These races, who are called An-ārya, 'ignoble,' in opposition to Ārya, 'noble,' had been gradually driven southwards or towards the hills by the Aryan settlers. They probably made great resistance in the North at the time the Rig-veda was composed. They are there called Dasyus, Yatudhānas, &c., and described as monstrous in form, godless, inhuman, haters of Brahmans, disturbers of sacred rites, eaters of human and horse flesh (Rig-veda X. 87, 16; Muir's Texts II. 435). In the epic poems they are generally called Rakshasas or evil demons, the relentless enemies of gods and good men and of all sacred rites1.

of the more southern hill-tribes are partially connected with the Drăvidian, especially the Tuda, Kota (two dialects of the Nil-giri hills), Gond, and Khond (Ku). The Ramūsies and most of the Korawars speak a patois of Telugu. The Male-arasars (hill-kings') of the Southern Ghats speak partly corrupt Malayalam and corrupt Tamil. The Lambadies, or gipsies, speak a dialect of Hindustani. Among the barbarous. tribes of the South are included the Vedārs of the forests of Ceylon.

1 In one place (Rāmāyaṇa III. i. 15) they are described as black, with woolly hair and thick lips. The following is from III. i. 22, &c. : 'Mendevouring Rakshasas of various shapes and wild-beasts dwell in this vast forest. They harass the devotees in the settlements. These shapeless and ill-looking monsters testify their abominable character by various cruel and terrific displays of it. These base-born wretches (an-ārya) perpetrate the greatest outrages. Changing their shapes and hiding in the thickets they delight in terrifying devotees. They cast away the sacrificial ladles and vessels (śrug-bhandam), pollute the cooked oblations, and defile the offerings with blood. They utter frightful sounds in the ears of the faithful.' Virādha, a Rākshasa, is said (Rāmāyaṇa III. vii. 5; Muir II. 427) to be like a mountain-peak, with long legs, a huge body, a crooked nose, hideous eyes, a long face, pendent belly, &c., like Death with an open mouth.' The Nishādas of the Purāņas, though described as dwarfish, have similar features, and are no doubt intended for the same race. In the same way, in describing races unknown to the Greeks, such as the Cyclopes, Laestrygones, Centauri, &c., Homer and other Grecian writers are given to exaggeration, and relate the most absurd fables.

arose.

It is to the subjugation of these non-Aryan races by heroic Āryan leaders who were Kshatriyas, as well as to the rivalry between different tribes of the settlers themselves, that we owe the circumstances out of which the two great Epics Whether the celebrated Aryan warriors of the Rāmāyaṇa and Maha-bhārata were identical with those of the Itihasas of which mention is made in the Grihya Sūtras and in Manu (III. 232) cannot be proved; but this much is clear, that the exploits of the three Rāmas, Arjuna, &c., became, soon after Manu's time, the theme of song, and that these heroes were in the first instance represented as merely men of great strength and prowess, whose powers, however extraordinary, were not more than human. The oral descriptions of their deeds and adventures by public reciters formed the original basis of the two great Epics, and were naturally the peculiar property of the Kshatriya and conquering class. Probably these narratives were in the first instance delivered in prose, which became gradually interspersed with the simplest forms of metre, such as that called Anushṭubh or Śloka1.

It is easy indeed for the most cursory reader of the Rāmāyaṇa and Maha-bharata to trace a substratum or basis (mula) of simple heroic narration underlying the mass of more recent accretions. But to what date is this first frame-work of the poems to be referred? And again-When occurred that first process of brahmanizing which obscured and transformed its original character? And lastly-When was the structure completed and the

1 The oldest part of the Maha-bhārata has a section entirely in prose (see note 1, p. 372). The invention of the Sloka is attributed to Valmiki, the reputed author of the Rāmāyaṇa, with the object doubtless of establishing his claims to be regarded as one of the earliest and most ancient of Indian poets. This metre is found in the Veda.

whole work moulded into a form similar to that we now possess?

With regard to the first of these questions, I have now to submit five reasons in support of the view that the earliest or pre-brahmanical composition of both Epics took place at a period not later than the fifth century B. C., as follow:

1. The Rāmāyaṇa records no case of Sati. In the Maha-bhārata, Mādrī, wife of Pandu, is made to immolate herself with her husband 1, and the four wives of Vasu-deva and some of Krishna's wives do the same; but it is remarkable that none of the numerous widows of the slain heroes are represented as burning themselves in the same manner. This shows that the practice of Sati was beginning to be introduced in the North-west of India near the Panjab (where we know it prevailed about 300 years B. C.), but that it had not at the time of the earliest composition of the Rāmāyaṇa reached the more eastern districts. But if one Epic records no Sati, and the other only rare cases-notwithstanding the numerous opportunities for referring to the practice afforded by the circumstances of the plot-it follows that we ought to place the laying down of the first lines of both compositions before the third century B. C., when we know from Megasthenes that it prevailed generally even as far east as Magadha.

2. The first construction, or, so to speak, 'first casting' of the stories of Rama and of the Pandavas as poems with definite plots, seems to have been pre-buddhistic quite as clearly as it was pre-brahmanical-by which I mean, that it took place anterior to the actual establishment of Buddhism as a rival system. Only one direct mention of Buddha and Buddhism occurs in the Rāmāyaṇa, and the verses in which it occurs (II. cviii. 30-38), and in which Buddha is compared to a thief, are admitted to be an interpolation and not part of the original poem. Nor can it be proved that any such direct reference occurs in the original Mahābharata. Nevertheless, there are numerous allusions (not bearing the stamp of later additions) in both Epics, especially the latter, to that development of rationalistic inquiry and Buddhistic scepticism, which we know commenced about 500 years B. C. 3

2 Mausala-parvan 194, 249.

1 Ādi-parvan 4896. See also 3030. * Note particularly the infidel doctrines expressed by the Brahman Jāvāli (see p. 353), and Book I. 12. of the Bengali recension of the Rāmāyaṇa, where Sramaņas, or Buddhist mendicants, are mentioned (see also p. 133).

3. It is evident from the Aśoka inscriptions that the language of the mass of the people in Hindustan in the third century B. C. was not pure Sanskrit. It consisted rather of a variety of provincial Sanskritic dialects, to which the general name of Prākṛit is applied. If, then, the first redaction of these popular poems had taken place as late as the third century, is it likely that some forms of Prakrit would not have been introduced into the dialogues and allowed to remain there, as we find has been done in the dramas, the oldest of which-the Mriććhakaṭikā— can scarcely be much later than the second century B. C.? It is true that the language of the original story of both Epics, as traceable in the present texts, is generally simple Sanskrit, and by no means elaborate or artificial; but this is just what might have been understood by the majority of the people about five centuries B. C., before the language of the people had become generally prakriticized.

4. When the story of the poems was first put together in a continuous form, it is clear that the Dekhan and more westerly and southerly regions of India had not been occupied by the Aryans. But we know from the Asoka inscriptions that the empire of the kings of Magadha and Palibothra in the third century radiated in all directions, as inscriptions are found in the Panjab, at Delhi, in Kuttack, and as far west as Gujarāt.

5. The Greek writer, Dion Chrysostomos, who was born about the middle of the first century, and was especially honoured by the emperor Trajan, mentions (Or. LIII. 555) that records existed in his time of epic poems, recited by the Hindus, which had been copied or translated from Homer. These statements, as Professor Lassen has shown (Ind. Alt. III. 346), must have been taken from the accounts of Megasthenes, who lived at the court of Candra-gupta (see note p. 231). They indicate that poems resembling the Iliad were current in India at least as early as the third or fourth century B. C., though it by no means follows that the Hindu poets borrowed a single idea from Homer1.

1 The passage in Dion Chrysostomos is as follows: 'Oпóтe кaì naр' 'Irdois ᾄδεσθαι φασὶ τὴν ̔Ομήρου ποίησιν, μεταβαλόντων αὐτὴν εἰς τὴν σφετέραν διάλεκτόν Te Kai pwvŋv (Reiske's Edit. p. 253). There seems too great a disposition among European scholars to regard the Hindus as destitute of all originality. I cannot but agree with Professor Lassen that Megasthenes was mistaken, though obviously the story of the great war between the rival tribes, and that of the carrying off of Sita by a South-Indian chief, have, of course, points of resemblance to the Iliad, which may have suggested the idea of plagiarism. The sufferings of king Dhrita-rashtra are like those of Priam, and the lamentations of the wives of the slain heroes after the

These points seem to merit consideration in fixing 500 B. C. as an approximate date for the first or prebrahmanical and pre-buddhistic versions of the two poems. The names of the authors of these original versions appear to have perished, unless it be held (which seems highly improbable) that the story of Rama must be assigned to Vālmīki from its very first existence as a Kāvya.

We come next to the second stage of their construction. We have suggested the fifth century B. C. as the probable date of the rise of Brahmanism, as depicted in Manu (see p. 215), and with it of Buddhistic scepticism. The ambitious Brahmans who aimed at religious and intellectual supremacy, gradually saw the policy of converting the great national Epics, which they could not suppress, into instruments for moulding the popular mind in accordance with their own pattern. Possibly, too, they may have hoped to turn them into important engines for arresting the progress of Buddhistic rationalism. Accordingly, I conjecture that in the fourth century B.C. they commenced re-constructing and remodelling the two great Epics. They proceeded, in short, to brāhmanize what was before the property of the Kshatriya or warrior caste. This process was of course committed to poets who were Brahmans, and was not completed all at once. Those songs which described too plainly the independence of the military caste, were modified, obscured by allegory, and rendered improbable by monstrous fable and mythological embellishments. Any circumstance which appeared opposed to the Brahmanical system, was speciously ex

battles between the Pandavas and Kauravas are like those of Hecuba and Andromache, while the martial deeds of Arjuna and Duryodhana resemble those of Achilles and Hector. According to Professor Weber the passage in Dion contains the earliest notice by other writers of the Indian epic poems. He is, moreover, of opinion that the Indian poets really took ideas from Homer.