sign is constantly confounded with the temporal sign, whilst the personal endings likewise sometimes exhibit as much irregularity and defectiveness as they do in the Nilgirian verbs. Nevertheless, judging by analogy, and resting on the wonderful similarity of genius and character pervading all the languages of the sons of Túr, I should not hesitate to say that the cultivated Dravidian and the Nilgirian tongues are framed on the same model as that above described as belonging to the northern, and that the samples above cited from Badaga and Kurumba are palpable proofs of it, notwithstanding the silence of all Dravidian grammarians touching the generic or class sign (transitive, intransitive, &c.) of their verbs. For example:I have no doubt whatever about Badaga Kurumba Kurumba May be analysed precisely as are Active voice. Turkic sever-d-im I loved (him) Hungarian var-t-am I waited for (him) Kiránti (Báhing) tip-t-ong I struck (him) Háyu top-t-um Háyu *há-t-um struck himn gave him Kuswar tha-tha-im ik-an }I I struck (him) Active. and numberless others of which I shall have, ere long, to speak in full. That is to say, I hold it for certain that all these verbal forms consist of, 1st, the root or crude; 2d, the transitive and preterite sign; 3d, the personal ending; and that, moreover, the second of these elements may, in every case, be *Hátum is active and passive in Hayu, and is regularly derived from the imperative transitive ha-t-o, give to him or give it, which is common to Khámti and Háyu; and this leads me to add that the so-called monosyllabic tongues, like the simplest Himalayan ones, and the Tibetan and Burmese, exhibit in their imperatives the compound structure instanced in háto, e.g., shat shod = kill, i.e., kill him or it, in Lepcha and Burmese, where final t vel d is the well-known objective pronoun seen in all the above samples taken from the highest-structured class. Newári has sháta for the preterite second and third persons active and for all persons passive; expressly because the "t" denotes the object or transitiveness of the action. So also Háyu has si-t- in the same sense, and si (sh) to in its imperative, which is modified by an enunciative sibilant, but shows the transitive "t" as before. resolved into the third pronoun, current or obsolete, and used objectively. Kuswar baba-ik = his father, compared with thatha-ik = strike (i.e., him, the object), settles the last point even more clearly than Samoiede lata-da= his stick, and Magyar Cicero-t Ciceronem.* = Having mentioned the wonderful analogy of these tongues, I will give a telling instance. In the Hayu language of the Central Himálaya and in the Mantchú we have khwachambi or khwachammi = I feed, that is to say, feed myself; for khwa, vel khóa is the root, chá the reflex sign, and mbi vel mmi the personal ending, and one, too, that in both tongues is invariable, though Hayu appears sometimes to drop the iteration in the second and third persons, khwachammi, khwá-chá-m, khwáchá-m, I, thou, he, feed (self). Now, that root, reflex sign, and personal ending should thus concur to absolute identity, and that sense also should be as identical as form in two unconnected languages, is simply impossible. It follows, therefore, that we have people of the Mantchu race forthcoming now in the Central Himálaya close on the verge of the plains! And, again, what shall we say to such grammatical coincidences as The answer is clear, that we have people of the Turkic stem also in the Central Himálaya, close to the verge of the plains of India. Nor need we doubt that such is the case in regard both to the Mantchúric and Turkic relations of the Himalayans, though the precise degree of such family connections can hardly become demonstrable until we have (what is now, alas! wholly wanting) a just definition of the Turánian family, and of its several sub-families, to test our Himálayan analogies by. The Mantchuric and Mongolic groups of tongues were long alleged to show no sign of pronomenalisation. It is now known that that was a mistake. Other still maintained distinctions will, I anticipate, disappear before the light of fuller knowledge, when it will plainly appear that not mere and recent neighbours, such as * Müller apud Bunsen, I. 319. * are alleged to be the Tibetans proper of our day (Bodpas), or they and the Ugrians, formed the Turánian element of Indian population, from the Himálaya to the Carnatic, but successive swarms from the one and same great northern hive-whether Turkic, Mongolic, Mantchuric, Ugric, or these and otherswho passed into Indo-China as well as India, and directly into the latter, as well as through the former into the latter, by all the hundred gates of the Himalaya and its southern offshoots. Simple as the Mongolic and Mantchuric languages are wont to be called, they seem to me to possess entirely the essential Turánian characteristics; that is, in like manner as they have endless noun-relational marks without any distinct declension, so they have a rich variety of sorts of verb (but all reduceable into the two great classes of action, or that of things and that of beings, equal neuter and transitive), and this peculiar richness united with great poverty of voice, mood, and tense, whilst the participles partake fully of this character of the noun and of the verb; that is, they are poor on one side but luxuriant on the other, and throughout the whole Turánian area perform the very same function or that of continuatives, being employed to supply the place of conjunctions and conjunctive (relative) pronouns. The Central Himálayan languages, but perhaps more especially those of the pronomenalised type, all present these characteristics with perfect general fidelity and with some instances of minute accord, besides those cited above, among which may be mentioned the hyper-luxuriant participial growth of Kiránti and of Mantchu, both of which have ten or rather eleven forms of the gerund, and these obtained by the very same grammatical expedient! There is another very noticeable peculiarity common to the Himalayan and Nilgirian tongues, which is the emphatic distinction of the first person in conjugation, thus, piuthtstini, Toda, I strike, stands apart from puithtsti, thou, he, she, or it strikes, which are all the same. So Newári has daya in the present and dayu in the past for I strike, I struck, as opposed * Are not Ugric, Uighur, or Igur, the same? and would not the identical name with the common characteristics (pronomenalised) of the tongues go far to identify the Ugrians with the E. Turks? to the common terminations yu and la respectively for all striking present and past of every other kind save by the first person, da-yu, da-la, any body or thing save me strikes or struck. Hence these forms are used to constitute the passive, as in jita dála, of the sequel. Again, the hardening or doubling of the sign consonant of the intransitive verb in order to make it transitive, a principle supposed to be so peculiarly Dravidian, is quite familiar to the Hayu and Kiránti tongues. And again, the Báhing dialect of Kiránti is fully characterised by that indiscriminate use of the transitive and neuter signs for which the Tamil language is so remarkable. Another common characteristic of the Dravidian and Himálayan tongues is the double causal, e.g., bokko get up; pokko cause to get up; pongpato cause to cause to get up-in Báhing. Dun become; thun to cause to become; thumpingko = cause to cause to become = -in Váyu. = = Another common and radical feature of the Dravidian and Himalayan tongues is the amorphous character of their vocables, which become distinct parts of speech, as noun or verb, by the suffixing of appropriate particles. Thus kan, the eye, and to see; so neu, goodness, to be good, good, whence neu-gna, I am good; neu-ba, the good one, &c.—of Báhing. I, however, at present forbear to touch on more of these common characteristics of the Dravidian and Himálayan tongues, because they are so apt to run into the common property of all the Turánian tongues. But I may just add that Hoisington's Tamulian traits (in the "American Or. Journal") are nearly all found characterizing the northern languages. The general absence of a passive, the partial or total absence of tense distinctions, and the combination of the present and future when there is such partial distinction, as well as the denoting of tense by annexed adverbs (to-day, yesterday, and to-morrow) when there is none, are Turánian traits common to the (not to go further) Altaic, Himálayan, Indo-Chinese, and Tamulian tongues. Thus the Toda and Kota verbs are always or generally aoristic, and the three tenses are expressed by the above adverbs of time, used prefixually. Precisely such is the case with the Bontáwa dialect of Kiránti and with the Hayu, whilst the Báhing dialect of Kiránti discriminates the past tense from the other two by the use of an appropriate infix, which is at once the transitive and temporal sign. If such be not visibly the case with the Badaga, Kurumba, and Irula dialects, we may yet discern the cause, partly in the carelessness of barbarians, partly in that fusion of transitive and preterite signs which cultivated Dravidian also exhibits, and not less Ugrofinnic and Turkic. But in the tin-d-é of Badaga and Kurumba, and tid-d-é of Kota I ate, as in the mad-id-é of Kurumba I made, not to cite more instances, I perceive that identical preterite sign (t, vel, d) which marks it in Báhing (tib-á, he strikes; tib-d-á, or tip-t-á, he struck), as in endless other northern and north-western tongues. = = I will add a few more words on these important points, for I conceive that the passive of the cultivated Dravidian tongues is clearly factitious, and suggested by contact with Arianism. There are still extant long works in Canarese, says Mr. Metz, in which hardly one instance of the use of the passive voice occurs, and the fact that the uncultivated Dravidian tongues have it not, is, I think, decisive as to its adopted character in the cultivated. Again, there can be no doubt that the negative conjugation of the cultivated Dravidian tongues presents the primitive form, and that form is aoristic; e.g., mad-en, I do, did, or will, not make. In Himálaya and Tibet and Sifán the passive is wanting. Its absence is wholly or partially supplied by the use of the instrumentive and objective cases of the pronouns for the active and passive forms respectively. Even Khas still adheres to this primitive and indigenous form, overlaid as that tongue is by Arian forms and vocables; and I have myself not the least doubt that the anomalous né of the preterite of Hindi and Urdu is nothing but the commutative equivalent of the Khas instrumental sign lé. A Khas of Népál invariably says, by me struck for I struck, and me struck for I was struck; and, moreover, there is still the strongest presumptive proof, internal and external, that this, the present preterite, was a primitive aorist, and the only tense in Khas. Those who are fully conversant with the spoken Prákrits of the plains can testify that the same traits still cleave to the |