā = Vāyu; Bḍ = Brahmāṇḍa; Br = Brahma; Hv = Harivaṁśa; Śv = śiva; Lg = Liṅga; Vṣ = Viṣṇu; Ag = Agni; Gr = Garuḍa; Bh = Bhāgavata; MBh = Mahābhārata. I treat the Harivaṁśa as a Puraṇa, which is what it is really.

^{page 886 note 2} See JRAS, 1910, p. 33Google Scholar; but the Bḍ appears to make her Gādhi's wife, see infra, n. 4, v.r. †.

^{page 886 note 3} Gādhi in the Epics and Purāṇas, Gāthin in brahmanical books. The former seems to be merely the Prakrit form of the latter.

^{page 886 note 4} This portion is given in Vā 91, 65–66, 85; Bḍ iii, 66, 35–37, 56; Br 10, 27–29, 48; 13, 91 (only first line); Hv 27, 1429–1431, 1450; 32, 1765 (only first line); MBh xii, 49, 1720, 1721, 1744. The text collated therefrom runs thus:—

Paurukutsy abhavad* bhāryā Gādhis † tasyām ajāyata

Gādheḥ ‡ kanyā mahābhāgā nāmnā Satyavatī śubhā §

tāṁ Gādhiḥ Kāvya-putrāya ∥ Ṛcīkāya dadau prabhuḥ ¶

tataḥ Satyavatī putraṁ janayāmāsa Bhārgavam

tapasy abhirataṁ dāntaṁ Jamadagniṁ śamâtmakam:

where * Vā, Paurukutsâbh°, Br generally Paurā yasyâbh°, MBh omits this line: † Bḍ Gādhes: ‡ Bḍ pūrvaṁ, Vā pūrva: § Vā all accusatives: ∥ So Br; Vā putraḥ Kāvyāya, Bḍ putra-kāmāya, MBh and Hv Bhṛguputrāya: ¶ Here follows the story of the two carus to explain why the kṣatriya Viśvāmitra became a brahman, and why the brahman Jamadagni, who was born at the same time, had for a son the terrible warrior Rāma. Vṣ (iv. 7, 6, 16) gives the account in prose, and Bh (ix, 15, 4, 5, 11) briefly; they are of no value for collation.

^{page 887 note 1} This is positively stated in MBh ix, 41, 2300.

^{page 887 note 2} This portion is given in Br 10, 55–57; Hv 27, 1457–9: 11. 1, 3, 5 disconnectedly in Vā 91, 87, 93; Bḍ iii, 66, 58, 65 (and differently but equivalently in MBh xiii, 4, 246–7): 11. 1, 3 in MBh xii, 49, 1745. The collated text runs thus:—

Viśvamitraṁ tu dāyādaṁ Gādhiḥ Kuśika-nandanaḥ

janayāmāsa putraṁ tu tapo-vidyā-śamâtmakam

prâpya brahmarṣi-samatāṁ yo 'yam brahmarṣitāṁ * gataḥ

Viśvāmitras tu dharmātmā nāmnā Viśvarathaḥ smṛtaḥ

jajñe Bhṛgu-prasādena Kauśikād vaṁśa-vardhanaḥ

where * Hv saptarṣitāṁ, Vā and Bḍ read this half line jagāma Brahmaṇā vṛtaḥ, MBh corrupt: Vṣ iv, 7, 16; Gr i, 139, 5; Ag 277, 17; Bh ix, 16, 28, 29, say Viśvāmitra was Gādhi's son.

^{page 887 note 3} Macdonell, & Keith, , Vedic Index, ii, 311–2.Google Scholar

^{page 888 note 1} The genealogies suggest that he was succeeded by his son Aṣṭaka, since the two following lines are found in Vā 91, 103; Bḍ iii, 66, 75; Br 10, 67–8 and 13, 91–2; Hv 27, 1473 and 82, 1775–6; where they are placed as a separate statement. They appear to refer to the dynastic succession, because they are given as such in Br 13, without noticing Viśvāmitra's other sons.

Dṛṣadvatī-sutaś câpi* Viśvāmitrāt † tathÂṣṭakaḥ

Aṣṭakasya suto Lauhiḥ ‡. Prokto Jahnu-gaṇo mayā:

where * Br 13 reads Viśvāmitras tu Gādheyo; † Br 10 Vaiśvāmitras; ‡ Vā corruptly yo hi. Vṣ iv, 7, 16–18; Ag 277, 18; and Bh ix, 16, 36 merely name Astaka as a son.

^{page 888 note 2} If Viśvāmitra had not been a kṣatriya of the highest rank, could he have ventured to contend with the great brahman Vasiṣṭha, and have successfully established his claim to brahmanhood? Otherwise, the immense body of tradition (and marvel) concerning his contest with Vasiṣṭha could never have grown up.

^{page 888 note 3} Muir gives this story in his Sanskrit Texts, I, 87.Google Scholar

^{page 889 note 1} Vā 88, 78–116; Bḍ iii, 63, 77–114; Br 7, 97–5, 23; Hv 12, 717–53; Śv vii, 60, 81–61, 19; Lg i, 66, 3–10 (giving only 11. 1–3, 6, 8, 9, 11, 13, 14, 66, 69–71).

^{page 889 note 2} Vṣ iv, 3, 13, 14 and Bh ix, 7, 5, 6 mention the story curtly and also agree generally so far as they go, but are of no use for collation here.

^{page 889 note 3} Śv mahātmabhiḥ.

^{page 889 note 4} So Bḍ; Vā °sān; Lg hatvâmitaujasam: see p. 897, n. 2.

^{page 889 note 5} Br kṛtôdvāhā hṛtā caiva.

^{page 889 note 6} So Lg; Bḍ na; Vā sam-.

^{page 890 note 1} Br sāhasāc.

^{page 890 note 2} Vā saṅkarṣaṇa-.

^{page 890 note 3} So Śv; Br kāmârtaḥ; Hv Kāmāt hanyāṁ sa.

^{page 890 note 4} Br, Hv adharma-śaṅkunā tena; Śv adharma-saṅginaṁ taṁ tu.

^{page 890 note 5} Vā pitā; Bḍ taṁ sa.

^{page 890 note 6} Vā, Śv 'vadat.

^{page 890 note 7} Vā, Bḍ ekaḥ.

^{page 890 note 8} Vā dhār°.

^{page 890 note 9} One Br MS purasya svasya cântike.

^{page 890 note 10} Vā mukto.

^{page 890 note 11} Br sāgarânte tu.

^{page 890 note 12} Vā, Bḍ śiṣṭānāṁ.

^{page 890 note 13} Śv vikrīṇantīṁ, svam atmajam.

^{page 890 note 14} Vā, Bḍ °buddhir.

^{page 890 note 15} Śv anukrośârtham.

^{page 890 note 16} Vā, Bḍ baddho.

^{page 890 note 17} Vā vīryeṇa.

^{page 890 note 18} Vā, Bḍ tasya vratena.

^{page 890 note 19} Śv poṣayāmāsa vai tadā.

^{page 891 note 1} So Bḍ; Vā apacat; Śv câkṣipat. Br, Hv read this half line, māṁsaṁ vṛkṣe babandha saḥ, which is good.

^{page 891 note 2} Hv avahat, Br avasat.

^{page 891 note 3} Br, Hv Tasmin vana-gate nṛpe, which is good.

^{page 891 note 4} Śv tīrthaṁ gāṁ.

^{page 891 note 5} Vā, Bḍ manyunā.

^{page 891 note 6} Vā rudaṁs.

^{page 891 note 7} Br, Hv tyajyamānaṁ, which is good.

^{page 891 note 8} Br nivār.°

^{page 891 note 9} Br bahunā.

^{page 891 note 10} Hv, Śv ha; Br na.

^{page 891 note 11} Vā, Bḍ evaṁ.

^{page 891 note 12} So Bḍ. Vā tān vai. Br, Hv tasmād.

^{page 891 note 13} So Hv. Br. dharmaṁ.

^{page 891 note 14} Śv Vasiṣṭhāt tu na taṁ śaṁsati.

^{page 891 note 15} So Hv. Br Bhārata. These vocative expletives have probably ousted some such expression as so 'vadat or acintayat.

^{page 891 note 16} Vā °vrate.

^{page 891 note 17} Vā, Śv °ṣṭho.

^{page 891 note 18} Vā guru.

^{page 891 note 19} Bḍ tapaḥ.

^{page 891 note 20} Śv sa ca.

^{page 891 note 21} Vā, Bḍ 'budhyad upāṁśu-vratam asya vai.

^{page 891 note 22} So Hv. Śv hy aparitoṣāya, Br aparitoṣaś ca, Vā côparate yo yat, Bḍ tu paramo roṣaḥ.

^{page 891 note 23} Br vihiṭāṁ; Vā bahudhā.

^{page 891 note 24} Śv kṛtavān vai; Vā, Bḍ kṛtêyaṁ ca.

^{page 891 note 25} Vā, Bḍ tato.

^{page 891 note 26} Bḍ na vārayat.

^{page 892 note 1} Bḍ naṣṭe (for rāṣṭre).

^{page 892 note 2} Śv asyânte vai 'bravīn muniḥ.

^{page 892 note 3} Śv udvahad, Br avahad.

^{page 892 note 4} Vā, Bḍ dhenuṁ.

^{page 892 note 5} So Vā, Bḍ. Śv dāśa; Br deśa; Hv misreading it as daśa, “ten”, inserts two lines of mistaken explanation—

mattaḥ pramatta unmattaḥ śrāntaḥ kruddho bubhukṣitaḥ tvaramāṇaś ca bhīruś ca lubdhaḥ kāmī ca te daśa.

^{page 892 note 6} Vā, Bḍ dṛṣṭvā.

^{page 892 note 7} Vā, Bḍ taṁ tadâtyajat.

^{page 892 note 8} Hv different but equivalent. Br, Śv merely Vasiṣṭha uvāca.

^{page 892 note 9} Śv krūraṁ.

^{page 892 note 10} So Vā Śv. Bḍ apohya vai; Br, Hv asaṁśayam.

^{page 892 note 11} Śv naśyetāṁ.

^{page 892 note 12} Śv purā.

^{page 892 note 13} Lg reads —

sarva-lokeṣu vikhyātas Triśaṅkur iti vīryavān

Vasiṣtha-kopāt puṇyâtmā rājā Satyavrataḥ purā.

^{page 892 note 14} Br anena.

^{page 892 note 15} Vā, Bḍ tadā.

^{page 892 note 16} So Br, Hv and Śv and add the line—

sa-śarīro vraje svargam ity evaṁ yācito varaḥ

But Vā Bḍ read guruṁ for varaṁ.

^{page 893 note 1} Bḍ, Śv jāte.

^{page 893 note 2} So Hv: others different but equivalent.

^{page 893 note 3} Bḍ yoj°.

^{page 893 note 4} After this Vā inserts two lines—

Vindhya-pārśve mahāpuṇyā nimnagā giri-kānane

tasya snānena sambhūtā Karmanāśā śubhā, nadī.

^{page 893 note 5} Br differently but equivalently. Here Br, Hv end.

^{page 893 note 6} This line is only in Vā Bḍ.

^{page 893 note 7} So Vā. Bḍ °tîmaṁ ślokam.

^{page 893 note 8} This verse only in Vā Bḍ.

^{page 893 note 9} This verse only in Vā.

^{page 893 note 10} The name Vidarbha, mentioned in the Vā Bḍ and Lg, cannot mean “king or prince of Vidarbha”, because (1) the kingdom of Vidarbha did not come into existence till later; (2) such an insult offered to a king or prince would have been avenged by war, yet there is no suggestion of any such reprisals being feared, while it is distinctly suggested that Vasiṣṭha might have mitigated the punishment; and (3) the term kulapāṁsana implies that the prince had dishonoured his rank, and the rape of a mere city maiden was a disgraceful offence. Vidarbha may be the name of the citizen, if it is possible it might have been handed down.

^{page 893 note 11} They lived not far from the city presumably (see p. 890, n. 9), because they would have found their livelihood as dog-killers and dogeaters chiefly by being near the city. At this day there are castes, who are employed to kill off the dogs in a town that have multiplied and become a nuisance.

^{page 894 note 1} So Br, Hv, Śv, Lg and Bḍ. Vā has altered the meaning to the absolute opposite by reading samprāpiteṣu, showing that it has been deliberately tampered with, so as to exaggerate Satyavrata's guilt and consequently to justify Vasiṣṭha's want of pity: and it and Bḍ imply the same in 1. 40. Compare also divaukasaḥ in 1. 2; where Lg reading amitaujasam refers no doubt to the officiating priest (see p. 897, n. 2), but Vā and Bḍ by reading “gods” have given an impossible exaggeration to Satyavrata's violence.

^{page 894 note 2} Where the alteration of the latter half in Vā and Bḍ yields little sense, and Śv is obviously corrupt.

^{page 895 note 1} The Vā omits the negative in 1. 49, and reads falsely that he did oppose the banishment.

^{page 895 note 2} So Br, Hv and śv. The Vā and Bḍ say he proposed to anoint Satyavrata himself afterwards, but it is hardly probable that Vasiṣṭha would anoint the prince whom he had allowed to be degraded by years of association with the lowest outcastes, and would thus lay himself open to possible vengeance from that prince. The difference however does not affect the story, for he never carried out his proposal, whichever it was. Satyavrata's son would have been growing up in the royal seraglio, completely under Vasiṣṭha's custody and training, for the story implies that Satyavrata was banished alone.

^{page 896 note 1} MBh xiii, 137, 6257; and xii, 234, 8601.

^{page 897 note 1} Indra alone was considered a possible personal name, if the ascription of Rigveda x, 38 can supply any testimony.

^{page 897 note 2} The suggestion on p. 903 would support this inference. In this connexion 1. 2 of the text of Vā Bḍ and Lg is noteworthy (p. 889). Satyavrata, when interrupting the wedding ceremony, assailed “the gods” (Bḍ divaulcasaḥ; Vā divaukasān) or amitaujasam (Lg). The former reading is absurd, but the latter is intelligible as referring to the priest who was performing the ceremony, and suggests that the former reading should be Divaukasam. Divaukas is nearly equivalent to Devarāj, and its use here (for the sake of the metre) would be intelligible, if the priest were Vasiṣṭha himself and his name was Devarāj. The name Devarāj would thus elucidate a corrupt word; but this touch to the story occurs only in Vā and Bḍ which exaggerate Satyavrata's misdeeds, and Br, Hv and Śv, which are more trustworthy, know nothing of it.

^{page 897 note 3} This is probable. He was obliged to leave them when undertaking a long course of austerities, yet would not have deserted them nor left them unbefriended. Kosala was one of the most powerful and safe kingdoms, and he could expect good treatment for them there since he was related to the royal house of Ayodhyā; see p. 886.

^{page 897 note 4} “Marshy region near the sea”; several are mentioned in East, South and West India.

^{page 897 note 5} But MBh says he gained brahmanhood at Rusaṅgu's tīrtha on the Sarasvatī (ix, 40, 2270–9; 41, 2307).

^{page 898 note 1} It is explained (1. 25) that the boy got the name Gālạva because his mother bound him by the neck (gala). This is no doubt one of the many fanciful derivations to be found in the Purāṇas, and indeed the whole of this incident may have been made up as an explanation for the name.

^{page 898 note 2} See p. 893, n. 11.

^{page 899 note 1} This explanation of the name may be doubted, just like that of Gālava in p. 898, n. 1.

^{page 899 note 2} The line which the Br, Hv and Śv insert (see p. 892, n. 16) is obviously an interpolation, for all three proceed to say that Viśvāmitra's response was, not to raise Satyavrata to the sky (a priestly notion), but to restore him to the throne and offer sacrifice for him (the natural desire of an heir-apparent).

^{page 900 note 1} His position was the same in the eyes of the people, if he proposed to restore Satyavrata (as Vā and Bḍ say); see p. 895, n. 2. It may be added that his priestly regime could hardly be pleasing to the kṣatriyas.

^{page 900 note 2} This would explain why he did nothing more than threaten Satyavrata for the personal affront in killing the cow, if that incident be true.

^{page 901 note 1} e.g. Sindhudvīpa, and Devāpi, , MBh ix, 41, 2294Google Scholar. The genealogies, even in the Vāyu, say that Kāṇvas, Gārgyas, Sāṅkṛtis, Maudgalyas and Rathītaras were of kṣatriya origin.

^{page 901 note 2} Vasiṣṭha and Viśvāmitra are discussed in Muir, 's Sanskrit Texts, I, 75 ffGoogle Scholar., but the various Vasiṣṭhas and Viśvāmitras are not distinguished clearly. It is hopeless to discuss them except with the chronological aid of the genealogies. Many Vasiṣṭhas are mentioned in tradition, and four are prominent in kṣatriya tradition, namely, in chronological order, (1) this Devarāj, (2) the priest to Sagara, king of Ayodhyā, (3) the priest to Kalmāṣapāda Saudāsa, king of Ayodhyā, and (4) the priest who helped Saṁvaraṇa to recover the Paurava kingdom from the successors of Sudāsa, king of Pañcāla: but in some stories the first two are not always distinguished nor the second two in others, while in brahmanical stories they are often jumbled up together. Similarly with this, the first and greatest, Viśvāmitra, and his descendants. There was rivalry between the later Vasiṣṭhas and Viśvāmitras.

^{page 901 note 3} Narrated in Muir, 's Sanskrit Texts, I, 98.Google Scholar

^{page 902 note 1} The MBh fable is narrated in id. p. 95.

^{page 902 note 2} Let us put ourselves into their position. What would have concerned and interested them was prince Satyavrata's touching story, his sufferings at Vasiṣṭha's hands and his ultimate success through Viśvāmitra's aid—not a portentous conflict between Viśvāmitra and Vasiṣṭha about brahmanhood, wherein he was but a puppet. The latter is a brahmanic view; and a late one too, embodying the strict brahmanic ideas of a subsequent age (see p. 901, n. 1).

^{page 902 note 3} Fortunately the general disregard of history, which the brahmans entertained, though it has supplied no real history of ancient times, has permitted them to preserve a large mass of kṣatriya tradition, the bearing of which on their own stories they did not perceive.

^{page 902 note 4} This may imply that he did not survive long, and it is probable, because he must have been shaken in mind and body by the ordeal he had undergone.

^{page 903 note 1} This is what paurāṇikā janāḥ (1. 73) must mean. These words obviously cannot refer to the Purāṇa. compilers, because the Vā and Bḍ cite the verses as more ancient than themselves.

^{page 903 note 2} Especially if Triśaṅku died soon after his restoration, in consequence of the hardships inflicted on him by Vasiṣṭha.

^{page 903 note 3} The reverse is not probable. If Vasiṣṭha's name was really Indra (see p. 897), the misunderstanding was inevitable.