texts can have any doubt who the borrower is, as the Purâna regularly substitutes easy readings for difficult ones, and adds numerous explanatory verses. Besides, Nârâyana, as well as Kullûka 1, quotes verses of the Bhavishya-purâna from a section on penances not found in the accessible MSS., which likewise are clearly intended to explain the text of our Samhitâ. All this is however useless, as for the present it is impossible to determine the date of the Purâna even approximatively. Professor H. H. Wilson 2, who has a very mean opinion of the book, declares that it cannot lay claim to a high antiquity, and seems to consider it a production of the ninth or tenth century A. D. Professor Aufrecht's discovery 3 that the Matsya-purâṇa, which mentions a Bhavishya-purâna in 14,500 verses, contains actually several sections which have been borrowed from the portions of the latter work preserved in the MSS., makes Professor Wilson's estimate improbable. For the Matsya-purâna was considered a canonical work about the year 1000 A. D., and used by Albîrûnî for his work on India. Though it, therefore, becomes probable that the Bhavishya-purâna is much older than Professor Wilson was inclined to assume, the data thus gained are much too vague for inferences regarding the age of our Manu-smriti.

Equally unsatisfactory are the results which an examination of the quotations from the Manu-smriti, found in various Sanskrit works, yields us. Perfectly indisputable quotations are not very common, and they occur mostly in works of comparatively recent date, e. g. in the Yasastilaka of the Digambara-Gaina poet Somadeva, 959 A. D.3, in Sankarâkârya's Sârîrakabhâshya, 804 A. D., and in Kshî

See e. g. his remarks on Manu XI, 101, and Nârâyana's on XI, 131.

2 Vishnu-purâna, vol. i, pp. Ixii-lxiv, and Reinaud, Mémoire sur l'Inde, P. 396.

3 Catalogue, p. 43.

I owe the knowledge of this fact also to the kindness of Professor Sachau. See Professor Peterson's Report on the Search for Sanskrit MSS., 1883-84, PP. 42-43.

• Deussen, Vedanta, p. 36. With respect to the date of Sankarâkârya's work, I follow the Hindu tradition, which places the birth of the author in 788 A.D. According to the statement of the late Vagñesvara Sâstrî, with whom I discussed the passages which he adduces in the Âryavidyâsudhâkara, p. 226, the sampra

rasvâmin's Amarakoshodghâtana 1. Other cases, where we find verses from the Manu-samhitâ quoted in ancient works, are made inconclusive by the vagueness of the reference or by the circumstance that the same passages occur also in other works. Thus we find Manu VIII, 416, with a slight verbal difference at the end of the first line 2, in the Sabarabhâshya on Mîm. Sû. VI, 1, 12. Though the exact date of the latter work is uncertain, we know that it preceded Kumârilabhatta's Tantravârttika, and its style, which closely resembles that of Patangali's Mahâbhâshya, makes it probable that its author lived not much later than the beginning of our era. Hence its testimony would be of the greatest interest, provided it were perfectly clear. Unfortunately the Bhâshya introduces the verse merely by the words evam ka smarati, and thus he records or states in the Smriti,' without specifying the author. As the doctrine of the verse which declares a wife, a son, and a slave to be incapable of holding and acquiring property is found, though expressed differently, also in the Nâradasmriti, Vivâdapada V, 39, it may be that Sabara took the passage from some other work than the Manu-smriti. Again, though Patangali in the Vyâkaranamahâbhâshya on Pâini VI, 1, 84 adduces Manu II, 120 without any variant 3, it would be extremely hazardous to conclude that he quotes from our text of Manu. For the Mahâbhârata (XIII, 104, 64-65") has exactly the same words.

dâya, referred to in his work, is that of Singeri, where also documentary evidence for its correctness is said to exist. Hence I hesitate to accept Mr. Telang's conclusions, who places Sankara in the latter half of the sixth century, Mudrârakshasa, Appendix, and Ind. Ant. vol. xiii, p. 95 seqq.

1 Aufrecht, Journal of the Germ. Or. Soc. vol. xxviii, p. 107. The date of this author, who used to be identified with the teacher of Gayâpîda of Kasmîr (779-813 A. D.), seems, according to the latest researches, more recent.

* See the edition in the Bibl. Ind. vol. i, p. 61:

a yaw निर्धनाः सर्व एव ते । यत्ते समधिगच्छन्ति यस्य ते तस्य तद्धनम् ॥ At the end of the first line Manu has त्रय एवाधनाः स्मृताः ॥

See vol. iii, p. 58 of Professor Kielhorn's edition. I may add that the same work on Pânini II, 3, 35 (vol. i, p. 457, Kielhorn) quotes another verse, the first line of which agrees with Manu IV, 151a, while the second entirely differs. In this case, too, the Mahâbhârata XIII, 104, 82 has a version closely resembling that of Manu.

More important are some allusions to the laws of Manu found in several works of considerable antiquity, and in inscriptions. Taken by themselves they would, indeed, not prove much. But considered in conjunction with the results of the three chief arguments, they certainly furnish a confirmation of the latter. The clearest case, perhaps, occurs in the Kirâtârgunîya of Bhâravi, a poet, whose fame on the evidence of the Aihole inscription was well established in 634 A.D., and who, therefore, cannot possibly have lived later than in the beginning of the sixth century, but may be considerably older. He makes (Kir. I, 9) Yudhishthira's spy say, 'He (Duryodhana), conquering the six (internal) foes, desiring to enter on the path, taught by Manu, that is difficult to tread, and casting off (all) sloth, since by day and by night he adheres to the (prescribed) division (of the royal duties), shows increased manly energy in accordance with the Nîti.' At first sight it might seem as if this passage contained nothing more than an expression of the ancient belief according to which Manu settled the duties of mankind, and among them also those of kings. But if we keep in mind the inference made unavoidable by Medhâtithi's statements regarding the ancient commentaries and by the character of the Brihaspati-smriti, it becomes more probable that Bhâravi alludes to the seventh chapter of Bhrigu's version of the Manu-smriti, which declares vinaya, humility or self-conquest, i. e. the conquest of the six internal foes, to be one of the chief qualities requisite for a king, and which carefully and minutely describes the employment of each watch of the day and the night. Other much less explicit allusions occur in the land-grants. It will suffice to adduce those found in the commencement of the Valabhî inscriptions of Dhruvasena I, Guhasena, and Dharasena II, to which I have called attention some time ago. The oldest of them is dated Samvat 207, i. e. not later than 526 A. D.2 There it is said in the description

1 See West and Bühler, Digest, p.46, and for the inscriptions, Indian Antiquary, vols. iv, p. 104; v, 28; vi, 11; vii, 67, 69, 71; viii, 302. For other passages, see Hopkins, Journal of the American Oriental Society, vol. xi, pp. 243–246.

2 This is on the supposition that the era of the Valabhî plates began in 319 A.D., the latest date ever assigned to it.

of Dronasimha, the first Mahârâga of Valabhî and the immediate predecessor of Dhruvasena I, that 'like Dharmarâga (Yudishthira) he observed as his law the rules and ordinances taught by Manu and other (sages).' Strictly interpreted, the passage says nothing more than that in Dronasimha's times various law-books existed, one and the chief of which was attributed. But, considering what we know from other sources, it is not improbable that it refers to our Samhitâ, which is acknowledged by Brihaspati as the paramount authority. This is all I am able to bring forward in order to fix the lower limit of the Manu-smriti. But the facts stated are, I think, sufficient to permit the inference that the work, such as we know it, existed in the second century A. D.

For an answer to the question whether our Manu-smriti can go back to a higher antiquity, and how much older it may be, we have at present very scant data. Its posteriority to the twelfth and thirteenth Parvans of the Mahâbhârata teaches us, as already stated, nothing definite. But there is a passage in its tenth chapter, vv. 43–44, which has been frequently supposed to convey, and probably does contain, a hint regarding its lower limit. There the Kâmbogas, Yavanas, Sakas, and Pahlavas are enumerated among the races which, originally of Kshatriya descent, were degraded to the condition of Sûdras in consequence of their neglect of the Brâhmanas1. As the Yavanas are named together with the Kâmbogas or Kâbulîs exactly in the same manner as in the edicts of Asoka 2, it is highly probable that Greek subjects of Alexander's successors, and especially the Bactrian Greeks, are meant. This point, as well as the mention of the Sakas or Scythians, would

1 The verse contains also the name of the Kînas, which formerly has been taken to be valuable as a chronological landmark. More modern researches have proved this view to be untenable; see A. von Gutschmid, Journal of the German Or. Soc. vol. xxxiv, pp. 202-208; Max Müller, India, what can it teach us? p. 131; Rig-veda, vol. iv, p. li.

2 See e.g. the fifth rock-edict, where the Yona-Kamboga-Gamdhâra or Gamdhâla are mentioned as Asoka's neighbours, the most distant being placed first. 3 The earliest mention of the Sakas probably occurs in a Vârttika of Kâtyâyana on Pân. VI, 1, 94, where sakandhu is explained by saka + andhu. According to the traditional explanation the compound means 'the well of the Saka king.'

indicate that the Slokas could in no case have been written before the third century B. C. This limit would be still further and very considerably contracted if the mention of the Pahlavas were quite above suspicion, and if the deductions of my learned friend, Professor Nöldeke1, regarding the age of this word were perfectly certain. Pahlava and its Iranian prototype Pahlav are, according to the concurrent testimony of the most distinguished Orientalists, corruptions of Parthava, the indigenous name of the Parthians 2. Relying on the fact that the change of the Iranian th to his first traceable in the name Meherdates, mentioned by Tacitus, and in the word Miiro, i. e. Mihira, on the coins of Kanishka or Kanerki3, Professor Nöldeke concludes that the form Pahlav cannot have originated among the Iranians earlier than in the first century A. D., and that it cannot have been introduced into India before the second century of our era. If this inference were unassailable, the remoter limit of the Manu-smriti would fall together with its lower one. But, with all due deference to the weight of Professor Nöldeke's name, I must confess that it appears to me very hazardous. For, first, the foundations of his theory are very narrow: secondly, one of his own facts is not quite in harmony with his assertions. However late we may place Kanishka, he cannot be later than the last quarter of the first century A. D. Kanishka was not a Parthian, and his coins probably were struck in the North of India. Hence it would appear that Iranian word-forms with the softening of th to h were known in India towards the end of the first century. Moreover, the word Pahlava occurs in the Girnâr inscription of Rudradâman 1, which was incised shortly before the year 72 of the era of the Western Kshatrapas. This era, as has been long ago conjectured, and is now incontestably proved by Mr. Fleet's important discoveries, is

1 Weber, History of Indian Literature, pp. 187-8, note 201a.

2 Olshausen, Parthava und Pahlav, Mâda und Mâh (Monatsberichte der Berliner Akademie, 1877), and Nöldeke, Journal of the German Oriental Society, vol. xxxi, p. 557、

3 Sallet, Die Nachfalger Alexanders des Gr. p. 197.

Ind. Ant. vol. vii, p. 261. Rudradâman's lieutenant at Girnâr was the Pahlava Kulaipa (Khoraib ?), son of Suvisâkha.