-as -ea, -e, -i). This postfix is clearly Dravirian glossarially and idiomatically. It is not found in the Ultraindian systems, save in some varieties of the common terms which have a great appearance of being contractions of the Kol full forms, and thus support the opinion that the latter were the original. The system is based on definitives or demonstratives like the Dravirian and all the other archaic Aso-African systems, and as the same definitives were common to most of the archaic formations, the Kol terms, like the S. Dravirian, present many resemblances to foreign numerals. These will be found in Appendix A to the next chapter. In this place I will enquire how far the Kol terms are related to the South Dravirian.

The root for 1 is mo, which contracts to m by the elision of the Vowel before the vowel of the postfix (mo-i, m-ia, m-ea, m-ia-d, m-i-dh,† m-i. The Ultraindian and Peninsular forms preserve the labial vowel (o, u,) and in some cases remnants of the Kol postfixes; po Angami Naga, bo Karen, muo, ma-i Mon, mo-e Kambojan, Ka, Chong, mo-t Anam, mu-i Binua. The root, as we have seen, is preserved with the labials m, b, v, p, and with the

a remnant of the modified south Gangetic or Bengal division of the ancient Dravi. rian race. Their dialectic peculiarities of a secondary kind must have been of much later origin than the first spread of Dravirian to the eastward, for the early forms of the pronouns found in Australia are the pure Dravirian. The numerals cannot be explained as a mere dialectic variation of the South Dravirian,—but it is to be remarked that the South Dravirian developed numeral system itself has no claim to stand on the same archaic footing as the pronouns. The quinary system was not in existence when the formation first spread with its proper pronouns, into Asonesia. The general character of Kol shows that the language must have existed as a separate one from a very remote period. There must have been at least two great and independent Dravirian nations or races, the southern, now represented by the Gond-Tamil peoples and languages, and the Gangetic or Bengal, now represented by Kol. The possession by the latter of a somewhat peculiar numeral system, although a sufficiently remarkable circumstance, is by no means anomalous, especially if the race occupied the lower Ganges and were a maritime and trading people. The more marked deviations of Kol from the homogeneity of the Dravirian system of pronouns and numerals appear to be referable to the ethnic revolution occasioned in the Gangetic basin by the entrance of the ChinoUltraindian and Chino-Tibetan race.

* In South Dravirian it occurs under the forms -ya or -ia, -iya, yo -a, -y or -i. The fact of the Kol dialects taking one of the common Dravirian possessives in their numerals and the southern group taking another, is one of those which establish an archaic separation of the two branches. Both systems go back to a period prior to the concretion of the possessives with the numeral roots.

root.

+ The superadded dental (-d, -dh) appears to be the Kol possessive -t, and its presence implies that the other possessive -ia, -i had become concreted with the It is remarkable that the Brahui term for 3 has a similar secondary dental, mu-si-t (mu-si being obviously a variation of the Dravirian muji Tul.) The Telugu vò-ka-ti, has also a secondary -ti. The exceptional Kol -d,-dh probably indicates the influence of a S. Dravirian dialect. The Auam mo-t preserves the Kol postfix.

same vowels o, u, in Dravirian proper, vo keeping its place as the unit in Toda and Telugu, and appearing in all the dialects with other forms in 10 and 3, in the latter being identical with the Kol-Ultraindian mu.

2 is bar (bar-ia, bar-ea). It is preserved in the Ka and Chong bar; the Binua mar, ha-mar; and in the contracted Mon ba, Binua ma, Kasia ar, Kambojan p-ia (unless this be a misapplication of the Kol 3); be (or b-e) Simang, hei (or he-i) Anam. The term has N. E. Asian and African affinities. The r element corresponds with the S. Dravirian 2, ira &c, (ara in some forms of 6). The b may be the m of 1 repeated, as in the binary basis of some other systems, but it may, with much greater probability, be identified with the v of avar, avara, avaru, varu, "they" (i.e. the 3rd pronoun followed by the plural definitive, which is glossarially the same as the dual numeral definitive). The Male bar, (also war, ber) found as the plural postfix in the 3rd pronoun, gives us the exact form of the Kol term for "two". It may thus have been originally the dual or plural form of the labial definitive which forms the unit, the first two terms of the numeral series being equivalent to "this", and "this dual," or "this -plural,” i-e. “ these”. But the idea of duality or plurality may have become attached to the definitive from its use as 2, in which case the application of bar or ar as a plural definitive would be secondary. The Kol dual postfix -ing, -ng, -n, appears to be a variation of ir, er &c. 2, similar to the Uraon en, but preserving i as in several of the 8. Dravirian forms.

3 is op, contracting to p (op- ia, p-ia). In Ultraindia it becomes pu-i, pa-i Mon, ba-i Kambojan, ba Anam, wu-i-p Sĩmang (inversion of pu-i), p-eh Ka, Chong, am-p-i, am-p-e, am-p-et Binua. The term is a modification of the labial unit of Dravirian and Kol; and the Dravirian 3, mu, mi, has the same root.

4 is upun, opun, pan, pn, in Ultraindia pun, bun, puan &c, probably a variation of the labio-liquid 2, which occurs with similar variations in other formations, bar, bur, pun &c. This explanation appears preferable to the analysis op-un, up-un, p-on, that is, the term for 3 followed by a definitive representing 1, identical with the S. Dravirian on, un, 1. The Kol term is different from the South Dravirian.

5 is mona, mone, moi, mo, muna, mun. It is confined to the Kol and Gond, the Ultraindian terms being different. The first clement of the term mo, mu, has the form of the labial unit found in the Kol and S. Dravirian 3. The second element na, ne, n, r, may be the South Dravirian 2, ra, r, er, ren, na, 1 &c (in 2, 4 &c) or 1, (as in 3), or a mere postfix or final. The term may thus be simply a unit, mo, as in the S. Dravirian system, with a consonantal final, as in 2 and 3, or a postfixed definitive, as in the S. Dravirian 3, mu-ru &c, to distinguish it from the mo of 1, (comp. the Ho mo-ya or mo-ia 5, and the Bhumij mo-y or mo-i 1); or it may be 3, 2, or 4, 1. The analogy of South Dravirian gives some special weight to the first suggestion, and the foreign affinities shew that it is well founded. It results that mun is only a variation of the same definitive that forms 4, 3 and 2, and of which 1 itself is probably a contraction.

6, tur, turu appears to be of undoubted Ultraindian origin [see App. A to Chap. VI].

The remaining terms are clearly Dravirian. They have no Ultraindian affinities.

7. The Kol term like the S. Dravirian appears to be quinary. The Sonthal iair is evidently the full form and the others contractions, the iya, aya, ia and cia representing the ayi, ya &c. of the S. Dravirian 5, and the final -ir, -r, the r of 2, so that ia-ir or ya-ir is still 5, 2.

8 is ir-al, ir-l-ia (in Gond, by inversion, ilh-ar, el-ar-ia). The ir is the S. Dravirian 2, corresponding with e of the S. Dravirian 8. The South Dravirian na-l or n-al 4 is 2 dual. The Kol ir-al is in form dual and may have been the second 4, but it is more probable, from the analogy of S. Dravirian, that the element 2 has reference to 10 and not to 4, that is, the full term was "2 from 10" as in some of the S. Dravirian names. The final il, 1, al, may represent 10, for the Kol 10 has the same final.

9 ar, ara (ar-ea, ar-e, ara-iah) has an external resemblance to the S. Drav. 6, ara Mal. (aru, aji &c.), but as there is nothing to shew that the term is trinal, and as the adoption of the Malayalam postfix-ra as part of the root would make the term of much later origin than the other Kol numerals, which must have been formed before the S. Drav. postfixes cohered with the roots, it may be inferred that the only common particle is the initial a, represent

ing 1. The Kol a-ra would thus appear to be 1 from 10 like the Dravirian term, and, if so, ra, r probably represents 10, as al, il, 1, does in the term for 8. The form of the postfix in 1, ea, corresponds with that in 10, whereas in the other numerals it is generally -ia.

10, gel-ea (in a Gond dialect gil, gul), is a peculiar term. The guttural is not found in any of the preceding numerals. It may possibly be related to the plural ko "these," and, if so, the dual king [ko+ing] is a similar example of the o coalescing with the i of the dual. In fact g-il or gel and k-ing or k-in (for n replaces ng in some forms) would thus be varieties of the same combination. If gel, gil be an integral substantive root it has no affinities in the S. Dravirian or Kol numeral systems. The Tamil kodi 20 is a different term, nor has it any relation to the Gurung kuti "one score", Gyarung kati "one" Magar, Lepcha kat "one", Naugaung Naga katang "one", Tengsa Naga khatu, in all which the guttural is a prefixual def.

The Australian and some other pre-Malagasy systems of Asonesia are more archaic than the Dravirian, for they have not yet raised a quinary or denary superstructure on the binary foundation. Some have only the two primary terms for 1 and 2, which are repeated for higher numbers. Others have a term for 3. Some use plural particles and words in combination with the term for 2, 3, or to express higher indefinite numbers. The more common binary roots have Dravirian affinities.

1. The labial occurs in li-mboto Goront. ri-moi Ternati, ipeh Bruner I., mo-tu, i-mu-ta N. Aust., peer Peel Riv., mal Karaula, and in the Australian compound terms ngun-bai, war-at, dom-bar-t, ka-marah, wara-pune, wo-kul, wa-kol &c.

The labial is the Dravirian unit, definitive and 3rd pronoun, and in Australian it is also common as a 3rd pronoun, and in some languages as a definitive postfix. The final 1, r of several of the Australian varieties-pronominal as well as numeral-appears to be the liquid terminal and postfix which is so common in Australian languages and is also a Dravirian and Scythic trait. West Australian has bal "he," "it" &c. In the Karaula mal, the definitive appears in the same form as the unit, and the Bijnelumbo war-at, Peel Riv. peer, Kowrarega wara-pune, Moreton Bay ka-marah are similar instances.

The Dravirian nasal definitive and3rd pronoun is also Australian, and in some languages it is the unit. Thus niu, ngi, no, are forms of the 3rd pronoun in Kamilarai (phonetically varied by the sexual and directive postfixes), and ngin, guin, nga, are forms of the same root in Wiradurei, the former language possessing also the labial 3rd pronoun, in fem., dual and plural forms. In the Wiradurei ngun-bai, 1, ngun is the nasal 3rd pronoun, and, what is interesting to remark in reference to the possessive form of the Dravirian numerals, it is not the nominative guin or ngin but the poss. gung. The second element of the compound, bai, is probably a contraction of the labial def. which appears in the form bari in the 3rd person of the imperative. Ba, wa, bala are also used as the assertive absolute. In Kowrarega as in Wiradurei the nasal def. is found in some forms of the 3rd pron. and the labial in others; nu-du "he," na-du" she," pa-le "they-two".

The Car Nicobar heng, hean, Simang ne, Borneon nih, indi, unii, enah, Philipine una, ona, uon, enot, Mille juan, New Caledonian nai, nait, Erub ne-tat, may be Draviro-Australian, but it is also explainable as a common insular definitive (identical with the Dravirian) applied to the expression of the unit.

The liquid definitive found as a postfix in Dravirian as in Scythic, does not appear to occur in the known Australian languages as the 3rd pronoun, unless it be identical with the nasal. In many of these languages 1, r is a dual and n a plural postfix in pronouns. In some vocabularies 1, r occurs as the unit,-lua Gnurellean, (whence youa Pinegorine), loca Raffles Bay, roka Terrutong. But these terms may be contractions of ngoro, ngolo, kolo &c, with the -ka post. The Raffles Bay 3, oro-ngarie (1, 2) suggests that lo-ka, 1, was ngoro (as in the Kamilarain goro 3, wa-kol 1), and ori-ka 2, ngori-ka.

The guttural occurs as an Australian numeral element both in 1 and higher numbers, but it is doubtful if any of the forms are referable to a guttural definitive. The naso-guttural 3rd pronoun of Wiradurei, ngin, takes the form guin, and, as the unit, ngun, ngung,-apparently identical with the possessive gung. The West Australian gyn, keyen, 1, resemble it, and the terms in other languages that have ng, g and k may be also variations of the same root. The Kamilarai ngoro (in 3) and kol (in 1) appear