being intercalary letters)-were once polyandrists, and are still regardless of female chastity, whilst the Tibetans were and are notoriously both.

Mr. Metz, on the subject of the dialectic differences of the Nilgirian tongues, observes:

"The differences of the several languages of the hill tribes consist, not so much in idiom as in mere pronunciation. But that is so great that the same or nearly the same word in the mouth of a Toda, with his pectoral pronunciation, can scarcely be recognised as the same in the mouth of a Kota, with his dental pronunciation. The Badaga and Kurumba dialects are midway between the former two with regard to pronunciation, only the Badaga is a little more guttural than the Kurumba. "There is some difference even in the speech of the several branches, or remotely located groups, of any one tribe. For instance, those of the Badaga tribe who, like the Kangaru or Lingaits, emigrated from Targuru and came to the hills at a later period than the others, speak a purer Canarese than the common Badagas. So also the Todas among themselves have differences of pronunciation according to the different districts they inhabit; for instance, some pronounce the s quite pure, others like z, and others again like the English th. And in like manner the Kurumbas round the slopes of the hills have so many little variations in their speech according to the situation of their villages (Motta) on the south, east, or west side of the hills, that it is difficult to say what the real Kurumba tongue is. In Malli, the chief Kurumba place on the south slope, the language is much mixed with Tamil."

I will now conclude with a few remarks on the grammatical traits exhibited by the subjoined papers.

PHONOLOGY.

As much as is forthcoming on this head has been expressed in the vocabular part of this paper and the remarks appended to it. It may be advisable, however, to repeat here that the presence of the English th, and its frequent substitution for s and z, and the equivalence of the two latter, are so far from

being exclusively Toda, as Schmid supposed, that they are common in Indo-China, Himálaya, and Tibet. Tibetan abounds in sibilants, having, besides the s, ch series, an equivalent z, zy, dz series. The former is possibly borrowed. At all events, z, zy, dz, and ts, tch are very much commoner in use than the Arian s, ch, series. The second z, represented by me by zy, and equal to the French j in jeu, is the same with the Tamil zh of Ellis and Elliot. It is a very prevalent sound, and equally prevalent is the French u, or eu in jeu aforesaid. Neither is ever heard from an Arian mouth; but the Himalayans most infected with Arian ways and habits are now gradually substituting Arian j for their own z, and Arian u for their own eu. D is also taking the place of their hard and aspirated z (dz and zh), and thus the Tibetan word zhí-ká-tsén and Newári Zhi-khá-chhén, the name of the capital of Tsáng, has become Digarché with those who use the popular and spreading Khas language, which language we hereby perceive also preferring sonants to surds (g for k), whereas the written Tibetan and Newári, like the Tamil and Toda, having a preference for surds.

But Tibetan is spoken with all the variety of hard and soft pronunciation noticed by Mr. Metz as characterising spoken Toda and indeed the whole of the Nilgiri dialects; and as there are few things more normally Turánian than the wide extent of legitimate, habitual commutability between the consonants and between the vowels also of the languages of the family, so I consider that to lay so much stress as is often

The etymology of this word is curious and important with reference to the evident identity of the term Tibetan. And it is hardly too much to say that the family identity of the two tongues (Newári and Tibetan) might be rested on it.

It means in Newári "the four-housed," zhi or zyi being four, khá the generic sign for houses, and chhén being house. De Körös has said nothing about that most fundamental sign of the Turánian tongues, the generic or segregative signs; but I have good reason to assume that this is one of the several serious defects of his grammar, and that Tibetan ká is = Newári klá, as zhi = zhi, and tsén = chén, though khyim be now the commoner form of the word in written Tibetan. Zhi-khá-chhén or Zhi-ká-tsến Turanice = Digarchén Arianice, is the name of the capital of Tsáng-why styled "the four-housed" I cannot learn. But three such elements, composing one word identical in form and in sense in two separate languages, involve the family oneness of these languages.

done on merely phonetic peculiarities is a great mistake on the part of Turánian ethnologists, and one apt to lead them much astray when in search of ethnic affinities. For example, the Myamma is questionless one language, notwithstanding that its phonetic peculiarities in Ava and in Arrakan are very marked; and a particular friend of mine, who is "genuinely Saxon, by the soul of Hengist," can by no means deal fairly by r, sh, or th, but calls hash has; shoes, soes or toes or thoes; brilliant, bwilliant; there, dere; thought, tought, &c.* A Londoner is not less Saxon, surely, because he is wont to "wow that weal, wine, and winegar are wery good wittals."

ARTICLE.

Mr. Metz says there is none whatever, but I feel pretty sure that the usual equivalents are recognised, viz., the numeral one, or the indefinite pronoun some, any, in lieu of the indefinite article; and the demonstratives in lieu of the definite, as also the segregatives van, val, and du, or an, al, and ad, for the three genders, or ál and pé for the major of gender, used as suffixes, and widely applicable to nouns (qualitives)—less widely and uniformly to verbs. We should always remember that the so-called segregatives or generic signs are essentially articles, definite or indefinite according to the context.

ADJECTIVES.

All qualitives which seem to embrace, as usual, the nominal (genitive), pronominal, participial, numeral, and adjectival, appear to be used both substantivally and adjectivally, and, when employed in the former way, to add to their crude, as a suffix, the appropriate generic sign, which, in the case of the participle, gives it a relative sense or an agentive, just as in English, the or a striker, or the or a striking person (or thing), and the or a hard thing, arc equivalent respectively to the person who strikes and the thing which is hard. But the latter form of speech is quite Anti-Turánian.

• "Three fresh fishes in the dishes" is, in the mouth of the same friend, "Tree fes fises in the dises."

=

Qualitives are always prefixed when not used affirmatively or substantivally. If placed after the noun they become affirmative, including in their sense the substantive verb. Man (is) mortal. That (is) mine. This the striker = this is the person who strikes.* He (is) loving one or lover = one who loves. That one (is) the black = that is the black one. Give me the black the black being or thing—a difference which must be expressed, and with the sign of gender, too (an, al), in the former event. This person two person this one is the second person (rend-al),† &c. Gender is fully marked in qualitives by the use of the suffixes van, val, du, or an, al, ad = hic, hæc, hoc. But these forms are very imperfectly reproduced in the verb, indeed can hardly be traced except in Badaga and Kurumba, where the following is unmistakable evidence of them.

English.

He strikes

She strikes

It strikes

=

The major and minor of gender in beings, not things, seem to be denoted by ál and pé suffixes-words having still the independent signification of man and woman. In Toda, moreover, adum marks the common gender as a separate pronoun, and tan as a conjunct prefix. I am not sure as to the major and mninor of gender, because the verb does not exhibit them in the peculiar manner of the cultivated Dravidian tongues or otherwise.

* In Newári it would be, ú-hma daya-hma, which is in every particular of idiom Dravidian, hma being the van or ál suffix of the above tongues, and its affixing to the verbal form rendering that a relative participle.

+ Here final al is not the contracted sign of the feminine suffix aval, but is the name for inan used as a suffix.

The prefix ta, with or without the nasalisation tan, tang, and with or without the causulate equivalent ka vel ga, is widely prevalent to the north and south, as I have noticed in a recent paper; and so also the other equivalent a vel e, witness ta-pe, ka-pa, ta-ga-pa-n, a-pa-e-ri, g-ri, ta-g-ri, tan-d-ri, a-yi, ta-yi, tan-g-yo, for man and woman in Gyarung, Kassia, Kiránti, Bódó, Kúcch, Tamil, Lepcha, Uraon, &c. Those who deny family connection between the Himalayan and Dravidian tongues are requested to pause over ta-g-ri (Lepcha), and tan-d-ri (Telugu), for man, and a-yi vel ta-ye, in both tongues, for woman-roots, ri and yi, vel i.

NOUN.

The papers furnish no sample of declension, but it may be safely inferred that it is simply postpositional with cases ad libitum, or none at all, according to the view taken of declension. Gender is marked either by separate words, such as man, woman; cock, hen; or by sexual prefixes like our he-goat and she-goat; or, lastly, the generic word bears also a male or female sense, when the feminine or masculine gender, as the case may be, is distinguished by the fitting sign prefixed. So Burmese sa means child and boy, and mí-sá, or female child, means girl.* I know not whether the suffixes van, val, and du, or ál and pé (pen, pem-the latter equal major of gender), are added to substantives as well as to qualitives, but I think not. Instances occur in Telegu, but not generally in the Dravidian tongues, nor in the northern.

The major and minor of gender (quasi, hic et hæc facilis; hoc, facile) are common in the Himalaya, Indo-China, and Tibet, but I have nowhere in the north found the fullydeveloped masculine, feminine, and neuter of the south.

In regard to number, the Nilgirian nouns are very defective, having no distinct and uniformly employed dual or plural inflexion or sign. But they seem to follow the cultivated Dravidian in so far as having no dual, but having the double, or exclusive and inclusive, plural at least in the separate pronouns and in the personal endings of the verb. Irula has not even the latter. In the Himalayan tongues it is often difficult to make out disjunct dual and plural forms of the substantive, even when the distinct and conjunct pronouns exhibit an exclusive and inclusive form both of the dual and of the plural of the first person, with correspondent verb forms as is the case in the Kiránti language. The source of the defective plural sign of nouns is to be sought in the fact that Turánian vocables generally, in their crude state, bear the largest and specific or generic meaning-a peculiarity well exemplified by the English word sheep. In the Nilgiri tongues neuter nouns

The mi is often suffixed. Thus ta and ta-wa, a child, is tu-mi, a girl, in Hayu and Kiránti.