the Province of Bengal. They cultivate the steep sides of hills in settlements of their own, of about twenty houses, but intermixed with other tribes. They repudiate all connection with any other tribe, and maintain, that their Language is separate from all others, and that they are the earliest human beings of the locality. They are a most frimitive people in habits and customs. They had till lately no knowledge of iron. They neither spin nor weave, nor have the least knowledge of pottery. They practise the Jhum system of agriculture. The women used not to wear a particle of clothing, but bunches of leaves before and behind, hanging to a girdle of beads. They were deterred by superstition from wearing clothes, and believed that, if they did, they would be devoured by tigers. Hunter mentions that within the last few years a large supply of cloth has been distributed by the State, and engagements taken from the men, that the women should henceforth wear clothing. There is nothing but Vocabularies of their Language. They are Pagans, and have no Character. Neither tribe nor Language are likely to survive.

KORWA.

In the centre of a dependency of Chútia-Nagpore, called Barwah, live the wretched forest-cultivators, the Korwas, separated from the cognate Kolarians, and intermixed with other tribes; but it is admitted, that they are the earliest settlers, and were once masters of the country. Their number does not exceed fourteen thousand, and they lead a savage and nomadic life. They are Pagans, wholly illiterate. Nothing exists but Vocabularies of this Language, and a long duration cannot be expected for it.

KUR.

The Kur and Kurka dwell in the Central Provinces, on the Mahadeo Hills, and Westward in the forests on the Rivers Tapti and Narbudda, up to the Bhil country. On the Mahadeo Hills they prefer to be called Muasi. They are

Pagans, and, though residing amidst Gonds, their Language is Kolarian. Vocabularies are supplied by Hislop and Dalton. They are wholly illiterate. In the Districts of Hoshungabad and Bétul their number exceeds fifty-nine thousand. This Language will scarcely survive very long. SAVÁRA.

In the Madras Province, and surrounded by Aryan and Dravidian neighbours, we come upon another small Kolarian tribe, speaking a distinct Language. They are known as Savára, or Sabara, or Sowruh, and supposed to be the Suari mentioned by Pliny and the Sabaro mentioned by Ptolemy. They are found on the West and back of the Mahendra mountain in the Ganjam district of the Madras Province, and their Language-Field is duly marked off on the Language-Map of the Census Report of the Madras Province. They are said to number about eighteen hundred, to dress in leaves, though they have picked up a little civilisation from their Uriya and Telugu neighbours. They are quiet and industrious, and dwell in villages. There are some still wilder members of the Family in the hills. Of their Language little is known beyond Vocabularies. They are Pagans. Dalton, in his Ethnology of Bengal, mentions this tribe by name as occupying the country betwixt the Khond Hills and the Godavery, and retaining a primitive speech; but he adds, that the Bendkar Savaras speak Uriya, and conform to the customs of Hindus of the lower castes, and dwell in the State of Keonjhur dependent on the Cuttack district of the Bengal Province. It may be expected that this Language will be crowded out by Telugu or Uriya. In the Madras Census Report, the Sowruh of the Jypore district in the Madras Province are described as semi-Hinduised, and have forgotten all knowledge of any Language but Uriya.

MEHTO.

The Ethnological Committee of the Central Provinces indicate a tribe called Mangee, or Mehto, in the hilly

tracts of Belaspur, who are Pagans, and who, judging from the short Vocabulary supplied, speak a Kolarian Language.

GADABA.

The tribe of Gudba, or Gadaba, inhabit the Eastern portion of Bustar in the Central Provinces, and Jypore, a dependent State of the Madras Province, where they are numerous. Glasford notices them. Their Language clearly belongs not to the same stock as their neighbours the Gonds, of the Maria Dialect, but to the Kolarian Family. It is interesting to find a Kolarian Language imbedded among the Dravidians down in the South-East. Glasford, in his report of the Bustar District, supplies a Vocabulary. Some of the words are identical with words of the Koorku, Kol, and Sonthal Languages. This same tribe is found again in the highlands of Guddapur, of the Ganjam District of the Madras Province. They are Pagans. We cannot anticipate a long life to this Language. In the Madras Census Report they are connected with another tribe called Kerang-Kapus, who speak the same Language.

MAL-PUIIÁRIA.

Dalton mentions the existence in the Ramgurh Hills of the Birbhum District of the Bengal Province of a tribe, who call themselves Mal-Puhárias, but who are altogether different from the Rajmuháli Puhárias, or Malers, of the Dravidian Family. A Vocabulary was collected by Coates from a prisoner in the gaol, but the words seem to be as far removed from Kolarian as Dravidian. They are Pagan, and have their separate customs. It was necessary to enter the Language somewhere, that it might not be overlooked. The classification is entirely provisional. They are also called Naia-Dumka. This Language, and that of the Mehtos, is entered for the purpose of exhausting the subject, and accounting for all the Languages, of which Vocabularies have been supplied.

CHAPTER V.

TIBETO-BURMAN FAMILY.

I APPROACH the Tibeto-Burman Family with some misgivings, for the Field is imperfectly explored, it is unusually extensive, and the classification is new, and I have no authority to follow, as in the three preceding Families. Little has been done in the way of classifying and arranging since the date of Max Müller's letter to Bunsen on the Turanian Languages a quarter of a century ago, yet in some parts of the Field our geographical, ethnical, and linguistic knowledge has so extended, that a reprint of that letter would do more harm than good. It is my present task to indicate, what has been done, and what remains to be done, and I see signs that something more will soon be done.

The interior grouping of the members of this enormous Family must for the present be based upon geographical considerations, and upon no other. It extends from the River Indus and the frontier of Dardistan, already described in the Aryan Family, in a South-Easterly direction to the River Mekong and the Isthmus of Kraw, in Siam. It embraces the whole length of the Himalaya range and the kingdom of Tibet, and portion of Yunan in China beyond. It is admitted, that there is a linguistic affinity connecting seven groups out of the eight, which make up this Family. The old phrase of Hodgson, "Tamulic," must be abandoned, as based on an error admitted by that scholar; the term Turanian is decidedly objectionable, as implying too much; the pro

posed subdivision of Max Müller into Gangetic, and Lohitic, would at best only apply to part of the Field, and is inappropriate. We must fall back upon a compound name, formed from the two leading Languages of the Northern and Southern branches of the Family. It is a positive fact, that Tibetan and Burmese are the only two great literary and political Languages of the Family.

It will be observed that in the Himalaya range within the Province of the Punjab and the North-West Province, the Aryans seem to have pushed the Tibeto-Burmans across the great Watershed, and the Languages of this Family to the West of the River Gogra or boundary of Nepál are all Trans-Himalayan. It is more convenient to exhaust the groups this side of the Himalayan Watershed first. The following groups are suggested, as a convenient mode of grappling with the subject:—

I. NEPÁL GROUP, II. SIKHIM Group,

III. ASSAM Group,

13 Languages. 16 Dialects.

I Language. I Dialect. 16 Languages. 23 Dialects.

IV. MUNIPUR-Chittagong GROUP, 24 Languages.

V. BURMA GROUP,

VI. TRANS-HIMALAYAN GROUP,

VII. CHINA Group,

VIII. ISLAND Group,

8 Dialects. 9 Languages. 10 Dialects. 8 Languages. 23 Dialects. 6 Languages. None. Dialects.

3

Total,

87 Languages. 84 Dialects.

This seems the only way of bringing this enormous Language-Field, comprising eighty-seven Languages, and eighty-four Dialects of those Languages, under review. The arrangement is entirely provisional. It is not pretended, that the list of Languages is exhausted, that Vocabularies exist for all the entries, or that the habitat of every tribe is indicated. In another particular there is extreme laxitude. Many names have been entered as Languages, which closer scrutiny may reduce to the rank of Dialects of other Languages. On the other hand, many, now entered as Dialects, subordinate to some