= =f ars! Ga ae, lox ' aul I i. . bal a it Pr qt male i nN # a ' ay ee ul 7. ye Le aR E ie PS AD 71 | ive a] _ if i= a 7 A 4 A a | A ag ‘| : S l ti | ie 1 he abt ra ; Le i} : 1 Lisi | rH he a 1 ; f Piel digit : | hl uleae ort Fil Wi i he | ' ib , ite | f é ' al i iF mea aa { | uy ity ' r [ i @ ! " vs | =e eat a al | sect eft are Pe at fi lf ; = 5 al Pay + Pe 7 “) i |, «4 ‘ * Cae a "WW 7 L _ ak i Tah eo f é} tt mi ; at 5 ‘\ i 4 1S Pye oe! | i 20. oo ate 7 The | Ad . aN } | les ; if nit cen 6 7 J a if \ ‘i | vt * i = | ed ih i re | r el 15 i i 1 “ally it | || ae = i j Ee i = my Pin } | 7 % [ iv t| ay re ; . " AT te he etal git Ne fc ey | St iy, ne pert ae (Ue i ie | for Fr Te Le at Ch pamela aliel ery) Pint, ALS lg (Le et pie iy i Va nik hap al t k ‘ BS ae eat ete a ite) . } tN ey u i oy ts fi fe rie Fai | yin 7 i 4 iu Hi eral | 45 a i =4 LF et arte ue he a yet ' s te is 7 / a ret H i‘ ue BF tata i th a i by , : ite | | 7, fy ahs i ; vit | }, ie : ull Y eeksy es er “ al i i . Pha jay 9 Mi, ee ak ae cho en ra aad ih i ae A ig ry. ard ts, Aaa bs ia! Eran . 3 i 7 i> pie a = 4 - —a / Tee i wal il : 7 Wet eed nh hee 4, ; "ed aes Oe Na mies Vela mite fe i te ae * nani Pa Fae nae A Oe aa | oe ae faq # Wt Pol at Me pay lt mt he Ais =! eal ie nPiie iF Yew if cual | on t @ Sui ey aha f a, 4 a | on Al th ae t ity Nite. a ies ex hh Hy Pa eT ea Hf i ie Ba yin : =) h ie hia ny i , - rt i ioe aoe ale ae Po. a ——- = Se sa Pyciier Ji in 4 ! i ites aa is Ave mili fu vai U — ny aL uy ee seg Ve! P| \ Paces i een “5 iP il og “A [aa . ont ru ea ai if ali 1 8 1984 FUMIGATE RAFFLES LIBRARY SINGAPORE, S.S. aes ni ; mines t4 ‘ as es fj . ie Wee a Ui i fey 4 1 eqny Rava WHYALLA bo WA Bikes BW Yrs oe y yh ~ Wi i ¢ 4 c ww eeaguy yeob! ork NARA ile 1 NN ahh } isi at * - (L16- OS 7 é a Ke . Loge | ee ad Singapore JE te -—- s45R é ] ; Y ‘oi ike f fA Oe Lory ao i SOR ee ie romhing / 42] E Th HoLoay OF THE TuJo- PACE < | lle cin Wea RR. LoGanw fat VoL, 7. * ‘ a | a ’ # ‘i am ae iets ad " 1 ] 4 | . fi Wet oa ne suing. ie ‘Saw a! a ao. eel ba ll ee ie ag a1 Pl a hart fhe ; | + = ral ioe ae Lae \ af 2 F Sy _ a — Ln 7" = i ein — se ® — se 4 Wi = é I bite — ‘eas pais Those renders who have accompanivd me from the first do not need to be reminded that the publication of this work has extended over severu! years; and that the increasing light thrown on the comparative characters of ench family of language, during the revision of the sections relating: to it, has disclosed defects in the preceding ones. The continuity of publica tion and equality of treatment, originally intended, have been prevented by frequent and, at times, prolon ks in the attention | have been able to give to the subject, and, insome degree, by absence from the place of printing. The consistency which the work had when tirst rapidly written, a5 a statement of the opinions to which | had been led by a review of the other linguistic groups with reference to the Oceanic, has been lost by the lapse of six years, during which ethnology has not stood still, while L have endeavouring to bring these opinions to the test of a more searching enquiry into the peculiarities of the different groups. A final revision, on the completion of the work, can alone restore its uniformity, by bringing all its fucts and inferences into ars with the knowledge of the tine at which its publication may beconcluded. It seems nevessary, however, on the separate issue of the present portion, nearly two years ‘efter its eurlier peared, to warn the reader that some of its glos- arial detuils are at variance with the more accurate acquaintance with the Himalaic and Dravirian roots which | have obtained from the minute eompurisons in chap. vi. ‘These errors will be best understood by a re- ference to that chapter, and especially tothe comparative table of Dravirian and Himalaic roots which will be found in it. Some of the most impor+ tant will be here noted, in addition to errors of the press and of haste, In some places I have used the word Himalaic ina large sense, and 28 the paragraph explanatory of it was omitted in the proper place, it is necessary to meution here that, for want of a better term, I have applied it to that large group of cognate languages and tribes which have imme- morially clustered in and around the Himalaya and the ranges subordinate to it, and the preservation of the native character of which must be chiefly ascribed to the protection afforded by these mountains against the more powerful and civilised races of Eastern Asia—Chinese, Scythic, Dravirian and Arian. An extract from a letter to Mr. Hodgson (15th July 1856) will illustrate the application of the name, “ That my Mon-Anam group was the Bengali of the pre-Tibetan era using Tibetan for the present Scythoid branch) and conterminous with the Vindyan Dravirian dialects is demonstrated ; but [ am not prepared to admit that Dravirian has not a distinet archaie ingredient, not derived directly either from the Mon-Anam or the Tibeto-Burman branch of what I have termed “Himalaic” till you can supply us with a more ap= ‘propriate name. I conceive the Draviro-Australian branch of Seythie or rather ofChino-Scythic, to be of vast antiquity, and to have long pre- ceded the descent of the Chino-Tibetan race from their erind- Fins layan abodes, Its strong Scytho-Caucasian element appears to me to show that itcame round the western extremity of the great dividing barrier between middle and southern Asia. ‘The Mon-Anam or Bast Himalaic stem was more Chinese and less Seythie than the later West Himalaic or Tibeto-Burman. All the earlier dispersed languages—that is, their nixed and sometimes hybrid descendants—have a core of primary ie roots, retaining # close resemblance to each other, arid to those of the te< eabularies that have remained in and near the primary abode of thé Mid- Asiatic tribes.. In this way I would explain the peculiar Ch tiese element of Himalaic, Caucasian ( ed. by the mountains), and Dra viro-Anstralian, and the secon Himalaic element of Caucasian Draviro-Australian and otherlanguages. The East Himalaic tribes pro- bably ocewpied much of what is now eastern Tibet and western China; and thouygh'the precise line of their first southern migrations can hardly he traced witlr certainty; it is most consistant with the general character of the Mon- Anamglossary, to infer that they first descended into the, Brah- maputra basin by the»routes afterwards,followed, by the cognate Tibetos Burman tribesjand thence spread over the Gangetic gallo maxing with the prior Dravirians, and, im the course of ages, eliminating the Drayirian phy- sical element, though retaining Dravirian prononns, numerals &c. Of course there may have been other more-eastern, migrations, but the Mon-Anam branch, which predominated and spread everywhere,in Ultraindia prior te the Tibeto-Borman, had its primary southern ho nursery in Ben- gal or the Bhramaputra-Gangetic valley, for its basis of Dravirian, and of ‘8 secondary or corrupt dialect. of Dravirian, could, have been obtained nowhere else.” ees - r ~ The name is convenient in distinguishing the various. elements of Asonesian ethnology. » The-Jatest of the three formations of the Indian province’ has appropriated its-only general name, which is radically Hi- mialaic.* This has rendered crooner adopt. a second name for that forme tion which would otherwise have the first. claim to the desi tiun of Indian,—the Dravirian, A thirdisrequired for the ,:ntermediate great formation of northern India and Ultraindia, ,Tibetan might be made to include the Indian and Transgangetic rs Sp the proper Ti- betan type; but Mon-Anam has native characters which cannot be con- founded with those of the more Seythoid Tibetan, and it is most. convenient to use a distinct name forthe formationas a whole, » December 1856. ; , maa Abpitroxs axp Corrrcrions. ». eh. v. sec. I. Pronouns and Generic Particles. Page 1. The calculation of chances here ascribed to Bopp, is Bunsen’s. Alluding to the hypothesis that families of language had many distinct erizins, he says that “the very roots, full or empty, and all their words, whether monosyllabic or polysyllabic, must needs ‘be entirely different.” «There may besides be some casual coimcidences im teal words; but the law of combination applied to the elements of sound gives a mathematical roof, that, with all allowances, that chance is leas than one in a million or the same combination of sounds signifymg the same ise object.” My objections to this position haye been greatly confirmed by my subse- | quent comparisons of Chinese, Himalaie and Seythic roots.” The number of the elementary sounds that éntered into primitive language appears to have been exceedingly small. The same monosyllabic roots & onetic) are repeated again and asrain, and meet us in every class of words. Like all otlier arts, languaye was, in it# earlier stages, rude and narrow. Only , Sind (whence Hind, Iud) ia Himalaje root for river. | Paterial thin#s were named; and to the undeveloped family of savages, ew even of these required names. The same name sufficed for many ob- jects having common properties. ‘The growth of the analytic fuculty must ave been very slow. Most new names were but old ones. in new shapes, Distinct sounds were not in general invented or imitated for new concep- or ~ "The conceptions and the names grew together from the old stock, he separation of families must have been the grand seurce ik a intellectual and linguistic, By this means chiefly the primary rootsacqui variety in [pect the lication, Each new family or tribe became the nursery of a new dialect ; and the intercommunication of these dialects gra- dually eariched each with ideas ond vocables. Xt wasonly by the aid of hun- dreds of sister-dialects that it beoxme possible for any one dialect, after ages of growth, to make an approuch to a language in-our sense of the word. In every period of time and in every ef languayes the same mutual action goes on. Hence, as the genealogy of every existing dialeet ascends to the beginnmy of human «peech in the world or in the race, and Passes through tong periods of barbarism and of a mirute subdivision of tribes, its yocabulary has had innumerable proximate eourves. its discoverable homogeneity is in proportion to the narrowness or exclusiveness of the circle of dialectic development and interaction, It may be at a maximum in a group that has always remained secluded, so far as the geography of any province admuts of this, and altheugh the seclusion has lasted so long that archaically distinct dialects have now few vocab/es in common. Page 6: The new series of Vindyan vocabuleries eoupiled for Mr. Hodgson bv Mr. Nevill (J. B. As. Soe for 1856, p. 46) have the cqynmon form of thelst pron. An “Kowdh, na Savara, nai-sa Gudaba, (uoie nyo poss., na-nu Yerukaln, (na-mu, na—mbu—ru, pl., the secoud form come bining the absolute aud the relative pl. purticles asin the Telugu mi-ru). The second pron. has the ¢om. form in Gadaba mo, Yernk. ni-ew (pl. ni-nga-lu= la se a—va-ru), Kondh yi-nw. In the Nilgiri series furnished by Mr. Metz, a-du-m is given as a sing. form of the 3rd pron, in Toda, along with a—du, a—va-n, the pl. being a—va-r a—du—m, Page 15. Savara supplies a new proof that the labial 2d pron. of Kol ie a plural form used in the singular, It uses the same pl. form, in both numbers, a~ma-—n thon, a-ma—n ye (pe-n Gad., a-va-ru Yeruk). Page 7 pat from bottom, after “Pl.” delete the comma and inseré a stop | 9 last line atter “only” insert a comma line 18 trom top, for “dialective ” read dialectic 18 from top, after “ adi,—” delete ,— and insert , 14 from top, betwixt “ mip ” and “mu” insert ; ine 17 from.top, betwixt ‘mu’ and “sometimes” ansert a comma line of the first note from top, for ‘‘preseut” read parent ine 13 frous top, betwixt “we” aud “Tamil” delete . insert , , line 6 from bottom, betwixt bn” and “am” delete - insert , 19 line 15 from tep, betwixt “Libyan” and ‘‘form” delete , line 18 from.top, betwixt “the” and ‘“‘affiaity” insert pronominal 20 line 11 trom top, after “ing” insert , » line 17 trom top, after “eng” tasert , 3, line 21 from top, after “plural” insert , yy line 22 from top, after “possessive’’ insert , » line 9 from a. of the note, betwixt “ineorporafion” and. “2.” 1 FE 4 = es 2 be aay SSssssesevsse33s3s 2 4 snsert iv Page 20 line i i pera betwixt “Kinawari” add “Ti. tan” ansert , » 21 line 17 from bottom, after “i” insert , ni yy 22 last line, after languages delete. (and insert , 4) »y last line of the first note, for vol, ii. read vol. i. } yy 23 line 12 from bottom, betwixt “ the” and *Indo-F ” ine sert Draviro- Australian and y 24 line 12 from botrom, after “on” insert , ; y» 25 line 4 from top, after “ Zimbian” insert Draviro » » line # from bottom, for “person” read pronoun » _»y line 8 from bottom, after “centre” insert philologically » 80 line 165 from top, after “American” delete aud Atrican, » 36 line 2 from bottom, delete nya-n »» _» lastdine, for “nga-n” ead. nga-n,. » 87 line 9 from top, for “obl.” read abl, ” nw = » jine 11 from top, for “Kam.” read Karn, =. 38 Jine 27 from bottom, before “double” ensert a, and after “Chi- nese” delete , nndansert , ‘ so» 38 line 7 from bottom, before “poss.” insert in a» _ 39 line 4 from bottom, for “Kani” read Khaxi adel. line 6 from bottom, kha being a fagtitious Siamese pronoun. Its proper meaning is slave, After last line, add chi Thochu, ji Newer, » 40 line 3 from top, for ‘‘Toung-lhoo” read Toung-thu, _y 50 tameshe is a/f in Bhotian * ‘The following additional notes received after the Sec, had been-printed “off, were inserted in the number of the Journal in which it appeared ;— ist Note, P. 29, 2nd line from top, after the word “ African &¢.” _™ In other Indo-European languages the labial is demonstrative &c. The Welsh 3rdpronoun mase. is ev, vo, eve, Ivo, @, wager he being the common sibilant or aspirate 3rd pronoun hi, si, i &c, Thisis a re «markable coincidence wit Semitico- Libyan, The coincidence between ‘the Welsh and the tian 3rd pers. masc. has been remarked by Dr, Charles Meyer, Sanskrit has «a demonstrative base ma (Bopp § 368), “Zend has ava, Sclavonie ovo, * this." The Welsh labial 8rd pronoun is more likely to be connected immediately with the Indo-European labial demonstrative and postfix than with the Semitico-Libyan postfix. It is probably a remnant of the period when the labial as well as the sibilant aud dental might be used as a 8rd-proroun and it js quite possible.that the former was then masculine and the latter feminine, Thatthe labial “hab become neuter as a definitive postfix, and that the dental had dis- placed it as a 8rd pronoun and lost its sexual function, would not be anomalous, The Semitico-Libyan famil esents similar phenomena, In some languages the sexual tnnetions of the two definitives have been lost. In others the feminite has displaced the masculine. If such changes ‘took place in Indo-European they must have “preceded the separation of of all the branches save the Celtic, which appears to havo carried west- ward the use of both ponouns. The Semitico-L Aa aon tse renders it probable that the sibilant or dental was obiginallly obsolute or common, 2nd Note. P.-29, line 8 from top, after the word “Anam &c.” * Jo the Gond wur “ he,” bur “who,” the plural r of Dravirian oecure = nd in the singular,-the pl. takingey,-k (wur-g,bur-k), This is probabl one of the dialectic confusions or form to in the northert dilate and it may have had its origin in the southern use of the plurals as ho- norific forms of the singular. It may, however, have been the fern. form of the po yy with-7 for-/. In Australian, Yeniseian, Scythic, WN. E. Asian and Semitico-Libyan the labial definitive so frequently occurs with a final liquid in the singular that it is necessary to recognize the existence of this formas a very archgie one (bal, bar, wal, val, mal, man, bari, buli&e. &e.) There are even strong reasons for holding that this particle and the liquid or, ra, ri, la &c were primarily identical and that the dual and plural function of the latter wassecondary, and acquir- ed from the use of the labial definitive in its various full and contracted forms (e. g. bar, bari, ba, ar, ri, li, ni &c,) as the numeral “two,” II. Numerals. The new Nilgiri and Vindyan series have led me'to adopt some modi- fication of the analysis inChap. v. The reasons will be found in the sec, on the Mon-Anam numerals in Chap. vi ie 1, The 8. Drav. on-ru 1, I now read o-nru, and identify the root with vo ofthe 1 of Toda vod, Telug. vo-ha-ti, Nilgiri vo-ndu, vo-dde, and the com. pa, ba of 1 -hu-du &e. (p. 56). Yerukala has vo-ndu =o-ndu Karn. ‘The Male Mee Ae are similar forms. Conyp. also the Telugu va-nda 100, The Kol and Mon-Anam mo-7 &c. 1 is the game root, with a different Drav. poss, postf. 2. era-du and the variations in 2 and higher numbers I now read e,ra- du, ed, iru, er kc. e,i being the sole remnant of the ultimate root, which in its oldest form had ra, ru, only as a postf. or second element, ‘but afterwards superadded -du &c,, probably on the earlier postf, cone creting with the root remnant, Uraon has e-no 2, ma-no 3. From the favile and frequent elison of the initial labial it is probable that the full form of the initial root was be, bi. (p. 60). The form be-ra &e, agrees with the Kol ba.r; and that bais the mitjal rootand r a second element or a ee sow from its occurring with the guttural postfix in Sa» wara, bag a form preserved also in 7 of Yerukala vo-gu, Kiranti bha- - g-ya and Mon bo-& (2 for5), and identical with the Telugu vo-ka of vo- a 1 wodthecom. 3. Dray. 10, It is probable that in be,ra &c, 2, ra was a second archaic definitive or unit. 4. If 2 be e.ra-du &c., 4 must be na-lu,i, e, the secondary element without the initial one, The Kol pee retains the root of 1 with the eoncreted consonant of the second element or primary post, The form opur-ia is probably from op-pu.n, i.e. op 3, pun 1, The Savara cone tracted form of 4, vo.n-j¢ (1 for 3, 1) 18 evidently the full form of the “Tuluva o,n-jt 1. : 5. The Gudaba mo-lla-yi confirms the analysis of the Kol mu-na, mo-r &e:as3(S, Drav, mu-du ke.) for 3, 2, 8 e-ntu (noten-tw) asin 2, The Kol irl appears to be radically i.r-}, a contraction of the 8, Drav, i.ra-du =i,ra-lu, The exceptional Gond, Telugu and Tuluva forms must, in conformity with the amended analysis of 2, be read a,na—mu-r, e.ni-mi-d?, e.na-me i.e, ani, ena, em 2, and mu.r, mi-di, me 10 inthe form of the anit found in the Kol 5, 5. Drav.3 &e, = « vi ©9. The Telugu tommidi must be to-mi-di, i. @. mi-di for 10 asin ent mi-dt 8, and to, a distinct root for 1, preserved also in the Chentsu to-ta, 9, and correspo Lo apryy Pee the Dray. dental 3d pron. and def, (p. 56). The Kol a-r of 9 has a for 1, as in the 8. Drav. 6. | sete identification of the 8. Dray. on 1 with vo.n, and e.ra 2 with be.ra &c, completes the proof of the. between the basis systems of 8. Dravirian and of indyan, between both and the primitive labial stem preserved in Australian, In its first form the system was simply the labial deine ar ie the labial and liquid, repeated er compounded. It is clear that the liquid in ba,ra, bira, ke, hal is very archaic, and that it early possessed a distinct numeral and force. In the Semito- “African and Seythic systems, in which dee acta was the archaic ahisrwa name, the liquid, changing to the sibilant &c., been more stable than. the labial initial, and there are strong reasons fren as the essential element in2. See App. to Sec. 6 of ch. v. and Africn Wemerelt, PP. 18, 19, 43. See also the zd Note supra p. v. Wie rit. Miscellaneous ‘Words (pp. 78 to 163.) — 96. The remark as to the Drayirian affinities of the eastern vocabula- _ fies of Lrania—Pashttu &e.—has been corroborated by Mr. Norris, who, in his edition ef Dr. Prichard 's Natural History of Man, i, 171, doubts the % Ariane by bnohard to Pashtu. 104. The flowing paragraph were accidentally omitted in printing this section but were the same number of the Journal, They should come in on page 104 and preeede the para. commencing “The following are illustrations from Semitico-Libyan, ka] As the Seythic languazes appear to have always been located Asia, and they still ve.a form that allies them srowely bon both with 7 monosyllabic and with all the harmonic Janguayea, it is pepbable te ae in mother tongues of the more calling £08, widely the World were intimately allied in Ayelet to the Suen thic ones, A eomparison of the ey ythic names for some of the parts of the body may thus be considered as the first step towards ascer- HH the true relations. of the names for We eae objects in any of the coer e aul as the Dra inn of he fe ri ve ght examination o ane or e parts o the - shows rae weare deuling w nde primitive vocabulary and has suffered dialectic changes almost without limit, The same Scary roots are found in all the languages from the Kamchatkan to the Hunga- rian. The same secondary or dialectic forms and compounds are found _ in numerous vocabularies of the same and of different groups, sometimes proaaens the same SEC and sometimes varying in this respect. _ Most of these variations are evidently archaic, They were formed in that early Gane of language when fixed conyentional names had not been appropriated to each part of the body, but. several were described by the _ same primitive roots, the distinctions being indicated by the addition of other words and partly, in all probability, by esture, In time varipus dialectic changes of the kind we haye i took place, and the same root became current in a multitude of forms and with different convention- _ allimitations of ee The history of these changes is probably too complex to be completely recovered, and the blendings and extinctions of dialects that must have socusssd’ Maer they commenced have obscured vid / and diminished the glossarial evidence. I ~ The Scythic’ roots for the principal parts of the body are the labial; the guttural ; the sibilant and dental ; and the liquid (n, /, r,),—that is, uff the pay sounds, Of these 4 roots the 3 last are not strongly « erin guished. Rand §; 8, Tand D; D,N, Land R ure evidently merely variation of eich other in several cases. The roots’ are monosyllables of 3 forms,—Ist, the consonant followed by a vowel, which varies some- times even in the same group; 2nd, the consonarit preceded by a vowel ; 3rd, the preveding forme followed by « final consonant, vocahised or not The terminal ¢orsonaint varies, and it appears in general to be purely phonetic or ron-radical, The most common terminal is the liquid pn, 1 orr. After its, t, is the most frequent, but a8 s andr are ‘much inter- changed, these two classes are not well distinguished,, The labial and and guttural are much rarer. -'The Ist form, pure or with a fina} corso- nant, is the most common, The 2nd is chiefly found in the Ugrian languages, and as similar forms are produced by the elision of an initial radical consonant, itis not always easy to decide whether the Ugrian con- eonant is radical or not, Redupticated formis of all the roots occur, Bes sides these forms, others occur in which a detiuitive is attached to the root, generally postiixually, The definitive is sometimes a simple vowel, generally prefixed, but most frequently a consonant, generally postfixed, and either simple with a final or initial vowel, or such a monosyllable with a final’ consonant. Double postfixes also octur. It is often difficult of impossible to decide whether the final Consonant is 4 postfix or part of the root, Where the root has a final consonant the servile character of the superadded consonant is in general free from doubt. “— - The following erg ay will illustrate this diversity of forms, Ist, pa, ba, wa, va, pi, pi ie ¢; 2nd, ap, ab, ip, ib, up, ub, op, ob, ke; Srd pan, ion, bar, bir, bur, pat, pit, pet, put, pas pus, apt, ipt, upt, in, pe pas, pis be Ve? deh, with a def., palan, piga, wilyt, diye wilugt, wilyulgt, burwn, pank, pak 5 5th mimi. ~ The ition ed table of Sey thie names of parts of the body is intended to facilitate the comparison of the roots.® I have thrown into it all the yo- cables thatare found in Klaproth’s Asia Polyglotta, for Head, Face, Eye, Hair, Mouth, Lip, Teeth, Tongue, Nose, Ear, Hand, Finger and Foot, and the arrangement aE eR phonetic. In several instances words that resemble each other in sound may be-varieties of roots independent in their origins. But there van be no doubt that a large proportion of those voca- bles that associate themselves phonetically in the table are also radically cognate. Such a table, for ethnological purposes, should contain not only all the names of parts of the body, but the roots in all their other applica- tions, and When a thorough Beythic philologist appears we may hope to obtains tables ot this kind. Without such comparative vocabularies of groups of roots in every family of language, itis impossible to ascertain with precision the various degrees of affinity which connect any given language or family with others. ' 119. This should have been 115, and followed by pages 164, 165, and 166. The page following 166 was omitted altogether in making up the forms and is here inserted. Itis p. 228 of Journ. of the Ind. Arch. for April-June 1855. © This will be given with similar tables for other families. iti ma, wi, mi &c.; 2, mil; 3, ma-bora, nu-bara &c; 4, This, mas, mit, ' met, mat, mad, mut &e 5 5, mikamek, mak, &e. [See Bye. } Making the hichest allowance for accidental coiucidenees it is clear that the same roots, the same phonetic formes of roots, and the same combina- tious of root and postfix, are found with Variable meanings not only in the different lanzuaves of the same family, but im different families, When the various Aso-Afvican families are compared with reference to their vocabularies of prumary words, they appear as if they were all dialects of one motier ri pee and it does not seein possible to account for phenomena so purely dialectic without concluded that each lancuage ascends, through various phases to an ultimate monosyllabic condition, and that, at some periud remote even in the purely monosyllabic era, their protophasta were, in reality, dialects of one language. The roots which we have been ney ay originally have been current in a single family, before they me dispersed amongst many, and by the separation of these received various dialectic applications. Without such original linguistic concentration or unity, followed by such division and dispersion, the facts cannot be reconciled, for the mere dis- semination of the words of dominant tribes in ages when the Old World. Was peopled in all its principal divisions could not account for an agree- ment so radical, so ati verend and so complex. It will explain many of the coincidences, but nothing short of the admission of one primary vocabu- lary having been preserved in separated families and been dialectically modified in its applications, can explain the whole. Amongst the secondary dispersions and diffusions it is clear that the ythic or prote-Scythic is by far the most important as it was universal. t rests on the Chinese or monosyllabic stage of vocables, and it enters that in which a definitive became attached to the roots. Many of its pe- culiar forms and combinations are found in all the southern and western rovinces, and it seems to bé a necessary inference that before the Austra~ ian formation was carried to Asonesia, Semitico-Libyan to Africa, or the Euskarian to Europe, they were comprised along with the archaic Caueasian, Tibetan, Scythic and Indo-Buropean, in a comparatively nar- row Asiatic geographical circle. -The only remaining formations, the qos and American, are expansions and deyelopments of proto-Scythig inlects. . 128. The note belongs to p. 129, and note + of p. 129 should be note * of p. 128. 152. line 3 from the bottom for pishik read to-pisa, and in the follow~ ing line, for to-pisn, revd musa, mus. 137. The substantive root in the name for the Buffalog is the Jabial. The liquid is the root for meter, e-ru-ma = water—cow. 138. In tanga cow of Jili (not Singpho) the root is nga (ta-nga). 140, The statement that in Chinese the root alone signifies buffaloa and the inference trom it are incorrect. In Chinese, as in Dravirian, the name for the buffaloe is water-cow (or ov), and it is only by contraction thatyu &c. alone is applied toit. . 141. The Deer oe is Wentical with the Bhotian god mare. 145. The sibilant name of Tiberkhad &e. is Tibeto-Ultraindian. _ 4155. 7th line trom toot, for moon, read silver. 157. del. Sth line from foot. : 164, 165, 166 are misplaced ; they follow p. 119. . « P 159 to 167 following 166 should be distinguished by an asterisk. aT) 185 182 189 190 Egrata (NX caar. v1 Secs. 1, 2, 3 and 4, 184 11th line from the bottom for these read those top for Gangitic read Gangetic 4th ae a last line, for r-lik read i-lik; syod, light, Turk. is a distinct root, 1 from the Tib, bod. , 10th line from bottom for all of read of all 14th line from top * | 7th line from top after Jilinsert , 11th line from top ajter being and manuer insert , 12th line from top after mee tmsert , 13th line from top after Hor insert , 15th line from top after range insert , 16th line from bottom a tan cohen eors ‘ 9th line from bounce Secmesieced roots 8th line from bottom efter Chinese, txsert — 7th line from bottom after former, insert — 6th line from bottom after Tariar, insert — 3rd line from bottom after dialeets insert , ‘ 2nd line from bottom a/ter Chinese insert , 10th line from top for portion read i lith line from top after vocalic insert , 12th line from top a/ter words dmser? , 16th line from top for which in read which is 14th line from bottom after Manyak insert: 8th line from bottom for brigi read brigt | 6th line from top after labial and Gyarung insert , 6th line from top de/ete b in and ansert bi, llth line from top after prefixes and infrequent insert , 13th line from top jor mo-; k-, read mo-; hi-, und delete cha-, ki-, 8rd_ line from top after trait insert , j 12th line from top after present detete ) and after generic insert} 14th line from top Yor localitive read locative 16th line from top for ka-, ta-, read ka-ta, 17th line from top for ta-, ta-, read ta-ta— 1 line from bottom, for Tibetan read Tibeto 11th line from top after erudes insert , 10th line from top for hazang read kavang 10th line from bottom after origin insert , 8th line from bottom after province insert , afir retained insert , after de insert , 2nd line from bottom after form insert , . 2nd line of the note after pronoun delete . and insert ; 3rd. line from top after definitive insert , 6th line from top after substantial ancert , 4th line from bottom. after pronoun mseré , 3rd line of the note after shui, insert sa, 15th line from bottom for ngo read ngo, 11th 7th 5th 5th 7th 7th 12th 15th last 11th, 13th 15th 11th 15th 17th 12th 7th Sri last 15th 5th 10th eth 19th 16th 15th 15th Tth Gth Sri Sth llth 15th 15th 3rd 2nd lith 4th 14th 16th 4th 7th 16th 7th 4th Sra last 21d ard line from bottom fr formations read forms, line from bottom after form insert , line Irom bottom a/fer Bhotian insert, line from top for Tanglhu read Tungthu, : line from bottom for nom, read nom.; for peu read poss.; yor n, and p, read n.p. line from top after chha insert , | line from top for -gen read -gen, aengey ? line from top delete T being a common Bhotian augment, ‘line of the note for anology read analogy, line from top, for pen read hen, line from top after khui insert , line from bottom after also insert , line from bottom for cha-ta read cha-tu, line from top after -sin insert , line from top far Nyertshmsk read Nyertshink, line from after -ra-ng msert , line from “bottom after Klaproth insert , line from bottom for seem read seen, line after Ultraindian) insert , line from top after &) insert , line from bottom for Kyan read Kyau, line from top for Kyan read Kyau, line from top for 2 read 2nd, line from top after kho dasert , . line from bottom aftr dialects insert , line from bottom after and inser , after the insert , after dla wnsert &e. line from bottom after Gurung, insert the line from bottom after &e. insert , and for Da read Drav. line from bottom after Seythie insert , line from bottom for lar, ‘od read lar, ler, line from top far -le read -ln line from top, for -kye read -kyi line from bottom after Naga insert , line from bottom after Abor-Miri insert , line from bottom for ngar read ngal line from top after det. and Bhotian insert , line from top before is insert , It line from bottom after plural insert 1st ; line from top after -ta tisert , line from bottom ar numeral insert , line from bottom before Tibetan tsert , line from top for Bhotian read Bhotians line from top for naga read Naga line trom bottom after ana-ta wmsert , line from bottom after tu insert , line from bottom after more insert or less line from bottom after Chinese wsert , line after pron. msert , At ven: line from top after Dravirian insert , line top (for respect read respects. » Alth » 24th 221 Sth 10th 222° 18th 18th 224 22nd 225 14th 226 17th 229 17th Sth 231 9th 234 6th 245 5th 937 15th 240 3rd line from top after particles insert , line from top after forming insert , line trom top after Chinese insert , after dialeets insert of line from top after European insert, , . line from top after Bhotian insert , line from top after languages insert , line from top after allied delete - line from the bottom, for The read It is line from for achik read gchig, line from the bottom for ni’ read ni line from bottom for gi read ngi line from the bottom for ne-ro-ka read ng-ro-ka line from the top after def. add g-, and after b- delete g- line from. the bottom for Athapas, can, read Athapascan line from the bottom for puli read p li line from the bottom after nyet, was line from the bottom after all, add the. line from the top for 3lread 25 — line from the top for affinitive r ad affinities line from top for , at read . At line from top'for by read ly *.* For a later analysis of the numerals, and many forms not given in Sec, 5, the reader is- referred to Sec. 7 The comp, to Sec. 6, vocabulary of Bhotian may be corrected by a reference - — — a oe Le ; . lh Rae Ti Pa | ¥ ; 4. Lond a rot Ps ee — =a hagipan © et Alea eet are i he = eu . Sat ees aL? i » - * « ‘ va ea — ee _ = - j 4 ip » © = , ie i ' = i> y ; a | yf! 5 “i oa 7 ; ls wl i « i - ie - = tg L ¥ —- ao : t. 7 * * ie a u = - ; Tides ain roa ee og 7. ye | 4 ma ” a ‘ ;: i al zee, vali oa ae ig ess ae we ie. ae dee, io 4 wis [ fl os 4 wy ——- q I 7 é } jim ‘i . a | is ‘= ‘ roe ' « . a “ +* = ' ~ a af * ’ , = ; ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. By J. R. LoGan: LANGUAGE. PART II, THE RACES AND LANGUAGES OF §. E, ASIA CONSIDERED IN RELATION TO THOSE OF THE INDO-PACIFIC ISLANDS. CHAPTER V. (Continued). ENQUIRIES INTO THE ETHNIC HISTORY AND RELATIONS OF THE DRAVI- RIAN FORMATION,—EMBRACING NOTICES OF THE PINO-JAPANESE, CAUCASIAN, INDO-EUROPEAN, SEMITICO-AFRICAN, Fd | + be sO EUSKARIAN AND AMERICAN LANGUAGES, ® - See. 11. GLOSSARIAL INDICATIONS OF THE ETHNIC HISTORY AND RELATIONS OF THE DRAVIRIAN LANGUAGES,+ 1, Proyouns ayp Generic Parricuzs. A comparison of roots, unless it embraces a wide field and is made with extreme caution, cannot lead to solid and satisfactory results. That great Iranian philologist Bopp has said that the chance is less than one in a million fur the same combination of * In the present state Soret: comparison of words for ethnic - poses must be Fano rfect "Bho snont distinguished Uae peianl not been able to avoid fonder when they have dntargel the the circle of pun A 2 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. sounds liaving the same precise meaning in unconnected languages, This calculation of probabilities is evidently based on a formation of which the syllables are frequently biconsonantal and the words polysyllabic. It is totally inapplicable to monosyllabic languages, or indeed to a comparison of ultimate roots in any formation, because these roots are generally monosyllables. In the Kwan- hwa Chinese, for example, the number of words is about 48,000, but the sounds to express those words only ainount, even with the tone flexions, to 1,2U3. So that each sound, on an average, would _ represent 40 different words if these words were allin use. By enlarging the number of monosyllabic languages for comparison, the number of homophons increases. But this is not all. In the progress of language the tones decay, become reduced in number and are ultimately lost. When the process of emasculation goes on without interruption, the vowel sounds are contracted to a very sma!l number. In Philipine there are only three distinct vowels, o and u, i and e being very commutable. Add to this that in parative studies in order to take in languages with which they are imperfectly nequainted. A complete investigation of the ethnic histery of a single root demands a thorough knowledge of all the languages in the world and no single philologist ean ever attain this knowledge. Hence it is only by combining and comparing the labours of numerous comparative linguists that the ethnology of roots will ulti- mately be perfected. Roots ramify through vocabularies in a very complicated manner changing not only their forms but their meanings, so that it is not possible, by merely turning over the Jeaves of a dictionary, to a-certain whether a given root exists in a particular langnage or not. We must know the phonology of the larguage, its phonetic and glossarial re'ations to other languages, and the kind of analogies that prevail throughout its glossary ani enuble us to trace the metamor- phoses of its roots. The only man who can pronounce whether a given root exists or not in a particular language, is a sound comparative linguist who hus devoted himself to a thorough analysis of that langnage. Until complete analytie glossaries are prepared, the comparisons of ethnologists must continue to be in great measure empirical, and must be received with a considerable allowance for errors, ‘The following comparisons require a large allowance not only from the necessity of the case but f.om the special disadvantag: s under which the collator labours. ‘They are limited to the classes of words mentioned in the Vrefatory Note to Part IT. A full ethnic comparison of the Dravirin vocubularies with those of other families would be a labour not for a single lite Lut for the ei hnologists of several generations. [Before sending this section to the press I received by the Jast mail steamer Chevalier Bungen’s Philosopliy of Universal History, to which Professor Max Miller has contributed two chapters on the Seyrhie, Dravirian, Tibeto-Ultra- indian, Thai, ani Malay languages. Some ol the glossarial details in-this section and in the next chapter bave I iind been anticipated by Prof. Miler, Where he has supplied data which were not accessible to me, I have aided a few notes which are distinguished by brackets. The supplement containing the compa) ative voca- bularies havirg been printed some tiwe since, [ have not able to subjoin any notes to it. I do not in this place offer any remarks on the coincidences hetween Prof. Miiller’s views on several points, and those previously published by me in the present series of papers. They will besufficiently obvious to ethnologisis who have read my 4th and preceding chapters, with the geweral remarks on Asonesian ethnology contained in the volume of this Journal for 1850). ETHNOLOGY OF TILE INDO-PACIFIC ISLAND3. 3 comparing different formations, and even the various dialects of the same formation, consonants and vowels frequently exhibit great instability, so great indeed that it can be asserted with perfect truth that each vowel is capable of being, by successive gradations, transmuted into all the others. The same remark applies to the consonants. In Polynesian there cannot be said to be more than 10 (in Raratongan and Mangarevan 8 ) consonants, the sonants having generally become eanfanuded with the surds, The dentals are transmuted into the liquids with great facility. They pass into the gutturals through the strong mutual affinity of the surds & and t, and into the labials through the liquids. Thus, to start with t. It may pass into g through k, on the one side, and through d, r, 1, n into m, b, v, f, p, on the other. Ils direct affinity to the sibilant and aspirate th, s, z &c. isso great that it frequently passes into them in many languages. Particles, whether separate, formative or flexional, are generally monosyllabic, and even to a lurge extent uniliteral in all formations. In the Burmah-Tibetan, the pre-Arian Indian, the African, the Turanian, and, it may be added, in the Iranian, words of all classes are radically monosylla- bles. It is evident, therefore, that the phonetic identity of a par- ticle in two or even more languages has hardly any value at all as an isolated fact, for comparative and ethnic purposes. It hap- pens, also, that a number of identical particles are so widely spread throughout most of the formations of the world that nothing can be learned from them per se, respecting the specific affinities of different formations. We arrive at this rule, that it is only by comparing particles in groups, and in connection with the entire phonetic and ideologic character of cach language, that positive ethnic conclusions can be attained. In the Burmah-Chinese languages there is little connection be- tween the particles. They are in general as isolated and independ- ent of each other as substantive words. In the Dravirian forma- tion, on the contrary, they are intimately connected both phoneti- cally and idiomatieally, and this greatly facilitates their comparison with those of other formations. In Dravirian we find a number of particles formed into a well marked system, presenting even flexional traits. For example the principal pronominal terms, as ex- hibited by the purer languages, or those of the South, are, na, “1,” and ni “thou,” n in the plural becoming m. Thus the three main 4 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. pronominal elements may be considered as flexionally related, and this gives to the Dravirian system a marked character. In addi- tion to this the root is reduplicated, with a change in the second vowel, or it is combined with a definitive particle. A. Pronouns.* Before attempting to trace the range and the affinities of the Dravirian pronouns, it is necessary to determine their proper forms, and mark their variations as accurately as possible. The root of the First Pronoun occurs under the full forms na (Tamil, Kurgi, Karnataka, Gond, in pl. Malayalam, Male), nga (Malayalam) and ne (Telugu). The vowel becomes o in some forms, The definitive -nu is postfixed in Karnataka, Telugu and Khond. The Gond agentiye nu-na appears to invert the relative position of the root and the definitive. The common Gond form, nak, preserves the true vowel of the root and postfixes the preval- ent definitive of that dialect, Tamil, Malayalam and Kurgi post- fix the contracted form of the definitive, -n. The root, as fre- quently happens in Dravirian glossology, loses its initial consonant in some forms, e. g. anu, an, a. “The form en may be an inversion of ne, but it is better explained as a contraction in which e is the radical element (en from nen or nenu like an, a from nanw, anw).+ * On the general subject of the Dravirian pronouns I may refer the reader to the valuable papers by the Rev. Dr Stevenson in the Journal of the Bombay Asiatic Society, and in particular to his article in the number for January 1852. My own lossarial comparisons had been independently made before seeing this paper, but it is due to Dr Stevenson to remark that one of the affinities which has considerabie weight in my deductions has been noted by Dr 8. although only as om isolated fact,—that of the 1st pronoun to the Chinese ngo, His general inference that the Draviriah protouns are of a peculiar type more allied to the Turanian than to the Sanskrit—unless it refer to the structure and not to the roots—is open to the remark that the Sanskrit roots are Taranian or Scythic while the Dravirian are not. The Ist pronoun, Dr Stevenson remarka, *‘ is allied to the lan es of Arabian and Syria on the one hand, and on the other with the Chinese family,” and also With the Tibetan.” The foreign affinities of the 2ad pronoun are not adverted to by Dr 8. The nain scope of his papers is to distinguish the Dravirian from the Sanskritic elements in the Guzarathi- li class of ey The honorific ap, apun, apan, &e., of these languages identifies with the Dravirian avan, Every student of the languages of India will find much matter of the highest value and interest in Dr Stevenson’s papers. His comparative vocabulary of the non- Sanskrit vocables in the vernacular languages of India promises to be a work of solid erudition, and its completion will be an important service to Indian and Asonesian ethnology. , ‘The resemblance between the Chinese, the Tibeto-Ultraindian and the Draviro- Australian pronouns was shown in the glossarial tables in my paper on the ‘Traces of an ethnic connection between the of the Ganges and the Indian Archi- pelago before the advance of the Hindus into India” read before the Royal Society of Edinburgh in January 1851, vide ante vol, vi, p, 654, t In chap. TV sec. 6, I have considered en, ne as seemingly the radical form, and in some cases assumed as portions of the root elements that I now refer to ETHNOLOGY OF THE INDO-PAQIFIC ISLANDS. 5 In the Tamil ya-n, Tuluva ya-nu, the contracted forms an, anu, take the common y prefix. The Todava one, on (pl. om) isa similar contracted variety, with the radical vowel changed too. In the broad form won, wom it assumes a quasi-consonantal augment, like some other words, e. g. on, won, “one.” The o form of the vowel also occurs in the 'Tamil verb form of the plural om, corres- ponding with the Todava; in the Gond oblique no-wa sing., and in the Male poss. ong-hi sing. om pl. (Tam, Tod.) The substitu- tion of o for a is characteristic of the Toda phonology. But as the Gond and Male forms cannot have been immediately derived from it during the era in which the more civilised Dravirian nations liave been interposed between the Todas and the Gonds, it is proba- ble that the o forms are very archaic and were at one time widely prevalent. The Tamil remnant in the plural of the verb postfix is a strong evidence of the antiquity of o. Possibly it is the original form, but the general character of Dravirian phonology makes it more propable that variations in the vowel existed from an early stage of the formation. In unwritten languages dialectic changes in the vowels are very common. - : The Second Pronoun has the full form ni in all the Southern dialects and in Gond, with and without postfixes (-nu, -vu, -er, -/), The forms nu and u [def. replacing pronoun] oceur in the Tamil piural poss,, and na in the Todava pl. The contractions i and ai are also found. Two kinds of pronominal plurals occur. In one the root post- fixes the ordinary plural definitives like substantives. All the relative or “exclusive” plurals of the 1st pronoun are thus formed. The second kind postfix m or flexionally replace the n of the singular by it. As it only occurs in the absolute or “inclusive” forms of the Ist pronoun, it is possible that in Dravirian, as in some other formations, one of the plurals of this pronoun is formed by the annexation or incorporation of the plural of the 2nd. In this view m would be radically a plural particle or root of the 2nd the definitive postfix. The great commutability of the vowels in both the proper Dravirian and the Kol dialects, with the agglutinated and conecretionary condition of the pronominal system, ren ers absolute certainty unattainable in minute analysis of forms. The Kol attinities favour the opinion that en, ne was the original form, while the Australian and the still more remote and primordial aflinities render it hardly doubtful that the most current agentive-form in the South na, nga with its variations in o, is the most archaic. Whether the three vowels a,o, e, were archaically flexional,—that is, marked different forms of the pronoun, agentive, oblique kc—is considered in a subsequent page. 6 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. pronoun only, and it would simply carry into those forms of the 1st in which it occurs the idea of “you” in addition to the original “I”, the Tamil na-m being thus literally “I-you”—i. e. “I and you.” Its displacement of the n in such forms as the Telugu me, Karn. and Toda am, Gond ma, would be a flexional change of an ordinary kind. In the Telugu me-mu both the root and the defi- nitive postfix are flexionally changed tom. Whatever may be the ultimate origin of this exceptional m, and whether or not it passed from the 2nd pronoun to the Ist, its true character, in the present condition of the formation generally, is that of a plural element confined to the 2nd pronoun and to the absolute or inclusive form of the Ist. Toda however presents a remarkable exception in its 3rd pronoun, which in the plural is atam or adam. This remnant of the undoubted use of -m as a plural definitive, and not merely as a form of the 2nd pronoun, when taken in connection with the Gangetic and Ultraindian remnants of a similar usage adverted to in a subsequent page, leaves little doubt as to its having been a very archaic plural particle in Dravirian, or in one or more of its branches. The phonetic identity of this archaic plural postfix with the archaic neuter (sometimes fem.) definitive postfix in m, b, p, v can hardly be accidental, The fem. lis also identical with the common plural postfix ia |, r. The Karnataka pronouns have ~vu inthe plurals of the Ist and 2ad persons and -ru in the 3rd as in the other languages save Toda (in Taluvu -ra becomes -li). This -ve is ilentical with one of the forms of the neuter definitive. In the use of these particles the dialectic confusion and irregularity are very great, and it is difficult to determine their true primary functions. The Vindyan languages present some remarkable dialectic pecu- liavities. ‘The Gond and Khond have the common a forms of the lst pronoun. The other nothern languages in their agentive forms have only that contracted variety of the e form which is the sepa- rate pronoun in Tulava (en), and occurs also as a possessive and verbal form in Tamil, Malayalam and Karnataka, and the plural of Kurgi, the full form being found in Telugu nene (pl. mem). In the more purely Drayirian Male and Uraon the Tuluva form is preserved unmodified en Male, enan Sing., en P]. Uraon. So n the possessives,—Sing. Uraon en-ghi, Pl. Uraon em-hi, Male ETHNOLOGY OP THE INDO-PACIFIC ISLANDS, 7 em-ki, Male preserves other varicties also, as om PI. (in addition to na-m) ong-ki poss. Sing. In the Kol dialects the vowel changes from e to i, ing, cing, aing, inge. These forms appear to preserve the original possessives of South Dravirian, to which in a later page I refer the e. It shouli also be remarked that the vowels i, e, aare definitives and definitive prefixes in Kol as in Dravirian generally. Compound vowels oceur both as a simple definitive and as a possessive. Kol has ia or ya poss, as in 8S. Dravirian and it has ayo, ay, ai &e as a definitive or 3rd pronoun, identical with the Tuluva 3rd pronoun aye (so ayi-no, “this”). It is foand also in Male. Compare the possessives ai-ye Bhumij, ahi-ki Male “his” &e, The change of ai into e, or e into ai, is easy, for e is but a condensed form of al. The Male and Uraon 2nd pronoun is the South Dravirian ni,— nin Male, nien Uraon. The Khond inw is Tuluv«(pl.) which again is a contraction of the Karnataka ninw. Gond has the full form with its own def. postfix in oblique forms, nik. Besides this form Gond has a peculiar agentive form imma, to which the Kol 2nd pronoun is allied, am Bhumij, Mundala, um Ho, umge Sonthal (ami poss.) The Kol duals and plurals present further variations of this labial 2nd pronoun, me, m, be, pe. Its probable origin is adverted to further on. The Dravirian plural element m is found in Khond,—any “1,” amu “we,” inw “thou,” mi “you”; Gond nak “I,” mak we,” nuna “1,” mar “we”; so in the oblique forms of the second pro- noun nit, S. mik, mek Pl., Male has na-m, o-m, “we,” e-m-ki, na-m-ki Pl. poss., Uraon e-m-hi PI. poss. The Kol plurals in m, b, p represent the Dravirian plural labial. In the Ist pronoun the relative plural takes the common plural def. -le, and the absolute only has the labial, under the form bu, * conformably with the South Dravirian idiom. _ It appears from these details that the original forms of the pronouns were na or nga “I” aud ni “thou”; that m was a plural definitive originally generic but afterwards restricted save in Toda to the 2nd pronoun and to the plural absolute of the Ist; and that the ordinary plurals of all the pronouns were formed by the plural definitives used with nouns, The form of the Ist pronounin * Inchap. IV. § 6, this form is not identified with the Drayirian plurals in m, but it is inferred that the labial element represents be, “you.” 8 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. en is a dialectic variation which must have prevailed in the parent Kol dialect as in Tuluva. The Gond imma of the 2nd person is evidently a secondary form (in which i is the common pronominal element) as the regular primary form nix is preserved in the oblique eases. The allied Kol labial 2nd pronoun must be of similar se- condary origin. The Kol dialects distinguish the dual from the plural in pro- nouns, as in substantives, the dual form being given by annexing the nasal to the plural. Thus the substantive pl. definitive iS ko, which in the dual becomes king [=ko+ing]; the pl. rel. of the 1st pron, is alle, which in the dual becomes alleng; the pl. of the 2d pron. is appe, which in the dual becomes abben. The dual particle is probably the Dravirian en “ two” (the Uraon form) but it may be a variation of the Draviro-Ultraindian plural el, le, li, ni &e., the dual being indicated by plural particles in some other families (Semitic, Scythic &c.) as well as in some Australian dialects.* The South Drayirian dialects with Gond Uraon and Male, do not possess a dual, Besides the indication of number and case, it does not appear that any other ideologic element is involved in the postfixes or flexions. The 3rd pronoun indicates sex by its postfixed defini- tives, the consonants being » masce., 7 fem. and d, th, ¢ neuter. There are no clear traces either of these or of a vocalic distinction of sex in the proper pronouns, which is the more remarkable from the sex definitives having, in the archaic stage of the formation, been used with substantives, and from their being found largely concreted in all the vocabularies as well as still partially current. — If any sexnal function can be ascribed to the pronominal postfixes, it would appear that the common forms now in use are masculine, -n and -nu being tho form of the postfix, If sexual forms were eyer current, we might have expected to find some traces of a euch form in the 2nd person, but / nowhere occurs as the post- ¥: ; The variations in the vowel of the Ist pronoun to e¢ and in that of the 2nd to « may have been glossarial. There are indications of this with respect to e, which however may have been the com- mon phonetic variation of the final vowel found largely in the vocabularies. If, as seems more probable, it had a flexional power, . * In Australian the plura] particle forms duals and one of its variations is -le. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 9 it would appear to have been possessive (and oblique), as it is now found in ali the Southern dialects, save Telugu, in those cases, or as the agentive postfix to verbs, which is radically possessive. In Telugu, by a dialectic variation, it occurs only in the nominative, the oblique cases taking the primary a. The Northern dialects, Uraon, Male, in their preference for e, follow Telugu, or more probably the Southern Tuluva, which has other special affinities with the Northern dialects including the Kol. It is probable from this that a (sometimes varied to 0) was the proper nomina- tive vowel, and that the substitution of the possessive e for it was a dialectic variation which spread from Telugu or Tuluva to most of the Northern dialects, or was internally produced by the loss of the ideologie distinction between the two forms. It is clear that the use of e in the possessive like that of m in the plural belongs to a very archaic condition of the formation or some of its branches. It is not probable that in any single branch there were originally two modes of indicating the plurals and possessives, and it is still less probable that both admitted of being combined. When we now find such combinations it is to be inferred that one of the particles is primary and the other secondary, the combinations having been produced by the blending of a foreign system of postfixes with the Dravirian or of two Dravirian systems previous- ly characteristic of different branches of the formation. ‘The anti- quity and wide prevalence of the ordinary plural particles in J, r &e are proved by their occurrence not ouly in South Dravirian, Kol and Gangetico-Ultraindian languages but in Asonesia. Sut one branch may have originally possessed labial plurals. The possessive in e whether postfixual or flexional must have preceded the use of the superadded possessive postfixes. “'The most probable explanation afforded by the Dravirian particle system by itself is- that the pronominal root na took the archaic possessive in i (in, ia . &e South Dravirian, Kol) and that this became e by the coales- cence of the root vowel a with the definitive vowel i (na-in—nen). But even the current possessive has sometimes e. Thus in Tamil we find ei, in Malayalam ye, in Dhimal eng &e. The u of the 2nd pronoun can hardly be explained as a merely phonetic variation of the radical. In the Auc. Tamil it occurs in the full form nu- in the possessive plural only nu-ma-du, the B 10 ETHNOLOGY OF THE INDO-PACIFIC ISLANDs. singular being ni-na-du. If any inference may be drawn from this, it is that it is possessive and probably plural. In Mod. Tam. it occurs in the possessive both of sing. and pl. u-na-du, 8. u-ma- du P. In the Kol dialects it is also found with a plural force under the form bu if my analysis of abu be correct. The absence of e or iin the possessive of the 2nd pronoun is accounted for vy i being the root vowel of the pronoun itself. In Chap. LV. it was stated that ‘the objective appears to be radically nu or un which is probably a variation of the possessive” (du, ru &e). Malaya- lam has w-de as well as tx-de as composite possessives. The archaic possessive function of w in the former is attested by in of the latter. I would therefore explain the pronominal 1.un and un as contrac- tions of ni-un. In the original system the roots and postfixes were free, and hence the same root admitted different postfixed or postplaced definitives. With the decay of this freedom, the variety in the definitives and the existence of double plurals, gave rise in the coneretionary stage to considerable dialective divergency and some confusion, as in all other pronominal systems using originally several elements for the expression of distinctions in each person. In the closely connected Southern dialects these variations are very marked, and in the Northern they take a still more irregular and seemingly capricious character. In the Tamil 1st persou we find the concreted forms yan, nan in the singular agentive, but in the singular possessive ena or en with the corresponding plurals possessive ema and nama. (I omit the poss. postfixes -du,-de,-di &c.) In the 2nd person we have ni both in the agentive and pos- sessive of the “‘ Ancient” dialect, but in the “ Modern” un or una in the possessive, corresponding with the plural possegsives in both numa Ane. (the full form), and uma Mod. In disintegrated and conereted systems, the original force of the secondary elements passes away, and hence serviles come to replace roots, one form to be substituted for another, generic definitives to receive a special restricted use, special definitives to be generalised or to be clothed with a new special power &c. Thus in Telugu in the singular the definitive -nw has become concreted with the Ist pron. and -vu with the 2nd, while in Karnataka —ny retains its position in the singulars of both and -vw is plural in both. Hence nivu is “thou” in Telugu but “ you” in Karn. The Telugu plurais are ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. lL equally irregular and cumulative, for the Ist person takes»—mu in addition to the flexional labialising of the root itself (memu), while the 2nd not only labialises the root but adds an ordinary plural definitive (mirw), The poss. presents yet another form of the Ist person na- sing. ma- pl. The sexual forms of the 3rd pronoun show similar changes. The proper forms are va-n, or va-nu mase. va-l, or va-lu fem. and du, da or di neut. pustfixed to the def. But in Telugu -du has become mase. (the neuter being varied to-di). In Karnataka the masc, has become va-m and in Telugu the fem. has become a-me. I have already remarked that Telugu also reverses the ordinary functions of the vowels in the 1st person, e being agentive (ne-ru) and a possessive (na-yoka). Asin Semitico-African and Indo-European languages, the postfixed agentive forms of the pronouns in some cases echo the definitive and not the pronoun.* This is almost uniformly done by Telugu, the Ist person postfixes -nu (from ne-nu), the 2nd person postfixes -vu (from ni-vz), the 3rd mase. -du (trom va-du), the fem. -di (from a-di-, now neut.) and the neut. -thi (from-a-thi). The concreted definitives of nouns show variations similar to those of the pronouns. Some nouns have the same definitive in all the dialects. Some have a masc. postfix in one dialect, and a fem. in another. In the Northern languages the dialectic irregularities are still greater than in the Southern, Gond having for “I” the forms na, nu, no,-an S.; ma, mo, -um Pl,; and for “thou”. im, ni S.; im, mi, me Pl. Male and Uraon have similar varieties. Male en “ 7’, ong poss. Sing., na-m, o-m Pl,,em PI. Posse, Uraon en-, eng- «7, em-in Pl. poss. The Northern forms in o resemble the Todava one, on, won Sing. om, wom. Pl, Todava frequently replaces the a of other Southern dialects by o (e. g. “eye” kon Tod., kan in the other vocabularies ; “milk” por, for pal; “six” ore, for aru). ' In the Kol dialects the Dravirian roots are still further con- fused. The foreign affinities of the Dravirian pronouns, are of two classes, the first embracing those indicative of an archaic extension of the formation beyond the present Dravirian province and the * The Kol le “we”, be “ you” are examples of the plural particles taking the place of pronouns, 32 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. second being of a primordial character and pointing towards the derivation of the formation itself. | The pronouns clearly indicate an early prevalence of the archaic Indian formation over Ultraindia and Asonesia, and the forms in which they are found in these regions show that the proper South Dravirian varieties are the oldest and purest, and were first and farthest spread to the eastward. They are found in all their inte- grity throughout the Australian sub-formation,*—the most ancient in Asonesia—and fragments of them are also preserved in other Asonesian provinces. The dialectic Kol system, on the other hand, is found in its integrity in the Mon-Anam formation, the oldest that is extant in Ultraindia, while it is also partially traceble in Asonesia. The Australian pronouns are nga “I” and ngin, nin, ngi thou,” with postfixed definitives as in Draverian,-nya,-ni,—ngi, -na, -te, -toa, -du, -pe, -i. Comp. the Drav. mu, -na, -n, -vu, and the common noun definitive postfixes, The common form of the 2nd pronoun, nin, is the Dravirian root combined with the contracted Dravirian postfix as in Karnataka, Kurgi and Male nin. In the Australian system the plurals are formed, like the ordinary Dravirian ones, by the plural postfixes, the Dravirian special m plurals being absent unless they are represented by -wa. Australian has a distinct dual formed by a Draviro-Australian plural particle -li, -le,—dli, -lin, Se Ist pron., -rang,-ra,-rle. &e 2nd pron. The 2nd has also-mwa and the id eg -wala in some languages. The Tobi nang, Ulea ngang, Pelew nak, Banabe nai, Tarawangai, [Austr. ngai], Rotuma ngo, ngou, and the Sumba nyungga of the Ist person, with the Onni ono, Tarawa ungoe, ngoe of the 2nd person, are also Draviro-Anstralian. The indication of sex in the 3rd pronoun distinguishes Tarawan and Australian from the proper Malayu-Polynesian languages and is one of the traits that connect the archaic pronominal system of Asonesia with the Dravirian. Although the plural forms in m are absent in Australian, it has absolute as well as relative forms of the plural of the first person. * The first indication of resemblances between the Dravirian and the Australian pronouns is due to Mr. Norris, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 13 In some dialects the former are produced by the union of roots of the Ist and 2nd persons, The latter is represented by the dual forms only. In the Malayn-Polynesian languages the two plurals and also the dual are found, and as they are not now Malagasy, although found in Semitico-African languages, they may be Dravirian traits. In some languages the dual and relative plural are not distinguished. The general character of the most ancient Asonesian pronominal system—as preserved in various degrees in the Australian languages, in Tarawan, Vitian, Tanan, in Polynesian and in some of the less impoverished Indonesian languages—is similar to the Dravirian, but it is more archaic, more complete and less concreted. The different elements are more numerous and more freely and regu- larly combinable. In the Australian system we find not only all the forms that are now extant in South Dravirian, as well as the dual and the pecaliar transition or agento-objective forms of Kol, but several others produced by the same power of compoun- ding elements in which these originated. This power is much less impared in Australian and the allied Asonesian systems, anil the inference is that in this, as in several other respects, they better preserve’ the archaic Indo-Asonesian type, and may hence suggest to us what the condition of Dravirian itself was before its forms had become diminished, confused and concreted as we now find them. In Avstralian the pronominal roots are compounded with definitives, singular and plural, with the numeral “ two’ to form duals, with mase. and fem. definitives in the 3rd person, and in all the 3 persons with each other, thus produeing not only abso- lute and relative plurals of the Ist person, but several other com- plex plurals. The Viti-Tarawan elemeats are still more freely compounded and their forms of this kind are consequently more numerous. The incorporation of numerals appears not to have been confined to “ two,” for in some of the Papuanesian languages a trinal is found, and in Polynesian the same form has lost its original meaning and beeome a generic plural. This highly agglomerative but erude pronominal system bas not been derived from Malagasy, and its presence in Asonesia is attributable to a prior formation, of Indian origin, similar to the Dravirian but 14 ETHNOLOGY OF TIE INDO-PACIFIC ISLANDS. more rich in forms because simpler and less concreted. It thus earties back the Dravirian type to a condition analagous to the American. To illustrate these remarks by going into details would be to anticipate so far the ultimate aim of our examination of Dravirian and the other S, E. Asian formations, and I must therefore refer the reader to the subsequent section on Australian. The merely glossarial connection between the Dravirian and the Australian systems embraces the pronominal roots, several of the agentive postfixes, plural postfixes and perhaps some vocalic flexions of the roots. The 2nd pronoun inseveral dialects changes its proper vowel ¢ to wu in the dual and plural, In some the a of the 1st pronoun becomes e in the plural, In Australian as in Dravirian and other compound agglutinative and partially con- creted systems, the pronoun is in some forms replaced or repre- sented by other elements, definitive, numeral Kc. The sexual distinction between the definitives m and is not found in the known Australian languages or in Tarawan, The Ist and 2nd pronouns do not take sexual postfixes, a fact telling against any surmise that Dravirian may have had them in an early stage. | The North Dravirian pronouns evidently preceded the Tibeto- Burman in the Mon-Anam languages and in Ultraindia generally. They are preserved in the pre-Malayan basis of the languages of the Malay Peninsula—Simang as well as Binua—and they have also spread to the Eastern Islands. The most common form of the 1st pronoun is similar to the Kol ing- with its variations eing, aing, inge—which is a liquid modification of the prevalent South Dravirian possessive en, occurring also in Uraon (eng). Both the Southern and Northern Dravirian en, eng andthe Kol form ing, which is probably the original, are dispersed amongst the vocabula- ries of South Ultraindia and the Malay Peninsula,en Simang; eng Chong, Kambojan; eing, ein, ye Simang; ain, oin, yan Binua; oci, oc Mon. In Indonesia the North Ultraindian form is perhaps found in Sunda aing, but this may be a Niha-Polynesian prefix with the irne pronominal root elided, The Timor ani and Kissa ba-nian are probably connected with it. The Sumba nyu-ngga is South Dravirian and Australian in form, but Gond has nu-na. The prevalent. Niha-Polynesian forms of the Ist pronoun are not Kol. ETHNOLOGY OF TITE TNDO-PACIFIC ISLANDS, 15 The Kol 2n]1 pronoun—which is mach more persistent and widely spread in the Mon-Anam languages than the Ist—is very rem ukable, and at first view anomalous, in its form. It is a labial, occurring under the forms imma Gon (agentive), am, um, umge, me, m, be, pe Kol. In the Himalayas the Kiranti am of the possessive am-ko is the only example of this root or form. Tn Ultraindian it is Mon puch, pi, bai; Kasia, me, pha; Anam, mei; Lau, mung, mau, mo; Chong bo; Simang, mo, bo; Trawg- ganu mong. The form ts rarely found in Asonesia in the agentive singular, which in the Niha-Polynesian languages, is, like the Ist pronoun, of Semitico-Libyan derivation through Malagasy. In the Timorian group,—which preserves the N, Dravirian Ist pio- noun in some of its languages and has other N. Dravirian traits— we find in the singular mue Solor, nyu ma Sumha, (nyu definitive as in the Ist pron. myu-ngga, which is-also Dyvavirian). It is common in the Niha-Polynesiin languages as a possessive under the forms mo, mu sometimes mi. It is found in the plural, either by itself or combined with another particle. It also enters into the exclusive or relative plural of the 1st pronoun. * In the N. Ultraindian and Mon-Anam languages it is excep- tional as a root for the 2nd pronoun, none of the pronomiaal systems of the formations with which they are connected, or which are found in Eastern Asia, using a labial root.t The nang, neng of the Chino-Tibetan and Ultraindian system is variable in Burman to mane, meng, but this mutation of the n of the root is confined to it, and its absence in the adjacent dialects of the same sub-formation, the forms of the pronoun in the conterminous Mon, and the recent Ultraindian spread of Burman even when compared * But as the m element may in some cases be the so-called companionative or may be a direct engraftinent from the Dravirian plural of the Ist pronoun in mi &e, it is enough at present to remark the prevalence of mu, mo, wias a subsidiary root for the 2nd pronounin Malayu-Polynesian, That as such, it is a Dravirian or Draviro-Uitraindian eugraftment on the Malagasy-Polynesian or Oveanic sys- - tem is clear from its being absent not only in Malaya-y but in the present Seinitico= Libyan system. + Although I consider the explanation in the text the correct one, it should be rernarked that several of the Uliraindian forms of the Seythico-Australian labial third pronoun and definitive have a close resemblance to varieties of the Jabial second pronoun, and that in some formations these two pronouns involve the same definitive. This is the case in Tibetan, Semitico-Libyan, Indo-European, Lesgian and Samoiede. Comp, mung “thou” Siamese, muny, ** He” &e Kambojun; pi *“thou” Mon, ke, pike “he &c.” Kambojan; ma“ he &c.” Dophia; bo‘ thou" Chong, Simang; wo “he &e ” Simang, Newar; bu, Miri e &c. 16 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. with the Naga-Manipuri branch of the same family, shut out the supposition that this accidental form was the parent of the archaic- ally diffused Mon-Anam, Vindyan and Asonesian pronoun. As the latter is neither Chinese, Tibeto-Ultraindian, Tatar, nor Mala- gasy, we are thrown back on the system to which the Ist pronoun belongs, and the widely prevalent plural power of the form in the Niha-Polynesian languages suggests that it is simply a Dravirian plural used for the singular, as happens in many other languages with the 2nd pronoun (e. g. the English “ you” for “ thou’). In Dravirian we find amongst current forms for “ you” miru Telugu, (midi poss.) where the plural m displaces the n of the root, (ni, nivuw) asin the plural of the Ist pronoun, e. g. Telugu nenu “1”, memu “we,” Toda an “1”, am “we”. Tamil emadu ; “ours”, umadu “yours”. With these compare the Khond anu “7”, amu “we”; inu “thou”, mi “ you”, the last term being identical with the Telugi mi of mirw. The nearest South Dra- virian forms occur in Todava ni-ma PI. and Karnataka ni-m P!., (Anc.), ni-vu (Mod.); ni-ma-du Pl. poss. (Mod.) The Gond i-ma is evidently a cowtraction of ni-ma. Tn some of the southera forms also, the root of the second pronoun is represented by the vowel only. The Kol variations of the proper radical yowel i to u and a are found in-some of the southern languages. .In Kol the singular forms are wm (as in the Tamil PI. poss. umadu) am, (as in Toda nama PIl.), me, m, variable in the plural composite terms to ba (comp, 8. Dravirian vu), be, pe. The connection between these and the Ultraindian mo, bo, pi, mong, mung &e, is obvious. The Telugu verbal postfix of the 2ud pronoun -vu exhibits the same substitution of the plural definitive for the pronoun. In the Semiticu-Libyan system, in which m has a plural power as in Dravirian, like examples occur of the fepiees- ment of the root by the plural particle. The Kol le “ we” is another example. } Of the Kol forms um, bu-am, me, (be, pe)—corresponding with the South Dravirian um, vu, am, mi—the first is the most widely spread in Ultraindia and Asonesia in the forms mu, mo, bo, mung &e. “In South Dravirian itis rare, but its oceurrence in the possessive plural of Tamil (um) and in the plural of Karnataka (vu) places its Dravirian origin and antiquity beyond doubt. & a) ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 17° The distinctive vowel u is found in the Tamil singular also (un). * Dravirian pronouns and pronominal traits are also found in the Gangetic and North Ultraindian languages. But as the Tibeto- Ultraindian pronouns are themselves radically the same as the Draviro-Australian, and as this radical agreement belongs to the most archaic dian affinities of Dravirian, it will be noticed in connection with these. For various examples; of Dravirian traits in the Gangetico-Ultraindiag systems I may refer tochap. TV - Here I shall only mention one, as it is illustrative of the archaic use of m as a plural definitive. The Naga pronominal system—which is a Tibeto-Burman super; structure on a Drayiriah basis—preserves the, Dravirian plural postfix in Namsangya ni-ma “we,” ne -ma “you.” The possessive of the Ist pron. sing. and pl. is i (from ni “I,” originally posses- sive now replaced by the Tibeto-Burman nga as a separate agen- tive term) but that of the 2nd pron, sing. as well as pl. is ma (from ne-ma). In Tengsa Naga me oecurs as the 2nd pronoun in the possessive mechi,—the separate form being the common East Tibetan nang. In Joboka Naga m is retained as the plural postfix although the roots are changed, Ist ku Sing. kem Pl.; 2nd nang Sing. hanzam Pl,; 3rd chua Sing. hom Pl. It will be remarked that while Namsangya like the Dravirian languages in general restricts m to the proper pronouns, Joboka like Tuda extends it to the 3rd also. The only other Gangetico-Ultraindian language in which this particle appears to be found is the Gurung which has it in all the three pronouns under the form -mo, The Newar -ping is probably another variety of it. The Mozome Angami. Naga -we of the Ist pronoun resembles‘Kol forms. In Angami ma appears to be combined with the liquidgplural particle of Dra- virian in all the pronouns -ra-ma. In Gayo mong (comp. Gurung mo) and ma occur as plural elements, and the Burman labial plural may be the same particle, * (Prof. Max Muller’s table of prononns, supplies two additional examples of the use of this form, In the Malabar dialect of Malayalom, the online Ben ct the singular ig um-(with postfixes), while the plural has both un and.um. In Brahui the nominatiwe plural is num (oblique numa), It ig abundantly evident that both ni and nu must have been current as forms of the sofia d pronoun from a very remote era of the Druvirian formation, and that the Kol forms and their Ultraindian ¢erivatives, so far from being really exceptional, are more distinctively and undoubtedly Dravirian than they might have been, considered had they adhered to the common aye forms, of the South, and thus resembled the Tibeto-Ultraindian forms with which they are intermixed in several Gangetic and Ultraindian languages. | o 18 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. The second class of pronominal affinities appear to appertain to the more archaic or pre-Indianhistory of the Dravirian or Draviro-Australian formation. They are very numerous if those of each pronoun Be considered separately. But the formations which have both of the’ Dravirian pronouns are much more limit- ed. The simple roots,are found best preserve di Chinese wnd in some of the more archaic or preScythic languages of America. The only other system in which@oth occur as the principal terms is the Tibeto-Ultraindian. These facts and the distribution of the different varieties of the roots in these and in other formations, lead us to the conclusion that the system is probably the most archaic and least mixed that is now extant. The Draviro-Austra- lian forms {stand in the same rank as the American in relation to the Chinese. Like American and proto-Scythic they belong toa secondary, harmonic, and post- positional formation, and not to a, primitive and generally preposi- tional one like Chinese. They have definitive postfixes like Ame- rican and Scythie and the full terms are in structure more imme- diately allied to the Scythic. The three formations stand on a similar footing in relation both to the primary Chinese formation and to the earliest harmonic development which it received. As regards the roots in particular, the Draviro-Australian na or iiga and ni or ngi have a more direct and complete affinity with the Chinese ngo and ni than the pro- _nouns of any other system. The adjacent Tibeto-Ultraindian* system is also Chinese and the Ist pronoun has the Draviro-Australian vow el a, which appears to have been early and widely prevalent, for it is found in some American languages (nai, nan &c), Korean (nai, na), Samoiede * In chap, IV I considered the original or integral Gangetico—U!traindian pro- noun system to be fundamentally Dravirian and distinet frum Tibetan, although different languages present modifications and iitermixtures, Thus the Naga was held to be a compound of Burma-Tibetun, South Uitraindian and Ganzetico- Draviriun traits. ‘The remarkable extent to which the roots and forms of uilferent formations huve been blended in the Ultraindian systems w ll appear when we examine the pronouns of the Mon-Anam or prepositional alliance. ‘lhe publication of Mr Hodgson’s East Tibetan or Situn vocabularies has not affected the general inferences at Whee had arrived, but they have made an importent modification in details, The 2nd pronoun ino [ considered to be Dravirian in all the Gangetic and Ultraindizn languages in which it occurs, the Tibetan root being totally diffe - rent, It now appears that the East Tibetan or Sifan 2nd pronvun is also a form of the n root, simimar to forms found in Ultraindian and Gangetic ae weere that have numerous other glossarial affinities with East Tibetan, In the text I have iniro- duced the necessary modification of my former view. - ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 19. (na, but this is probably a variation of the Scythic ma), Caucasian (na, Kasi Kumuk), and Semitico-Libyan (na, also no, nu, ne, ni, that is, all the vocalic varieties of which instances occur in Chinese, Dravirian &c.) The Tibeto-Ultraindian Qnd pronoun has also the broad form nan, na (the West or proper Tibetan has a different root), thus directly connecting itself, not with the slender forms of the adjacent Chinese and of Draviro-Anstralian, but with the archaic Scythic nan, na (Ugtian). The numerous Ugrian and other Seythic and N, E. Asian affinities of the Tibeto-Ultraindian vocabularies render it probable that this form of the 2nd pronoun is of archaic Ugrian-origin. The Dravirian slender ¢ form and the u form are also Ustian, ny, ny, nvagi, nyn, num. The affini- ty between the Ostiak form nyn and the Draviro-Australian nin is obvious. The nasal second pronoun is not the prevalent Scythie, Indo-European and Semitico-Libyan, form, which is im ¢, s &e. If the Seythic m of the 1st pronoun was an archaic variety of n— which is foand in Scythic, but as a flexion of m—the demonstra- tion of the affinity of proto-Scythic, with American on the one side and with Draviro-Australian and Tibeto-Ultraindian on the other, and of the derivation of the common roots of all from the Chinese formation, would be complete, Although it is clear that the Dra- viro-Australian pronouns are not derivatives from the Tibeto-Ultra- indian, but are to be considered as having like them an independen; connection with an archaic Mid-Asiatic system—Chinese in roots and Seythie in form—it necessarily happens that the forms of the common roots sometimes so closely resemble each other that it is difficult to say what their true origin is in certain of those Indian languages which are placed at the junction of the two formations and have other affinities with both. The@ibeto-Ultraindian nza of the 1st pronoun becomes in different languages ngo, ngai, (comp. Chinese ngei) ngi, nge, nye. It is distinguished from the full and more prevalent Dravirian form, not so much by the liquid nasal (i for n) which is also Malayalam, Kol and Australian, and appears from Chinese to have been the primary form, as by the absence of the definitive postfix. But the coutracted an. slender Dravirian varieties an, en, eng, ing are little distinguished from Tibeto-Ultraindian forms such as ngi, nge, nye, and it thus becomes difficult in all cases to decide whether varieties like the Mikir ne, — 20 EPGNOLOGY OF THE INDO-PACIFIC ISLANDS. Naga ni, Bodo and Garo ang, anga &c, are Tibeto-Ultraindian or Dravirian.* The chief distinction between the Tibeto-Ultraindian and the Draviro-Australian systems consists in the combinations, aggluti- nations and flexions which are found in the latter. But there is also a real difference in the forms of the roo. The proper form of the Ist person in Tibeto-Ultraindian is still nga, This was no doubt the original Indian form also, but from a remote period in the history of Dravirian as an agelutinative formation, modifica- tions of this form have prevailed, the principal being na, ne or en, ing and the contractions e andi. When East Tibetan languages tame under the influence of Dravirian phonology similar forms might be produced in them, but in general such forms appear to be of true Dravirian origin. It is not at all probable that so great a transformation as that of nga into i took place in any purel Tibetan language, while the archaic prevalence of ¢ in Dravirian and its original identity with ‘the e of en, eng are certified by numerous facts in different languages. When therefore we find in the obviously compound Naga system, with its flexional Dra- Viriah traits, not only the true Tibetan forms nga “1” and nang “thou” [Gyarung Ist nga, 2nd nan-] but in the plural Ist ni and 2nd ne, and in the possessive Ist i and 2nd ma, there can be no doubt that ni ‘and i are remnants of a Dravirian form of the Ist pronoun similar to the oblique South Dravirian, to the Kol and Limbu, and to the allied forms found in the older or prepositional languages of Ultraindia. Other Gangetico-Ultraindian examples * The caupenire table of the Dravirian pronouns will show the great difficulty of distinguishirg between the Dravirian and the Tibeto Ultraindian terms, I am if no means satisfied that the classification is correct in all cases. Some of the Himalayan and Ultraindian forms are, in mere phonetic form, as much al- lied to the southern as te the northern group. The principal facts that have guided me are these, The southern forms of the Ist pronoun in i, ¢ have been cae by the incorporation of the sive particle i &e. with the pronoun hey are consequently noes eae n the singular. The Tibeto-Ultraindian forms in i, have been produced by the incorporation a *cythic ant East Tibet n plural particle, ni, i, Ges Horpa) with the pronoun, as is évident from this par- ticle remaining as a postfix in several languages. The Himalayan and the allied Ultraindian forms in i are gt vm found regalarly in the plural only, Hence I consider the singular ninga Milchanang, inga Milch., Limbu, to be allied to the Dravirian ing, eing, eng, en &e. and not to the plural Garo ning ; and the plural ni Serpa, ani Limbu, ain Kiranti, in Murmi to be distinct from the singular ving Ho, ain Binua &e. A few forms in e, obviously Tibeto-Ultraindian (‘Takpa, Kinawari Tibetan, Mikir) are attributeb'e to the pore Dae tendency to re- place a by e, found in some of the ‘Tibeto-Ultraindian uages, as is more full noticed in the next chapter. Poin, . ETHNOLOGY OF THE 4NDO-PACIFIC ISLANDS. 21 of Dravirian forms oceur in the Milchanang and Limbu inga (identical with Kol and Mon-Anam forms), Garo ning and Sing- pho i (both Pl.), Mikir ne, Nagaung Naga nyi, Khari ni and the Naga forms noticed in a, previous page.* The Angamia [Manyak also] and the Gangetico-Ultraindian ang (Bodo, Garo), angka Kiranti are probably East Tibetan. The 2nd pronoun is more strongly distinguished in the two systems by its radical vowel, which in Draviro-Australian is i as in Chinese, while in Tibeto- Ultraindian it is a, as in some of the archaic Seythie forms. The Ultraindian mewbers of the Tibeto-Ultraindian family show other Dravirian affinities in their pronominal systems besides the occa- sional adoption or retention of Indian forms of the roots. Dravi- rian plurals, possessives and other particles occur in several lang- wages, Bodo, Dhimal, Naga &e (see chap. IV.) Not only the @ommon Dravirian plurals in | &e are found, but, as we have seen, the pronominal m. ¢ Amongst the primary affinities of the §. E. Asian languages and Dravirian may be inclnded the plural m and the possessive in é, ni &e, The formeris Chinese -mun, -men, -me, -mei, -pei and the latter is Tibetan (yi), Manyak (i), Burman (i), Limbu (in), Bodo and Garo (ni), as well as Seythic, Semitico-Libyan (i) Zim- bian (i) &e. The Chinese traits in the Himalayan and Ultrain- dian languages present great difficulties. Some are of compara- tively recent East Tibetan origin and in Ultraindia even more modern. Others appear to belong to a connection as archaic as that between Australian and Chinese roots. The Draviro-Australian or archaic Indo-Asonesian proninalom system with its numerous distinet elements and combinations, appears to be more ancient or less impaired than most of the sys- tems of other harmoni¢ formations of the Old World. Trom its general structure it must be ¢onsidered as cognate with proto- Scythie or Scythico-American. It is richer than Seythic, which has neither sexual forms nor any plurals save the “rdinary generic ones, with the absolute “ we” (formed as in Dravirian), although the Seythie power of combining such elements as the formation possesses is similar to the Drayiro-Australian, and the position of the subordinate definitives isthe same. In some of the * [ Brahui i. ] . 93 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Seythie languagesa dual is found (ante vol. viii. p. 70),* and as it is preserved in Kol, Australian &c. it was probably common to the archaic Seythic and Indo-Asonesian systems Double plurals oceur in Seythic as in Dravirian. ‘The transition forms of Kol and Australian are absent, but the Fin reflexive forms may be considered as analogous remnants of an earlier and richer condition of the Scythic system, when it had departed less from the Semitico-African types on the one side and the American on the other. The Scythic, Caucasian and Africo-Semitic habit of post- fixing the pronoun possegsively is preserved in the Kol pronomi- nal postfixes to names of kindred andin the ordinary Dravirian persons of the verbs+ (ante vol. yiii., p. 58). The Scythic postfixed n definitive of the singular is Draviro-Australian. The plural definitives in J, r, are also common to the two formations } but the regular m plurals—flexional and postfixual—of Draviriam are not Scythic. In some Ugrian languages the Ist pronoun has the n form in the singular and m (the root) in the plural, the former being evidently the definitive postfix left on the elision of the root ; and as m does not occur in the plural of the 2nd pronoun it cannot be considered that the Dravirian plural m has any Seythie affinity.§ The Scythic plural def. k is found in Gond (-A, -nk, -g) and Kol (ko). In the Gond pronouns, as in» some substantives, it is common and this is also the casein some Usgrian systems (nank Wogul) and in Semitico-Libyan. . Com- bined with the 1, r plural it is found in most of the Dravirian languages (—kal, -gal, —-kulu, -kan, &c., so -galai Dhim., -khala, * [It would appear that the dual is not limited to Lap for according to Castrén it is found in Ostiak and Samoiede also. It is formed by the guttural postfix ga, ka &e., which Castrén derives from ka or ki “also.” But is it not identical with the plural cuttural particle (ante vol. viii, pp. 66, 70)? “* In the Irtishian dialects of the Ostiakion, in Lapponian and Kamassian nouns and adjectives have lost the dual, and pronouns and verbs only have retained it. In the Samoied-Ostiakian it is the pronouns that have lost the dual.” Prof. Max Muller in Bunsen’s I’ hilo- sophy of Universal History, vol. ii, 461}. + The Asonesian habit-of postfixing the pronouns possessively ic mainly referable to a Semitéeco-Libyan source, through Malagasy. + The Dravirian plurals in nar, mar, are probably connected with the Seythic nar, far (Mongol, Turkish). ‘Those in ra, la are also African, and in Asonesia are thuscommon to the Drayirian and to the Malagasy derivate forma- tions. § The’ traces of a Inbial plural in Seythic are too obscure to be relied on. The piuee postfix, also dual, and in the Ist pron, -ne, -7 is dunl. These phonetic aftinities with Kol do not appear to me to indicate any glossarial eonveciion. The dual n of the Ist pronoun is evidently the ordinary Ugrian fletion of the pronominal root m, or it is the definitive left as its representative on contraction as in the Hungarian separate fori (en). Lap. hus p, o, rsa ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 93 ~kara Naga, combinations resembling the Samoiede, N. E. Asian and American gada, ganda &. The Seythic systems in their vocalic flexional plurals and some other traits, are rather Indo- European, Semitico-Libyan and Zimbian than. Dravirian in their affinities. But we have seen that Dravirian has some traces of yocalic flexion in the change of the agentive a, o of the Ist pro- noun to e in the possessive, and in that of i to u in the 2nd.* The two systems cannot be referred to the same formation, and the aflinities, great as they are, must be considered as collateral. They point to a common source, to an archaic postpositional formation at once more crude and more redundant in forms and combinations than Ugrian, Drayirian or even Australian. The Indo-European system in its possession of a dual number e of sexual definitive postfixes and flexions which extend to the rd pronoun, but not to the Ist and 2nd, resembles Draviro- Australian in some of the characters in which it is richer than Seythie. Dravirian in its retention of the sex distinction in the 8rd person of verbs is less abraded than Indo-European. In other respects the latter system is, in its basis form, analogous in roots and structure to the Scythic, although somewhat richer, and has no general affinities with -Draviro-Australian save what are observable in Seythic. It is more concreted and flexional than either, although similar flexions and irregularities occur in all three. The Semitico-Libyan system like the Indo-European, has dual and sexual elements, and in the latter it is richer than either, for it uses them with the 2nd pronoun, and there are even traces of them in the first. ‘The union between the pronominal elements and words used assertively, is more complex than in Indo-European or Seythic, as it has objective or transition forms like Draviro- Australian. The root of the Ist pronoun is Draviro-Australian, but that of the 2ndis not. The postfixed definitive k of the 1st person assimilates the term to the Gond forms in k (nak &c.) The Gond -k although now used in the singular is properly plural and Seythic, while the Semitico-Libyan is generally singular and probably masculine, but in Hottentot it is plural both in the 1st bd Ucro-Fin definitive of the singular changes its vowel to % in the obliqa casa. ins, mi-nu). This nay. be related to the Dravirian change of the root-vowel i to uw in some dow al and plurals, In many of the Semitico- African langsuges u is plural. . 2 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. and 2nd pronouns. The fact of both formations having # as a plural, * i (variable to ¢) as a possessive, and u as a plural element ean hardly be accidental, but the affinity belongs to the most archaie period in. the history of the two formations, like others that will be noticed afterwards. The common radical elements, with the agglutinative and flexional tendencies under which both formations, have been developed, have [produced several coin- cidences amongst the various forms which have concreted in both. Thus the possessive i ore represents the lst pronoun in several Semitico-Libyan languages as a verb postfix or prefix. The Mahrah pl. of the-1st person abu (comp. Hausa mu) is similar to Dravirian, forms’ (abu pl. absolute of Kol, &c.)+ The Dravirian formation has madical affinities with the archaic ones of S$. W. Asia, where it departs’ from S. B. Asian and Seythie in roots or ‘forms, and although these identical pronominal ‘terms have been independently formed in ‘both formations, thereoihcidence cannot be considered as purely accidental when it rests on a communivy of roots and, to a certain extent, of ideologic and phonetic tendency also, Brat HEY The Caucasian pronominal systems preserve affinities to those of formations in nearly all the great stages of development. The roots are varied and mixed. .The Iron in, an, on sing. of the Ist pronoun and the Kasi Kumuk na are not Seythico-Iranian but Semitico-Libyan, and Draviro-Australian. The plural ma, am, ab ‘is also phonetically, Semitico-Libyan and Dravirian, but it does not occur in the 2nd pron, and is probably Seythie glossarially. The root of the 2nd pronoun di &e, is ultimately a variety of the Chino-Dravirian ni but more immediately connected with Scythie ke, Caucasian has transition forms and attaches the pronoun possessively and assertively to other words, but it wants the com- plex duals and plurals as well as sex definitives or flexions, The nature and historical import of the ,aflinities between Caucasian and Dravirian are considered elsewhere. * Aas a definitive pootiix the labial is #euter and sometimes feminine in Dravi-~ rian ws in Sanskrit. In Semitico-Libyan it is plural and mage: + The Semitic plural and dual(Arahic) w definitive may possibly be connected with the dua! n of Kol. In Semitico-Libyan languages it is variable to d, 1, r, nd, nt, &e. and appears to be radically the same as the Scythic and Draviro-Austra- lian plural element in Ll. r. In Gangetico-Ultraindiun languages li becomes di, ni, nin, ning, &e. Horpaalso hasni. In the purer Scythic languages the redu- Heated lar, ler, of Turkish becomes nar, ner in Mongol. WN forme ure also found n Yeniseian (n, ng,) and Yukahiri (l,2,) (ante, vol. viii. p. p. 55 56.) ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. ' 2 Euskarian in the Seythic, Caucasian and Semitico-Libyan characters of its pronominal system has necessarily Draviro- Australian affinities also. The transition tendency is archaic Semi- tico-Libyan, Georgian, Zimbian, Australian, and American. The roots are varied and m is found in both the Ist and 2nd pro- nouns, but wighout indicating any special connection with Dra- viro-Australian. The Zimbian pronominal system is in many respects even richer than the Australian, but the roots are Scythic and Caucaso-Yenis- eian combined with Semitico-Libyan and numerous as the general affinities are, there are no special ones with Draviro-Australian. It is only in American that we find examples at once of a free and multiform combination of pronominal elements similar to the earlier Draviro-Asonesian, and of systems which, with this archaic © richness of terms, preserve the Chino-Dravirian roots. In some American languages the extent to which pronouns combine with each other and with:different definitives is still greater than in the outlying or insular members of the Draviro-Australian formation. As traces of a similar primitive freedom of combination, are found in most of the harmonie Aso-African systems, it is probable that a pronominal development analogous to the American was the ultimate source of the Scythico-Iranian, Semitico-African &c. and that the remotest and most sequestered branch of the Indo-Asone- sian formation has remained more faithful to it than the exposed continental systems. The close connection between the general structure and ideology of the Dravirian aud Scythie formations and the large glossarial affinity give additional importance to the fact that the Dravirian prououns are not the predominant Seythic ones. The prevalent Scythic Ist person is a labial, ma, mi, bi, &c, and the 2nd a dental, sa, si, tt. These are Iranian and Caucasian, the Ist being also found in Zimbian and the 2nd in Semitico-Libyan, N, E. Asiatic, and American languages. Both are evidently very archaic, but their diffusion over the Iranian, Scythic, and connected African area must have been later than the spread of the Draviro-Austra- lian and allied American terms which centre in the Chinese. The fact of the latter being found in widely separated and outlying ethnic provinces—Americu, N. E. Asia, Africa, S. India, Australia— D 26 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. combined with that of the Scythic and Iranian being the latest of the great migratory races, establish a high antiquity for the move- ments which dispersed the Dravirian pronouns on all sides from their probable centre in S. W. Asia, The Chinese is probably the most ancient integral formation to which they can be referred. They appear to have been diffused over a large portion of Asia and Africa as well as over America prior to the rise of the dominant historical races, and their spread over India, Ultraindia and Asonesia in the era of Draviro-Austra- lian civilization, now represented by the Australians, throws light on the ethnic condition of 8S. W. Asia at the period when a civili- gation of this character was connected with the most influential and diffusive formation, The roots only are Chinese. The Dra- virian and Asonesian forms of the pronouns shew that the lan- guages of this formation had already acquired a harmonic and postfixual character. The preservation of the same roots in Ame- rican, N. E, Asiatic, Seythie and African languages and the gene- rally Scythic structure of Dravirian, lead to the inference that they were associated in Upper Asia with an ideology of the Scy- thie kind before they spread to India and the farther east. The general conclusion is that the Draviro-Australian pronomi- nal system is not an offshoot from Scythic proper or from any of the other Aso-African systems, but is a remnant of the proto-Scy- thic era of the harmonic development, and a link between the Scythie and American ideologies and between Chinese and Ame- rican. In American the crude and pleonastic ideology of the early monosyllabic stage is preserved under a harmonic and agglo- merative phonology. In the Australian edndition of Draviro- Australian the pronoun system retains the same combination to a large extent. Traces of a similar crude and elaborate system are found in the other Aso-African formations, and they all present evidences in flexions, contractions and irregularities of different kinds, of having fallen away from a condition more elaborate and consistent in terms and forms. Although Scythie is amongst the most decayed and simple of these systems, some of its members which retain other American traits also, are possessed of vestiges of such a condition, while its affinities to Indo-European and other systems which preserve similar and more numerous vestiges, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 27 and the highly agglomerative character of the formation, leave no doubt that in one of its early stages the proto-Scythic pronominal combinations were as crude and numerous as the American or Australian. The Draviro-Australian system may be considered as proto-Scythic in its general structure and character, for even in Upper Asia that type is not limited to languages which posgess the proper Seythic pronouns. B. Definitives, (including 3rd Pronouns, Possessives and Directives. ) The possessives and directives are merely definitives, and as most formations possess nearly the whole range of archaic defini- tives, the comparison of isolated applications of them can seldom Icad to specific ethnic results.* The Dravirian and Australian labial definitive paz, wa ke is Tibeto-Ultraindian, Seythic, N. E. Asian (Kamschatkan), Cau-~ casian, African, Celtic (Welsh ve, vo &c); and it passes into ba, va, ma,am, um &e &e.+ In Tibeto-Himalayan languages it has a qualitive power, which is not found in Dravirian. } Ta, da (with vocalic variations) is almost universal as a defini- tive, and it passes through the surd form into ka, ga on the one side and through the sonant into la, ra, na on the other.§ In the form ni, in, &e it is the principal Dravirian possessive, and this is probably identical with the Tibeto-Ultraindian and Scythic poss. ni. The nasal possessive in, yin, m, ¢ kc is also Semitico-Libyan and Indo-European. The Turkish reduplicated forms nin, nun, nen, ning are found in Asonesia as well as the simple form ni. The same particle occurs in some of the Irano- European languages as an archaic possessive, as in the 2nd per- sonal pronoun in Zend ma-na (in Sanskrit euphonically ma-ma), Gothic mei-na, &c. In Guzerati both the simple and reduplicated cna Thee She bau act ce ane nwo Tie ld ty also refer to the same place for examples of the wide prevalence of most of the definitives found in Dravirian, and for indications of Scythic affinities. ' + See vol. viii, p. 63. ¢ Bot Gond exceptionally has wa, Ist pron. no-twa sing., mo-wa-n pl., 2nd ron. ni-wa sing., mi-wa-n pl, This form is Tibetan through Gangetic. Magar {st pron. ngo-v, (root vowel of nga modified by that of postl.); 2nd nu-100 (1D) 3rd, hoch-vz (nom. hos.) In the plural the full form um is used, corresponding with the Gurung mo, the latter however being plural not simply poss, § Ante vol. viii, p. 62. - 28 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. forms are found, ni, mo, nun. The same possessive is found in several of the Gangetic languages including Limbn, (ix )and Bodo (ni).* Of the other Himalayan possessives, the most common, found also in Male and Uraon, ko, ke, gi, &c is Tibetan and Chi- nese and the rarer ti, chi, so, sei, sa &c is Chinese (ti, chi &c), (see the Table). The transitive use of ka, hu, tu, du is very gencral (Iranian, Trano-Gangetic, (Hindi &c), Siamese, Chinese, Scythic, African, Asonesian &c, including Australian), But nearly all the defini- tives are so used. The Dravirian za, an, nu &e used as a definitive with pronouns &c, is applied in the same mode, and also as a def. prefix, in Semi- tico-African and Asonesian languages. As a def.—separate, pre- fixed or postfixed—it is found also in Iranian, Semitic, Scythic and American languages. The form in 7, J, is also common, and in some African, Asiatic and American languages it assumes pecu- liar forms such as id (8. African, Caucasian, N. E, Asian, Ame- rican). The common masculine and inanimate or neuter 3rd pro- noun is the dental definitive da, du, thu &e. The absence of the widely prevalent sibilant or aspirate defini- tive might be considered as a peculiarity of the Dravirian forma- tion, but it is frequently only a modification of the dental as in the Scythie and Semitic 3rd pronoun in ta, sa &e. The Toda athu, Male ath, shews the dental becoming aspirated or half sibilant. In the Uraon as-an (az is a postf.) the change is complete. The Magar hos is the same particle, and in the Sunwar hari it appears to be combined with a different one. The Burman thw, sw and the Murmi the, Gurung and Manyak thi, Naga ate are variations of the same particle, probably of Chinese derivation. The root is so widely spread that it is hardly safe to draw any conclusions as to the relations which its various forms | may indicate. The resemblance between the Dravirian and Chi- nese pronominal roots is completed by the Chinese 3rd pr. tha, tht which however has representatives in most of the formations of the old world. The Dravirian va, we, wu, of the 3rd pronoun (Australian ba, pa), is the same as the common pa, bu, wa, u &c. of the Tibeto-U!tra- indian languages, and in both formations is directly connected * Ante, vol. viii, p, G1. See the Table, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 29 with the N. E. Asian and Seythic labial definitive and assertive. It is also Caucasian, Semitico-African &c. In Dravirian it has a generic personal application, the postfix m rendering it masc. and the postf. 7 fem.,* but it also occurs as a common def. element. The labial as a postf. is usually neuter and sometimes fem. and in Tuluva it is the 3rd pron. neuter. The primitive form was probably the feminine flexion of the labial which also came to in- clude neuter. [See Tibetan, Anam &c.] The objective use of the neuter m is common to Dravirian with Indo-European. It is also objective in Caucasian and Seythic, Besides the dental and labial 3rd pronoun, Dravirian has a vocalic one i, ye, yi, found also in Kol. The more common Kol 3rd pronoun ni is also demonstrative and it is found with both functions and as a generic definitive element in many other formations, Scythic, Africo-Semitic, Malagasy, Asonesian. As a demonstrative it is Scythic, Tibeto-Ultraindian. The vowels are used as definitives in Dravirian, chiefly prefix- ed to other particles. All the 3rd pronouns above referred to take them (a-va, a-van, a-du, a-ye, a-i, i-ni &e. &c). In some Dravirian demonstratives and locatives ¢ has a proximate and a a remote force. Similar applications of the vowels are found in Seythic, Indo-European, Semitico-African, Malagasy, Asonesian &e. The affinities of the Dravirian possessive and directive system are too numerous and complicated to be referred to the influence of any other existing formation. They support its claim to an inde- pendent place amongst the most archaic of the harmonic forma- tions. The general character both in roots and structure is Scythic but with a leaning in some points to Semitico-Libyan and Caucasian—which again are Scythic in many fandamental traits. The Tibeto-Ultraindian affinities in roots are in general coinci- dent with the Scythic or with Chinese. The indeterminate and variable functions of several of the definitives have been adverted to in connection with the pronouns. A similar confusion takes place in all agglutinative languages in proportion to the number of well separated dialects that exist or to the force of those causes that evolve dialectic changes in each * Traces of a similar archaie application of these postfixes are found in Cau- casian. 30 ETHNOLOGY OF TIIE INDO-PACIFIO ISLANDS. language with the progress of time. In the Semitico-Libyan and Indo-European formations we meet with facts of a like kind. The same element may become singular, dual, plural, masculine, feminine, neuter, possessive, objective &c. in different dialects and even in different positions. Definitive Postfixes. The use of definitive postfixes belongs to the earliest stage of the inversive formation and cannot be said to be even confined to it, for some prepositional languages postplace the definitive or demonstrative, as Siamese and most of the Indonesian languages. In the Africo-Semitic prepositional Janguages definitives are com- mon as postfixes, and they occur in very archaic words, as in pronouns. Substantive terms are, to a great extent, composed of a root and a definitive postfix in the Scythic and North Asian, in many American and African, in the Caucasian and Indo-European languages and even in Semitico-Libyan. The Draviro-Australian, unlike the Scythic and Caucasian for- mations, distinguishes the gender by some of its postfixes, in this respect possessing ar Irano-Semitic character, The Dravirian inanimate or neuter posfix am, um, mu &c is identical with the Indo-European m, am &c of the objective which in neuter words is used as the nominative. This usage is Dravirian also, In Semitico-Libyan the labial has a mase. and plural force, and in some languages it is common or neuter. The feminine i, a, of Dravirian are likewise Semitico-Libyan and Indo-European femi- nine terminals. The masc. (sometimes neuter) power of -n, d and the fem. power of -/ are not Indo-European or Semitico-Liby- an, but the roots are preserved with the same powers in Caucasian words for “father” and “mother.” All the Dravirian postfixes are found in Seythic, Caucasian and Semitico-African yocabularies. It is deserving of remark that the wide spread definitive in s which is a common Semitico-Libyan, Indo-European and Seythic postfix to substantives does not occur as a Dravirian postfix unless i, d, zh, j may be taken to represent it. In the Seythic languages s frequently becomes ¢ and both take the sonant forms z, d which countenances this suggestion. The vocalic prefixes common in Scythic and African languages, and in some of the Indo-European (e. g. Greek) are rare but not ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 31 entirely absent in Dravirian as has already been remarked, Their archaic use is evinced by the various forms of the 3rd pronoun and demonstratives. But it must be observed that in the vocabulary the prefixual vowel is frequently a contraction of the root or of its first syllable, and that the general stracture of the words is Scythic more than Caucasian or Semitico-Libyan, the vocables of those formations being comparatively curt and elliptic and more often involving a prefix or infix.* In the Dravirian vocabularies the definitives are common but they appear to have lost their sexual functions in most cases. As they have also plural functions they may indicate namber rather than gender in many words, most words being primarily collective or plural and not singular. AJ, J, lu, ru, the feminine definitive, is common. The masculine -an, —na, —n occurs less frequently, but as-the lax and flexile phonology renders the x easily transmutable into d, or Jon the one side and into m on the other, and as in some dialects d is the current masc. form, postfixes that now appear to be phoneticafly fem. or neuter may originally have been masculine. The neuter (sometime feminine) labial occurs under varied form —va, —av, -v, —vu, —pu, —p, —ba, —-b, —ma, -mu, -am, -m ke. &e, The neuter definitive —du, —da, -di, -thi is much less common. The guttural -ka, -ga, —gu &e. is comparatively rare save in Gond. As in the dialectic groups of other formations different glossaries affect different postfixes or forms of postfixes, showing that a separation into dialects preceded the concretionary stage. At the same time many roots have the same postfix in all or several of the dialects, in some cases by the direct transmission from the pre-dialectic period and in others from the dissemination of the form of one dialect amongst others. * Bopp has remarked that Sanskrit, Pali and Prakrit in combining the final vowels of the primury forms with case-sutfixes beginning with a vowel interpose n euphonically, a phenomenon which is almost limited to this group of the Iranian lanyu » in which, also, it is most frequently employed by the neuter gender, less so by the masculine aud most rarely by the teminine (Comp. Gram. I, 5133). In the highly eaphonie Dravirinn languages consonants are interposed, an bocce ag nam, ram, dam, tam &e, and it is possible that the Sanskritic languages deriv this peculiarity from the influence of the languages of the Dravirian formation with which it came in contact in the basins of the Indus and Ganges. I do not here consider the question whether the agreement in these particles between Dravi- rian and [ranian was a consequence of the advance of the latter into the province of the former or of an earlier cause. The definitive is a common one, It occurs as a prefix in the Africo-Asonesian langu iges and as a postfix in the Caucasian and Ugrian, and it is evidently the common labial definitive. 32 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, Am, the inanimate or neuter definitive, is common-in the South- ern vocabulary, Tamil-Malayalam ; Ju, nu, du, tu &c., variations of lw, in Telugu, Karnataka and Tuluva. Where Tamil has Puy bu, Malayalam has often ba, Telugu va and Karnat. vu. Viis compuratively rare. It sometimes becomes bi, mi, b. The final vowels vary greatly. Tamil affects ei, Malayalam a, Telugu and Karnataka u and i, Tuluva e, while Tudava gene- rally dispenses with the vowel. In the purer Dravirian lan- guages of the Vindyan group, Gond, Uraon and Male, similar postfixes occur, They are distinguished by the frequent use of h, ka, kha. Double definitives sometimes occur, and they are proba- bly to be explained in the same way as the double prefixes of Kasia and other languages. But ina few cases one of the defini- tives appears to have been infixed. Thus tolu “skin” is also tovalu, and potu “sun” is also polutu. , The definitives which are used as plurals have been already considered. The Chinese and Tibeto-Ultraindian aflinities of the labial are shown in the Table. * The more remote were adverted to in discussing the pronouns, The common plurals in kal, gal, kulu, ngal, nar, kan, la, al, r,ir, lu, ru, &c. and k are Scythic, East Tibetan, Ultraindian and Gangetic.* The Scythie, East Tibetan, Ultraindian and Gangetic plurals in ni, in, i (flexional in several languages) although radicallyidentical with the Dravirian ir, la, &c, distinguishes the systems in which it occurs both from West Tibetan (Bhotian) and Dravirian.* The postfixed definitives belong to the foundation of the forma— tion, and their forms and variations carry it back to an era in which Dravirian like Seythie and the other harmonic Aso-African formations had only partially conereted these particles with the substantial roots. In many iustances where the roots are common to Dravirian with some of these formations, the definitives vary. (See the remarks on the Caucasian definitives, ante, vol. viii. p. 34.) In the comparative paucity of prefixed definitives Dravirian is Scythic more than N. E. Asian, Caucasian, Semitico-Libyan, Tibeto-Ultraindian or Asonesian, * See Table of Plural Particles, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 33 The most marked feature of the Dravirian system of pronouns and particles is its combination of Chinese and Tibetan roots with a Scythie phonology and structure and with some Scythic roots that are not Chinese. In its cruder and less agglutinative archaic form, of which Australian is partially a representative, its true place appears to be between Chinese and Scythic. The radical affinities of the system with Tibeto-Ultraindian are close and unequivocal. In roots the two are the same, and both are Scythico-Chinese, and much more Chinese than Scythic. The Dravirian and Australian forms do not appear to have been directly derived from Tibeto-Ultraindian, They havo several marks of independent derivation from an E. Asiatic source, Chinese and Scythic. The historical connection with Chinese must be of extreme antiquity and altogether pre-Indian, for the general character of Draviro-Australian is inconsistent with the supposition that the Chinese formation itself was the first to spread into India and become the basis of the Dravirian. This would involve the assumption that before the barbarous Draviro-Aus- tralians spread to Asonesia an original Chinese formation had been modified by an intrusive Scythic one in India. The con- nection is mainly with the Kwan-hwa or proper N. E, Chinese and not with the western, The supposition that Dravirian pre- ceded Tibetan in Tibet and is simply the product of the oldest Seythico-Chinese current from Tibet into India, Ultraindia and Asonesia, would make the close connection with Tibeto-Ultra- indian a direct historical one, for the latter would thus be in great measure a form of the archaic pre-Indian Dravirian in which, after the separation of Dravirian, the Chinese element had increased from contact with Kwan-hwa and the Scythic proportionally diminished. But the Tibeto-Ultraindian languages themselves oppose strong facts in phonology, glossary and ideology to such a hypothesis, and Dravirian has direct western affinities—Caucaso- African, Iranian and Ugrian—which would of themselves render it more probable that the formation was transmitted to India round the Tibetan region to the westward, and not across it. The affinities between the Draviro-Australian and the Tibeto- Ultraindian systems are the necessary result of their both being, Seythico-Chinese, but Scythic and Chinese are each of vast KE ot ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. antiquity and appear to have all along been in contact, so that mixed formations must always have existed and been in the course of production. The individuality both of Draviro-Aus- tralian and of Tibeto-Uliraindian not only when compared with each other, but with Chinese and the existing forms of Seythie, is so strongly marked, as to claim for each an independent existence from the most remote periods of Seythic and even of proto- Seythie history, At the same time the Tibetan languages haye been from era to era receiving new impressions both from Chinese and from more than one branch of Scythie; and the casternand northern dialects have been more exposed to these influenced phan the western and southern. The Tibetan languages, thus perenially modified, have, in turn, been earried into the Dravirian province from era to era, supplanting and modifying the Dravirian languages, so that— leaving the Arian and the direct Chino-Ultraindian elements out of view—India and Ultraindia now present Ist Dravirian lan- guages, litile if at all Tibetanised, but in which some Tibeto- Ultraindian elements probably exist although difficult to dis- eriminate (South Dravirian), 2nd Dravirian modified by Tibetan (Kol and, much more slightly, Male, Uraon, Gond), 3rd Tibetan in different forms (Bhotian or western, Si-fan or eastern) and of different eras and varieties in each form, with much blending amongst themselves, as well as with Mon-Anam and Chinese, and with a variable but comparatively weak Dravirian element, difficult to discriminate in most cases from that archaic community of roots to which we have adverted and from Tibetan having a Scythic harmonic tendency. In the Gangetic languages for example, an agelutinutive and harmonic character may be either Seythic throngh East Tibetan or Scythic though Dravirian. The facts and general probabilities “of every case must give the decision, where decision is possible. The three existing branches of the Draviro-Asonesian family— the Dravirian proper, the Kol and the Australian—have each had an independent development, and been exposed to widely different influences, internal and external, from a very remote period. The Australian pronominal system is the most crude, redundant and agglomerative, and the least flexional. The systems, both of ETUNOLOGY OF THE INDO-PACIPIC ISLANDS. oo Kol and Dravirian proper are more agelutinative, elliptic, and flexional, and their forms and particles are more confused and in dialects have wandered more from each other and from the original system. While Kol retains some forms that have dis- appeared in Drayirian proper, the pronouns have lost the primary agentive or separate forms which both the other branches preserve. In most respects the system is that of an impoverished dialect of Dravirian proper formed at an early stage of the latter, and since modified, by separation, and by,the influence of Ultraindian formations. The breaking up of the original system is so consi- derable that it was probably produced by the contact of the north- ern Dravirians nine race having a different pronominal ideo- logy. It is a dialect that could not have arisen so long as the native Dravirian idiom remained strong and pure, and is of the kind that grows up when a race becomes closely connected and intermingled with a foreign one, The range of the Kol terms to the eastward renders it probable that this modified system was not formed until the earlier Ultraindian tribes occupied the lower basin of the Ganges, blended with the Dravirian aborigines and produced a mixed lower Gangetic race and language. The Kol system must have arisen in one community which ultimately became predominant in Bengal, spread over a portion of the proper Dravirian highlands on the right bank of the Ganges and carried its pronous with its numerals over Ultraindia, Each of the purer North Dravirian languages—Male, Uraon and Gond—has also had its pronominal, its definitive or its numeral system slightly disturbed by the North Gangetic branch of the Tibeto-Ultraindian family or by the previously modified Lower Gangetic or Kol system. Thus some of the Kol numerals are found in Gond dialects. Gond has reeeived a Tibeto-Gangetic possessive particle into its pronominal system, and like Kol it uses the plural Jabial in the singular of its 2nd pronoun, while the general irregularities of its pronominal system speak to the shock it has received from the presence of foreign systems or of a foreign element in the languages of adjacent and partially intermixed tribes. Uraon and Male have adopted a Tibeto-Gangetie pos- sessive. The annexed Tables show the glossarial affinities of the Dra- 36 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. virian pronominal roots, and of the possessive and plural particles. The other directives areso much interchanged and confused with possessives in Dravirian as in other formations that I do not give tables of them. TABLE SHOWING THE GENERAL RANGE IN THE OLD WORLD Il. OF THE PRONOUNS FOUND IN DRAVIRIAN. Ist PRONOUN (‘fI”) CHINESE. ngé Kwan-hwa, Quang-tung "nigu Shang-hai (pl. ngu ni, or ni, 1+ you) ngei Kek (Cheo-hu) | ngal ry ngol “s gua Hok-kien, Hai-lam wo, wu Kwan-hwa i Tie-chu ua ” wa ” nung (occasional) yu Kwan-hwa Draviro-AUSTRALIAN. A. Australian and other Asonesian. nga-nya W. Australian nga-toa N.S. Wales nga-t 8. Australian nga-pe Encounter Bay nga-tu Kowrarega ne-ng Tobi na-k Pelew ngo Rotuma ngou s B. Dravirian proper. na-na Gond, Karnataka (poss.) Braliui (poss.) na-n Tamil, Kurgi, Brahui (pi.) nya-n Malayalam nya-n * ETHNOLOGY OF THE INDO-PACIFIC ISLANDs. 37 nae-nu a-n a ya-nu ya-n na-k na ne-nu e-na e-nu ing inge eing aing D. inga ninga eng Karnataka Kar. Ane., Gond (aq. postf. rT Gond (ag. posts: ) Tuluva Tamil Gond Telugu (poss. obi.) Telugu Tam. (in obj. obl.), Mal. (in obl.), Karn, (poss. obj.), Kurgi (ib.), Toda (ib. ) Kam. (ag. postf.) Karn. (ag. postf. in present ) Kurgi ( poss. obj.) Uraon Mal. (poss. -re or -de). Tuluva, Male, Tamil ( poss., ag. postf.) Toda (ag. postf.), Uraon ( pl.) Kurgi ( pl.) Malayalam (dat.) Karn. (ag. postf. in verb abs. past tense), Tu- luva (éb.) ! Brahui Tod. Male (poss. with -ki) Male (pl.), Tamil ( pl. ag. postf.) Gond (/poss.; in pl. mo-wan) Tuluva (pl. ag. postf. in verb, abs., past tense) Bhumij, Mundala, Ho Sontal Ho Ho Gangetic and Ultraindian. Limbu, Milchanane Milchanang Namsangya Naga (poss. ) Kambojan, Chong. 38 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. eing Simang cin ss en ” ain Binua oin or yun ” oei Mon oe ” ye Simang eyu Binua Ei. Asonesian. aing Sunda ani | Timor nyu-ugea Sumba anare Belo ba-nian Kissa ina Formosa III. Trevro-ULTRAINDIAN. ifgea Tibetan, Horpa, Gyarang, Naga (Namsang.), Kasia, Burman, Murmi, Gurung, Magar, Serpa. nga-yo Gyarung (double form, yo is Chinese, ) na Tibetan ngya ” ngat Burman (poss..), Singpho, Tengsa, Naga (poss. or pl. forms, Tengsa has a in pl., the Sing- pho pl. has i; a mixed system; the 2nd pron. in Singpho has both nang and ni in sing., ni in pl.) . nyi Naugaung Naga (pl. annok, mixed system) ni Khari Naga, (pl. akan, mixed system) aug Bodo, Garo, Naga (postf.), Kiranti (poss. ang ko) an Deoria Chutia alg-a Garo anka Kiranti (a modification of anga or two roots combined, see ka infra ) a Manyak, Naga (Angami), Mikir (pl.) ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, nge nye ne net eneshe eneatung i ni ani ainko in-n0 ugi ning ni-ma ni-khala yo i ngo ong ka go gu gi “eo ki kyi kima kem kai tai-le kha kau akau tau ti-sez ti-checha Takpa, Singpho (vbl.) hE Mikir, Tunglhw (pl.) Ladak and Kinawari Tibetan ( pl.) Tibarkhad (pl.) a3 + | » (4g. post.) Serpa (in pl.) Limbu (in pl.) Kiranti ( poss. p/.) Murmi ( poss. pl..) Guriing (in pl.), Kasia Fs ) Garo (pl ) Namsangya Naga (pl.) Tablung Naga (pl.) _ Gyarung ( pl. a Chinese pron. ) Singpho (pl. ) Abor- Miri Lau ( poss. in Laos) Thochu, Dhimal (a var. of nga), Lepeha (in poss. sing. kaseusa and in pl, kan-kurih. Comp. Kiranti an-kan ( pl. ) Lepcha, Sunwar Milehanang, Sumchu Tiberkad Mi'chanang (in pl. ki-shunq) Khyeng, Silong (ki in pl. with postf.) Kyau Joboko Naga (pl.) Kumi, Kami (comp. ngai Singpho &c) Muthun Naga ( pil. i, e, ¢ for k ) Lau, (Siam) Lau (Khamti, Ahom) Kari Naga ( pl. ) Tablung Naga, Anam (t for &) Mulung (poss. ) to-we ( obj. ) Tablung ( pl. ) 40 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. ku , Lau (Laos), Muthun and Joboka Naga kung Lau ( Shan ) | khwa Toung-lloo he-lam Mulung (sing. ) he-lan ( pi. ) IV. Caucasian. na Kasi Kutauk -n Iron ( postfix ) -in 9 : -on ” V. Evusxkarran, n ( objective ) VI. Semirico-Lipyan. [ See ante Sec. 6. Supplement to Sub-Sec. 4; the root is na, no, nu,ne, ni, an, in, &c, with a prefix or postfix or with both, but also occurring bare,—contracted to the postfix or to a vowel or consonant of the root or postfix, the latter also changing from k to g, h, t, 8.] VIL. Uartan. The Ist pronoun is the common Scythie labial, but in some cases the m changes to n. na Samoiede (Motor) VIII. N. E. Astan, na Korea nal as ad Yeniseian dy 9 y® ” al ” 3 ” IX. Amenican. ne _ Athapasean nan f. neeah Sioux (Winebagoes) ney ” ” ni Shoshoni I a} in Sahaptin &e, ETHNOLOGY OF THE INDOQ-PACIFIC J8LANDS. nai &e. I. Curnese. II. li lin urh, ’rh nai, nei nong na nyi, ni ndi li a du ju jo nyu 41 Chinook &e. 2nD PRONOUN (“THOU”). Kwan-hwa, Gyami, Horpa, Quang-tung, Shanghai, in pl. of Ist pron. 1fgu nior ni (i. e. I, thou) Kwan-hwa Bi a) yy (ane.) Shanghai Kek ( Cheo-hu) - Kwang-tung of Si-ning Hok-kien, Tie-chiu i” 3 Hai-lam Kwan-hwa DRAVIRO-AUSTRALIAN, ' A. Australian and other Asonesian. ngi-ngi ngin-toa nin-na ngin-te ngi-du ni-wu ni-medu nu-rang nu-ra nu-wala nu-rali ngu-r'le ngu-ne ono uligoe ooine Sydney N. S. Wales S. Australian Encounter Bay Kowrarega S. Aust. (dual.) W. Aust. ( pl.) W. Aust. (dual.) N.S. Wales (ib.) Parankalla (zb.) ” (pl.) Kowrarega (dual.) » (pl) Onin Tarawa Hawaii ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. B. Dravirian proper. ni ni-nu ni-n nin-na ni-en ni-v% ni-k i-nu i-ng i al i-r i-7t i-r-gal na na-nt un nu “Tu ré Tamil, Malayalam, Toda, Telugu ( poss.) Karnataka Ib. Anc., Kurgi, Male, Tamil (ob/.) Malayal. (obl.) Karn. (poss.), Male ( pl.) Uraon Telugu (pl. postf. in sing.) Gond Khond, Tuluva (in pl.) Male (in poss.) Gond (ag. postf.), Karn. (id.) Tamil (7.) Kar. (pl.), Gond (ag. postf: pl.) » (pl. ag. postf-) Tamil (id.) Toda (in pl.), Brahui (ob/.) Malayal. ( poss. with postf.) Tamil Tamil Anc. (with pl. poss.), Brahui (with. pl. postf-) Telugu (ag. postf.; pl. particle for sing.) Brahui (7.) Plurals with the labial postf. or flexion. ni-m nim-ma nu-m u-m mi-ru mi me im-at im-ar mi-wan Karn. Anc., Jb. Mo d. (poss. with postf.), Male (pl. poss. with postf.) Karn. (obl.), Kurgi Brahui, Tamil Anc. (poss. with postf.) Tamil Mod. ( poss. ) Telugu » poss.) Gond | bed 3» ( poss.) Labial Plural forms used in the Singular. vu Telugu (aq. postf.) ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 43 um-athu Malabar (od/.) im-ma Gond C. Kol. in-ko-ghi Mandala ( pi.) um Ho [Tam. um pl.| umma Bhumij (o0b/.) umge Sonthal am Bhumij, Mandala ami Sonthal (ob/.) appe com. pl. of Kol api Mandala ( pl. with postf.) me Ho (obj. postf. in verbs) [Gond] m » (t-) be » (pl) D. Gangetic and Ultraindian. am Kiranti (in poss.) pi Mon pueh bai ” pha Kasia phi ” (pl.) me Kasia, Tengsa Naga (im poss.) ma Namsangya Naga (poss, sing and pl.) mel Anam wont Kambojan mung Lau (Siam) mat yy (Shamti) mo », (Ahom), Simang bo Chong, Simang mong Malayu of Traugganu E. Asonesian mo also into composite plurals of the Ist pro- mi nouns nol Malayu-Polynesian, poss. and pl., entering nyu-mu Sumba moce Solor 44 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. iba ibu iwo Bali Kandayan Mandhar III. Trsero-UttTRAINDIAN. A. Last Tibetan or Si-fan. nan-re nan nang ngan nga na no nuwo huni ani i ni i ne neng nen meng men Gyarung Changlo, Kami Bodo, Garo, Mikir, Singpho, Burman, Khy- eng, Kumi, Naga (Namsang, Tengsa, Kha- ri, Tablang, Mithan), Magar, Changlo (in poss. ) Tiberkad 7 Gyarung (poss. pref.*), Dhimal, Mikir (in pl.), Singpho (in poss.), Naugaung Naga, Tengsa Naga (in pl.), Toung-lhu Manyak, Dophla, Abor, Deoria Chutia, An- gami and Mozome Angami. Namsangya Naga Magar ( poss.) Tiberkad Deoria Chutia, in poss. ni-yo [ ? Dray.] Takpa Dhimal (in pl, ny-cl, poss. ni-ng) Singpho (in pl. ni-theng), Khari Naga (in pl. ni-khala.) Takpa Namsang Naga (in pl. ne-ma), Angami Naga (in pl. ne-ra-ma; also in poss. sing.) Burman au 7} | F B. West Tibetan or Bhotia. [The root is not Chinese in form, but I place the series here in order to illustrate the mixture of systems in the Himalayo-Ultra- indian provinces. The original was probably nga, nge, ngyo, a * Ni is given in the Vocabulary (and copied by Muller) as the prefixual poss. form, but it appears to be a misprint as Hodgson in his notices of the grammar invariably uses na-, and in a note to the Vocabulary na- also occurs. i ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 45 form of the broad or E. Tibetan variety of the Chinese root still found in Tiberkad. In Thochu a similar change from ng to k has taken place in the ' kwa kwe ka chha khyod khe khyo khye khe-ne ke-n kha-na kha haw ha-yu chhu chhe kheu gai ai ki kbau ist pronoun. | Thochu a7 » (in poss.), Milchanang (also kas) Sokpa, Newar Tibetan wr. sp. a7 Serpa » (poss. pl.) Limbu Gurung Kiranti Lhopa (in pi.) Lepcha » (pl) Lhopa 3x (p0ss.) »» (poss. pl. as in Serpa) Sunwar Marmi Milchanang Ahom (pl. Lepcha form) [TV¥. Oavcasian. di Iron, closer to the Seythic. (V. Evsxariy. n Jem, perhaps def. only, (VI. Spmrrrco-Linray. The 2nd pronoun is the dental, as in Seythic, changing in some eases to & and also to th, sh.] VII. Ugenrtan. na-n na-nk neil ny ny-ngl Wogalian a AP oY ay 46 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. ny-n Ostiak nu-m Ostiak [ The other Ugrian languages pas the common Scythic dental and sibilant pronoun. * | VIII. N. EB. Asmy. * nun Korea un ” IX. Awerrcayn. yin Athapascan ni » Otomi (poss.) jan » na 3 nanuk 3 ne-be Cheroki niah Sioux (also dia, de, neh) inui Selish nan Kinai nin-he Kitunaha eno Naas nune 35 &e, &e. &e. TABLE OF PLURAL PARTICLES OF E, AND 8. E. ASIA. Chinese, E. Tibetan, Ultraindian and Indian. mun Chinese mei PP pel ” me Gyami mye Gyarung kamye -mo Gurung si-rmong Garo (2nd pron.) -ma Naga (Namsangya) ma-rang Garo ma-dang ae * Muller’s table supplies: nen, nenna, -n Ostiak Sere iin Yakuti ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 47 -m Joboka Naga, Dray. -mya Burman -we Angami (Ist pron.) -te-be Tengsa (2nd pron.) -ping Newar : [-nam Tibetan] N. Lt, Asian and LE. Scythic. a.|,n Yukabiri n, ug Yeniseian r (7) Koriak ra Japan ri Manchu yl Nyertshinsk lar, ler, r Turkish nar,ner Mongol n Ostiak jergi Manchu Seythic. i Fin i flex. in pron. Yukahiri, Hungarian, Turkish e Samoiede e Jjlexw. in pron. Manchu Chinese and Scythic. ki Chinese g,k,t,d Seythic (with different vowels), also Caucasian, Euskarian &e, tu, su Chinese [Manchu sa, se, si, Mong. s, Turkish z, variations to ch occur. The Scythic si- bilants are probably from t] E. Tibetan. b. ki, ko Thochu b. +a. k-lar Thochu a, +b. rigi Horpa a. ni ” dur Manyak [Mong. od, d, da, t &e. with r as in nar | Il. Dravirian and Asonesian. a. la Dray. 48 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. al Dray. lu 5 rw Ps, ir a re ‘3 de Fe r ” nar H mar ss n Gond (in pl. of pron. poss.-wa-n) b. + a. kal ¥) 3) gal .} | ” ngal ” » Kan 3 » Kulu x b. g ” k, nk a” (Fe ko Kol a. Ya Australian (plural or dual) rang ” rali 7 rle ¥ wa-la A li + dli * le Pe lin ” rin P dlu oa b. + a. ngalu ies galang ” b. nga <5 ra Aru rare ” aronga Polynesian Ill. Gangetico-Ultraindian. a, ra Takpa ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. rang arang rama madang era arai b. + a. khala kara galal el, al, | le li to-leli li, di, ni ni In, nh ning to-thete to-leli jo, njo u Chinese tse Garo, Serpa Abor Angami Garo Bengali Siam (3rd pron.) Tengsa, Tablung Naugaung Dhimal a3 Muthun Naga (1st pron.) Mikir Angami Naga Khyeng Murmi, Singpho Kiranti Namsangya (8rd pron.) Kasia (general) Singpho (pron. ) jlex in \st pron. Serpa, Limbu, Kiranti, Mur- mi, Gurung, Garo, Singpho Namsangya (flew. 2nd pron.) es (demonstratives) Magar [rigi Horpa] Bengali Abor (? ding from ning) Sunwar Singpho Bodo [dur Manyak] Tibetan Lhopa Burman bh] Angami (38rd pron.) 3 (2nd pron.) Tibetan Lepcha and Gangetico-Uliraindian. Chinese G 49 50 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. tang Chinese eshe Tiberkad esh, ish, osh Milchanang tehi Kumi [chi, si may be from ki] chi Garo (lst pron.) Kiranti (3rd pron.) Limbu (¢4) si-mong Garo (2nd pron.) sin Abor tam-she § Kanawari Bhotia fi. e. the Chinese double tang-tse. Comp, Tiberkad eshe]} ta-she wa tham-che Changlo atung Tiberkad te-be Tengsa (3rd pron.) to-thethe Angami (3rd pron.) the Toung-lhu checha Tablung Naga AFFINITIES OF THE DRAVIRIAN POSSESSIVE AND QUALITIVA PARTICLES. I, Scyruic. Dravirian. na, an, nu, ni, in, no ta, tu, thi, ti, ji, che, cha da, du, di, de ra, ru, ri, re Ja, lu &e. na, nu, no, nau &c. athi dana, tano tat, tad, dad a, i,e la, Ya, yO, ye, ei yo-ka ; da-ya u-da-ya u-dei-ya in-u-da-ya in-de Il. ETHNOLOGY OF THE INDO-PACTFIC ISLANDS. 51 u-dei a-du ea la, ai, @ Kol a-tana - t- Sonthal (pref) Scythic. ni, un, ung, en, na, an ning, nung, uigge Ny US» i, e, u inki Tana East Tibetan (? Bhotia yi, i) ni Sokpa i, ¢, Manyak [i Mongol, Manshu] Gangetico-Ultraindian. i Burman ni Bodo, Garo in Limbu un Kami ng Dhimal ne Mikir na Singphu, Murmi (also la) [Scythic na, an, a] la Murmi, Limbu, (qual.) Changlo (ib.) ra Limbu (qual.) lu, lo Changlo (qual.) © nang Namsangya Naga rang ” CHINESE PossessIVEs IN TipeTAN, GANGETICO-ULTRA- INDIAN AND N, DRAVIRIAN. Cuinese A. ku, ken Shanghai ge, e Hok-kien ko Quang-tung 52 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Tibetan. uk k-chi Thochu khyi, khi, kyi, hi Bhotia ga, ka »» (qualitive.) Gangetico-Ultraindian. gi Lhopa ga Changlo, Abor g Abor, Daphla ga, ka qual. Newar eu Newar ke, ku qual. Limbu ko, ku, ke Takpa, Kiranti, Sunwar, Magar, Dhimal, Khyeng khang Siam North Ultraindian. ki, Male ghi, hi Uraon CuineEsE B. tih, chi, te Kwan-hwa East Tibetan. ti Gyami k-chi Thochu Gangetico-Ultraindian. ti Serpa chi Tengsa Naga sei Tablung Naga 8a Lepcha SO, 0 Kiranti Dravirian. [Possibly some of the dental forms may be Chi- nese and not merely variations of the Scythic n.] ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, il, NuMgna.s. = As numerals are based on definitives, the principal test of their antiquity ina particular language is (heir mutual dependence, and their relation to the definitives preserved in pronouns, substantival prefixes or postfixes, directives &c. If their elements are the same that occur in these particles, and if the terms for the higher numbers are connected by composition or flexion with those for the lower, it may be concluded that the numerals are native, that is, belong to the earliest era uf the language, or of the formation of which it is a member or derivative. If the different terms have no connection with the other particles of the lancunge, it may be inferred that they are extraneous or of foreign origin; and this inference will be greatly strengthened if there is also-an absence of connection amongst the numerals themselves. But, in the latter ease, the heterowencous character may be either that which they had in the single foreign language of their immediate ori- gin, or it may be a consequence of successive displacements of old terms by new ones derived from several influential foreign languages, Tried by this test the Dravirian numerals must be considered as very archaic, and as native in the linguistic formation to which the ancient Indian languages belong. It may be remarked amongst their archaic characters that they are not only qnalitive as in other systems, but the roots are always clothed with a posses- sive or qualitive postfix, so that the series is literally “ one-of,” “two-cf”, “ three-of” &c.* * See Appendix A. Comparative Vocabulary of the Numerals of the Dravirian Formation. The following are examples of the terms. ti 54 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Rejecting the possessive -postfixes, the S. Dravirian roots appear to be 1, on, vo; 2, ir, er, ira, era, re, ra (euphonically ren &e); 3, mu; 4, nal; 5, ai, (euphonically ain, an, d-c.); 6, a; 7, e; 8 (2,10); 9.1, 10). If -du, -zhu, -~ju, -ndu, -nju &c., -ru, -lu, “nu, -tu, -da, -zha, -ta, -la, -ay,-di, -ji, -ti, -de, -d &c. are all merely Hlexional variations of the possessive postfix, as is evidently the case the root of 5 is az, and not ain, anor anj. If this view of the basis of the Dravirian numerals be correct, it follows that it was originally forme. from a few definitives, further distinctness having ultimately been attained in each term by slight variations or Hlexions both in the roots and in the common postfixual posses- stves, variations similar to what take place in all agglutinative aul Hexional languages. Thus in Tamil the postfix takes the forms -ru, 1, 3, G; -du, 2, 9; -lu, 4; -ju, 5; -zhu, 7; -tu, 8; -ta, 9; —in Malayalam -na, 1; -da, 2,9; -ar, 3; -ra, 6; -la, 4; sa, 5; -zha, 7; -ta, 8, 10;—in Taluva -ji, 1, 5,65 -d, 2; -lu, 45-1, 75 -nu, 5; -t, 10 ;—in Karnat. -du 1, 2,5; -ru, 3, 6 sel, 4 ; -lu, 7; -tu, 8, 9, 10 ;—in Telugu, -ti, 1; -du, 2, 3, 5,7; -lu, 4; -ru, 6; -di; 8, 9, 10 ;—in Todava, -da,1, 2;-du, 3; -n, 4;-j,5; -ra, 6 ; -ta, -t, 8, 9, 10, From the easy convertibility of most of these forms, any origi- nal regularity in their flexion—if such ever existed—was not likely to be preserved. But sonie of the languages maintain a, manifest . connection between I and G, and between 8, 9 and 10, the former being probably dependent on an archaic quinary scale, while the laiter intimates that when the scale became decimal, the lower numbers in the vicinity of 10 were named with reference to it. From the general character of the variations in the forms of the postfixes and the faintness of any traces of real Hexion, it is pro- bable that none of them had ever any function but the simple possessive. They are similar to the ordinary variations of the possessive, the consonant being d, t, r, I, n, j, zl, nd, nt, nr, and the vowel a generally, but sometimes a ( Malayalam), ori (Tuluva). [See the remarks on the final yowels affected by different dialects]. Lravirian. proper. Kol. 1. on-7w. 1. m-ia 2. era-. on-ba-d 9. ar-ea 10. pa-tu 10. gel-ea 56 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. It is found in some Ultraindian and many Asonesian languages as a definitive, demonstrative and unit. The South Dravirian on, 1, appears to be one of the two principal definitives, demonstra- tives, and 3rd pronouns of the formation, — The second, and in South Dravirian—as in Sifan, Gangetico- Ultraindian and Australian—more prevalent, 3rd pronoun &c., the labial, is also used in 8. Dravirian as the unit. It is a common definitive postfix, as well as 3rd prenoun. In the exceptional vo-dda Toda, vo-ka-ti Telugu, 1, vo is evidently the root and da, ka and ti possessive postfixes. The antiquity of the term is prov- ell by its having kept its place in 10 and the higher numbers. The Telugu guttural, as we have seen, is preserved in the Ancient Tamil 10, ba-hu-du (in 9 and 50), pa-ku-du (in 10, 20 &e). In 10, 20 &e, the labial root takes the forms ba, pa, va, in Mal. ; ba, pa, va, and mi in Telugu; bha, ha, and va, in Karnataka; and bo, po, vo, pe in Toda. In the Malayalam, Telugu and Karnataka 9, on, 1, becomes om, which assimilates it to vo, but the assimilation appears to be phonetic merely. In the term for 100 Telugu preserves the labial, va-nda. With reference to the variation of the vowel from a to 0, itmay be remarked that in the 3rd pro- nown the southern languages have va, av, am, &c. while Gond has wu-r, and that o, u, are found in Newar wo, and Abor bu. On the other hand Gyarung, Dhimal, Garo and Tung-lhu have wa and Dophla ma, while Takpa has pe and Bodo bi. In S. Dravirian the postfixed labial definitive has various forms, bu, bo ba, va, VO, vu, pa, po, pu, ma, mo, mu, um, am, &e,, the vowel having little stability. The other 3rd pronoun of the Dravirianformation,—as-an Uraon, ath Male, (asa-bar in pl., ahi-Aé in poss., ih “ this,” ah “ that’), it, id, adi, athu &e. S. Dravirian,—does not occur as the unit in any of the Dravirian or Kol numerals, but the Brahui as-it has it. That as is the root and that it is Dravirian appears not only from the postfix, but from 2 and 3 also being Dravirian (ira -t, mu-s-it). The absence of the sibilant as a Draviro-Australian unit is one of the most striking peculiarities of the system. Ra, e-ra, yer, i-ru, ir, re, en &c. 2, is one of the variations of the common def. da, la, na, &e. of which n, na, has pronominally been restricted to the masculine gender, and Ja, /, to the feminine. Ra ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 57 (variable to la, Je, &c.) is the plural form, and it may be derived from 2, or vice versa. In 8 (2, 10), the root for 2 has the forms e, ye, in Tam., Mal., Tod., en, yen in the other dialects, as in the Uraon 2 (en-otan). In 12 itis er, ira, ra and e orne. In 20 it is ira, ir, iri (in Toda ye, e, i, and in Karnataka i). The root for 3, mu, corresponds with the labial definitive, with the pronominal plural element, and with the labial root for 1, thus giving indication of a primary binary seale in which the term for 3 returned to the root for 1, (2, 1). In higher numbers (13, 30 &e¢ ) it generally retains the form inu. [un Dr Stevenson’s Karna- taka list 13 is had-im-b-ru, in which b represents mu and labia- lises the » of the conjunctive —in. The root for 4, nal, nar, non, (if we include the final of the first syllable of the term), appears to be a repetition and reduplication of ra 2 (i. e. 2 dual, as in many other languages). In 14 itis nal, n or an (pa-n-ka, pat-in-an-ku). In 40 it is nar, nal, Tt is pro- bable that the & pustfix wus adopted instead of that in x, /, d, &c. to distinguish it from the root, This is supported by the fact that in the higher numbers the other numerals lose the possessive post- fix, while 4 loses Au k only and retains 1, r. The closest foreiyn terms for 4 have a final |, n, &c. (nila, nol, nan &e.) The higher roots present little that is tangible. But there is evi- dently a connection between these very elliptic and undefined high- er roots and the two first of the lower series, 1, 2, 3. 5 is ain, yan, an, or ai, ya, ayi, ei. As the higher as well as the lower numbers are formed from three elements, on &c., mu &e., and ir, er, &c., it is not probable that ai, &c., involves any fresh root. As i, e, is only found in the root for 2, and represents it in some other terms, it may do so here also. In the Toda er-bod, 50, 5 is represented by er 2. The term in Toda at least, was therefore 3, 2, (as in Kol), and as the a of ai can hardly be a remnant of the term for 3 (unless muna-iradu was the primary form and not munru-iradw, which is improbable), we must explain ai, ei, as a phonetic varia- tion of e, if we consider it as 3, 2. In some forms a, ya, represent the e or i, and in Git is also represented bya. There is another and—despite the Uraon and Kol terms— more probable explana- tion of the 8. Dravirian 5. In many quinary systems the term for 5 is the root for 1, or a merely phonetic variety of it, on the same H 58 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. principle that 10 is named 1 in many denary systems, 5 was “ one tale,”’ counted on the fingers of one hand, as 10 was “ one tale,” reckoned on the fingers of both hands. One of the forms of the Dravirian definitive, demonstrative and 3rd pronoun which 1s used as 1 in the term on, nu &e, is yan, ayi, aye, ai &e, This would appearto be the root of 5 in the South Dravirian dialects, (Comp. Tuluva aye “ he” &e., ayi-no “ this,” ayi-nu, 5). A, o, 6, is still more elliptic than the ai of 5, and like it has the form of a mere definitive. The Toda form, o, is identified with on, 1, in L1, and the term would thus appear to have been a quinary one, 5, I. In the Appendix, al- though considering it probable that the root is a, I have refer- red it to ira, era, 2, the a appearing to point to it rather than toon ke. 1. But the Toda o-r, 6, has the proper vowel of 1, and it occurs in the same form in ll. The Tuluva and Gond a-ji, G, has the postfix of 1 (on-ji T., on-di G.) and not of 2 (-du T., -nu G.) The -ra of the Mal. 6 corresponds with the -na of 1, and not with the -du or -ndu of 2. (The postfixes of the other dialects are the same, or nearly so, in 1 and 2). The term for 6 would thus appear to haye been a quinary one, 5, 1, the word for 5 having been disused for brevity’s sake. In many other formations a quinary system appears superimposed ona binary and ternary one or ona compound of both, and it is only in the erudest glossaries that the term for 5 is retained in the higher numbers. The root of the Dravirian 6 is thus merely a variety of that for 1. The e, ye, of 7 bas the same character. It can only be referred to the e, ye, of 2 (5,2). In 8, e, en, again oecurs as the representative of 2, and the formation of this term as 2, 10 and of 9 as 1, 10, clearly indicates that the denary scale was superimposed on an older and more limited one, probably quinary as far as it went, 1; 2;3;4;5; 1,5;2, 5. There would also appear to have been a quinary 8 (i.e. 5, 3). In the Appendix the Gond form, ana-mu-r, is omitted. It resembles the Tuluva en- ame and the Telugu en-imi-di. In all these forms the labial unit of 10 has neither the form v asin I of Telugu and Todava, nor that of v, p, b, as in 10 and the higher numbers in all the dialects. It preserves the m of the Kol 1 and of the Dravirian 3. The Gond 10 has the form pa-da of Malayalam, while 1 has the form ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 59 un-ddi (on-ji Tuluva). The Gond mu-r of 8 appears to show that, when the term for 8 was formed, mu-r or mu-rw was the current form of that for 1. But for the e, en prefix in all the terms for 8 save the Gond, mu-r would be referable at once to mu-ru 3 (Karnataka ; mu-nuw Gond). In the same way the Telu- gu mi-dé and Tuluva me would be referable to a slender form of 3 which is actually current in Todaya, mi-n. The term for 8 would thus be quinary (5, 3) like 7 and G. The Gond an of ana-mu-r is the an of the Tamil and Malayalam 5 (an-ju, an-ja), so that there seems to be no room for doubt as to its true quinary charac- ter. ‘The e of the other terms appears to be as clearly referable to 2. The Telugu mi-di recurs in 9 (t-om-mi-di), where it must represent 1. The forms of 8 and 9 appear to carry us back to the period when the labial kept its place in 1 as well as 3, and had the m form in 1 also, The Todaya bo-d is a near approach to mo-do, mu-du, mu-ru. The quinary system, in its turn, would appear to have rested on @ primitive binary and ternary one; and the series of terms as we now find it has the following sequence of root elements:—1, and also I (two roots), one; 2, two; I (for 2, I), three; 2, 2, four; 2 (for 3, 2), five; 1 (for 5, 1), siw; (2 for 5,2), seven; 5, 3, also 2, 10, or 2, eight; 1, 10, nine; I, ten. To those comparative philologists who have not analysed and compared a large number of numeral systems, this reduction of the Dravirian to three roots (two primary terms 1 or I, and 2), combined by binary, ternary, quinary and denary methods, may appear exceptional and fanciful, but the fact is that nearly all numeral systems have been built up in the same mode by a succession of steps. The Iranian, the Semitic, and most of the other Asiatic systems, as well as the allied African, Malagasy and Malagasy-Polynesian, haye had a similar history, and under their present denary form preserve vestiges of the earlier modes of counting and forming the names. A large number of African and some Ultraindian and Asonesian systems still retain the quinary terms from 5 to 10 undisguised, and entirely or nearly identical with those for 1,2,3and 4, In most systems 10 is either 1, or 1 followed or preceded by another word. Various illustrations of these facts are given in the Semi- tico-African sub-section, and they are more folly considered in a separate paper on the numeral systems of the Old World. 60 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. The first direction which our search for facts that may help to clear up the obscurities of the South Dravirian system, should naturally take, is to the Kol, Gangetico-Ultraindian and Asonesian systems. There has evidently been some displacement and pho- netic modification of roots in the 8, Dravirian system, and in some points the correctness of our analysis cannot be considered as fully established by that amount of mutual elucidation which the S. Dravirian dialects themselves afford. The Kol dialects preserve a somewhat different numeral system. It appears to have prevailed in Ultraindia also prior to the inmtro- duction of the Tibetan and Sifan modification of the Chinese, for it is now retained—partially blended with the latter—in those Ultraindian languages which in pronouns and other words, have the strongest glossarial affinities to Kol.* A full list of the yaria- tions which the roots undergo, with some remarks on their distri- bution and the probable course of their diffusion, will be found in the next chapter. The Vindyan, like the South Dravirian, numerals postfix a pos- sessive definitive, but in place of varying in different terms as it does, to a greater or less extent, in South Dravirian, it is uniformly -ta or -ya (with a few slight phonetic changes and contractions * The maritime position and habits of the Mon or Peguans, the evidences of their having been at one time thechief traders to the eastward on the Bay of Bengal, and of having greatly influenced the other Uliraindian, the Peninsular and several of the Indonesian races, with the undoubted spread of Vinilya—Ultraindian yocables through their instrumentwlity to the east and south, led me to surmise that the words common to the Mon—Anam and the Kol yocabularies, had been carried by the Mons from Ultraindia to the Gangetic basin, rather than by an inland tribe like the Kols to Ultraindia, and this surmise appeared to be strength. ned by the peculiarities of Kol compared with South Dravirian. The 2nd pronown in parti- ‘cular, with the lower terms of the numeral system, appeared to have a character eompletely foreign, Amongst the mi-cellaneous words comm: nto Kol and Mon- Anam ecealaaticion some were, beyond all doubt, non-Dravirian and of Ultraindian and Tibeto-Ultraindian origin. In the Introductory Note to Part LI ( ante vol. vi, p. 658) 1 therefore remarked that the vocab!les of the Mon-Anam formation were not only found in Gangetieo-U!trainJian lan usages, ‘but to a remarkable ex- tantin the Kol dialects, proving thatthe Peguformationembraced Lower Bengal and a portion of the Vindyas, although the Dravirian basis was preserved in the lan- | of the letter”; and in See. 6 (vol vii, p. 200) itis said “the phonetic basis of Sie lenigatre (Kol) and many particles and words are Dravirian, but the pronouns, several of the numerals and a large portion of the words are Mon-Anem." At the same time, the influence of the Dravirian pronominal system in Ultraindia was in several places remarked. A more minute examination of the pronominal and numeral elements of Dravirian, of the foreign continental affinities of the formation, and of its remnants in Asouesia, with the: reference of the Kol 2nd pronoun like the Ist to Dravirian, have satisfied me thut wl ile Kol, owing to its position, has been influenced by the ‘libeto-Chinese formations, as the race itseli has ty the Tibeto-Ultraindian, the uttinities between it and the Mon- Anum vocabularies are mainly of primary Kol origin. The most probable conclusion is that the Kols are ETHNOLOGY OF THE INDO-FAGIFIC. ISLANDS. Ol —as -ea, -¢, -i), This postfix is clearly Dravirian glossarially and idiomatieally.* It is not found in the Ultraindian systems, save in some varieties of the common terms which have a great appear- ance of being contractions of the Kol full forma, 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 ather archaic Aso-African systems, and as the same definitives were com- mon to most of the archaic formations, the Kol terms, like the 8. 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-z, m-ia, m-ea, m-ia-d, in-i-dh,+ m-i. The Ultraindian and Peninsular forms preserve the labial yowel (0, #,) and in some cases remnants of the Kol postfixes; po Angami Naga, bo Karen, muo, ma Mon, mo-e Kambojan, Ka, Chong, mo-¢ Anam, mu-i Binua, The root, as we have seen, is preserved with the labials m, b, y, p, and with the a remnant of the modified south Gangetie or Bengal division of the ancient Dravi- rian race. Their divlectic peculiarities of a seconJary kind must Jive been of much later origin than the first spread of Dravirian to the ea-tward, for the earl forms of the pronouns fount in Australia are the pure Dravidian, ‘The numerals cannot be explained as a mere dialectic variation of the South Dravirian,—but it is to be remarked that the South Draviriau developed mumeral system itself has no claim to stand on the same archaic footing as the proununs. The quinary system was not in existence when the formation first spread wiili its proper pronouns, into Asonesia, The general character of Kol shows that the lynauage must have existed as a separate one from a very semote period, There must have been at least two great and peepee Dravirian nations or races, the southern, now represented by the Gond-Tamil peoples and languages, and the Gangetic or B l, now represented by Kol. The possession by the lutter of a somewhat peculiar numeral system, altho h a sufficiently rewarkable 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 homogencity of the Dravi system of pronouns and numerals appear to be referable to the ethnic revolution occasioned in the Gangetic basin by the entrance of the Chino- Ultraindian and Chino-Tibetan trace. ) ; * In South Dravirian it occurs under the forms -ya or -ia, -iya, yo -a, -y or -1. The fuct 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. th systems go back to a period prior to the concretion of the possessives with the numeral roots. t The superadded dental ( -d, -dh) appears to be the Kol possessive -f, and its presence implies that the other ve -ia,-i had become conereted with the root. It is remarkable that the ui term for 3 has a similar secondary dental, mu-si-é (mu-si being obviously a variation of the Dravirian mu-fi Tul.) ‘The ‘elugu vo-ka-ti, has also secondary -fi. The exceptional Kol -d,-dh probably ae influence of aS, Dravirian dialect. The Anam mo-¢ preserves the Kol post . . ¢ 2 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 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, fa-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 7 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 pro- noun 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 terins of the numeral series being equivalent to “this”, and “ this dual,” or “ this -plural,” i-e. “ these”’. But the idea of duality or peat 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 &e. 2, similar to the Uraon en, but preserving i as in several of the 8. Drayirian forms. 3 is op, contracting to p (op- ia, p-ta). In Ultraindia it becomes pu-i, pa-i Mon, ba-i Kambojan, ba Anam, wu-i-p Simang (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 Dravi- rian and Kol; and the Dravirian 3, mu, mi, has the same root. 4 is upun, opun, pan, pn, in Ultraindia pun, bun, puan &c, pro- bably 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 8. Dravirian on, un, 1. The Kol term is different from the South Dravirian. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 63 5 is mona, mone, moi, mo, muna, mun, It is confined to the Kol and Gond, the Ultraindian terms being different. The first element of the term mo, mu, has the form of the labial unit found in the Kol and 8. Dravirian 3. The.second element na, ne, n, r, may be the South Dravirian 2, ra, r, er, ren, na, | &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, asinthe 8, Dravirian system, with a consonantal final, as in 2 and 3, or a postfixed definitive, as in the S, Dravirian 3, mu-rw &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-¢ 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 pro- bably a contraction. 6, tur, turu appears to be of undoubted Ultraindian origin [see App. A to Chap. Vil ‘ The remaining terms are clearly Drayirian. They haye no Ultraindian affinities. 7. The Kol term like the 8. Dravirian appears to be quinary. The Sonthal iair is evidently the full form and the others pea the iya, aya, ia and eia representing the ayi, ya &c. of the 8 Drayirian 5, and the final -ir, -r, the rx of 2, so that ia-ir or eit is a 5, 2. is ir-al, ir-l-éa (in Gond, by inversion, ilh-ar, el-ar-ia), The i i the S. Dravirian 2, corresponding with e of the 8, Dra- virian 8. The South Dravirian na-l or n-al dis 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 8. 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, Drayirian 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-ich) has an external resemblance to the S. Dray. 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 8. Dray. postfixes cohered with the roots, it may be inferred that the only common particle is the initial a, represent- r 64 ETHNOLOGY OF THE INDO-PACIFIO ISLANDS. ing 1. The Kol a-ra would thus appear to be 1 from 10 like the Drayirian 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, cor- responds 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 g-el and k-ing or k-in (for n replaces nig 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 founda- tion. 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 Btuner 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 &e. The labial is the Drayirian 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 Austra- lian languages and is also a Dravirian and Seythic trait. West Australian has bal “he,” “it” &c. In the Karaula mal, the definitive appears in the same form as the unit, and the Bijne- tumbo war-at, Peel Riv. peer, Kowrarega wara-pune, Moreton Bay ka-marah are similar instances. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 65 The Drayirian nasal definitive andSrd 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 guéw or ngia but the poss. gung. The second element of the compound, bai, is probably a contrac- tion 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 asser- tive 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, Borncon nih, indi, unii, enah, Philipine una, ona, uon, enot, Mille juan, New Caledo- nian 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 Seythic, 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, ris a dual and n a plural postfix in pronouns. In some vocabularies ], 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: &e, 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 lomant, 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 66 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. to be merely variations of the Wiradurei ngun, gun, As is 2, 1, or 1, 2, ngoro probably represents 1 (the term for 2 having been dropped) ‘and is identical with the kol of 1. This is confirmed by the Wollondilly 8, in which kol appears (koll-uer) and the Karaula 3 (kul-eba). The Perth g-wdjal, 2; (gyn 1) is formed from the 3d. pronoun in g or ng precisely as the 8. Australian p-w'la@ 2 is from the labial 8rd pron. Kul, gal, kar, gar, gur, ngar, ngor are found as terms for 1 or representing it in higher num- bers in different languages. In Kamilarai gala, gira, are used for the assertive absolute as well as the labial definitive. The guttural without the liquid postfix is found as an element in several systems,” Raffles Bay, loka 1, orika 2, Corio koi-moil 1, Moreton Bay %a-marah 1 (the same compound), Jhongworong Xa-p, 1, (pro- bably a contraction of a similar term) 1. The Encounter Bay ki-tye “he” &c. appears to be another instance of the full gutturalising of the ngi preserved in Wiradurei, and analogous to the Perth gyn which also retains the slender vowel. The En- eounter Bay dual kengk appears to be a reduplication (the 1st and 2nd pronouns take dual postfixes -le, -wrle). The plural k-ar has the proper dual form, The Australian guttural unit explains the Kol 10(gel, gil, gul). 2. The E. Australian bula, bul-ea, bulo-ara, pul-ar, and the Northern la-wit-bari appear to have the Kol bar, bar-ia, bar-ea &c. The Mairasi a-mui, Bruner I. la-mui, have a singular resemblance to the Kol moi, Binua mui 1., They are varieties of the root found in Menado bua, Tidore ma-lo-fong; bu-lango Goront., pahi- wo Louis., bo Hunia, buiu Tupua, iu Gallia, In Australian the term appears, in some cases at least, to be the dual of the labial 8rd pronoun and demonstrative. For example, in South Austra- lian pa is “ he” &ec., identical with the Drayirian va, (ba, pa &c), and la, dla, dli, rla, urla &¢ are forms ofthe dual postfix. The dual of pa is purla [i. e. pa-wrla]. The dual of the demon- strative ia (Dray.) isa double forms i-dl-urla. In the possessive of the 2nd pronoun the dual is also compound ni-na “ thou,’’ ni-wa “ you-two,” n-a “ you,” ni-wa-dlu-ko “ you-two-two-of.” In several languages the numeral “two ” and the dual postfix isa com- pound similar to the 8, Aust, p-urla, “ it-two,” (i. ¢. they-two”). The Kamilarai buloara and Peel Riy. pu-lar are double forms like ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 67 it, and the Wiradurei bu/a, Moreton B. pwalak are single forms. MS Some cases the double form may be apparent only, for a final], is found in the singular in some languages, as we have above Mie, West Australian has bal “ he &e,” balal “ he-himself”’ bula “ they-two” general, i. e. friends, brothers and sisters, buléla if parent and child or uncle and nephew or niece, bulen, husband and wife. The reduplication of the dual also forms a plural bula-Jel “ they,’ but the substantival plural or collective postfix is also used in the form bal-gun, “ they.” The use ofa third pronoun dual to denote the dual of substan- tives, as well as of the 1st and 2nd pronouns, is not an exceptional trait in Australian ideology, for the plural of substantiyes is also frequently expressed by a 8rd pronoun in the plural. , Thus the dual of “ dog’’ would be “ dog he-two” or “ the-two, * i. e. “ these two,” and the plural “ dog he-many,” or “ the many” i, e. “ these.” The Kol idiom is so‘far different that the dual is in form a limitation of the plural, sita ho “ dog these,” sita h-ing “ dog these-two,” The affinity of the Kol bar and Australian bula is complete in both elements glossarially, as well as in the compound being simi- lar to the dual or plural of the 3rd pronoun. In the most archaic condition of the system the dual and plural power may have been transferred from the numeral to the pronominal use of the definitive. 8. The same element recurs in the Anstralian purla, murn, burui, warh-rang, mar-din, mur-ten, mu-dyan, ma-dan, 8, which resemble the Dravirian muru, munru, mudu, &c. The Australian terms are 2, 1, generally fully preserved, but in a few cases with the 2or the 1 aided Thus some of the above terms appear to have the root for 1. Mar-din, ma-dan &c is the labial unit (mal Ka- raula), with a nasal postfix as in the Bijne-lumbo war-at, and the contracted Wollondilly me-dung (dung for du, the common def. postfix), Limbu Apiu mo-tu. But without additional vocabularies both of definitives and their numeral and other applications it seems hardly possible to analyse these terms with preci- sion, for the labial enters into both 1 and 2. The nasal final in din, dan, may possibly be the common Australian plural postfix. In the Kowrarega ta-na “ these,” “ they,” it appears with the dental Ggnienets as the 8rd pronoun. ! 68 ETHNOLOGY OF THE INDO-PACIFIC I8LANDS. 4... The, Australianterms are binary (2, 2) like South Dra- yinan, but in general with little or no agglutination, 5, The Mairasiiworo.may be connected with the Vindyan mor. | 10, The Pagai putu, Totong mo-put, Keh wut &c. resemble the S. Dravirian patu, pot. &c. The term is a common archaic one (N. E, Asiatic, African), The following are examples of the pure binary systems of Australian and Torres Strait. Kamilarai, 1, wa-kol; 2, bulo- ara (ara is also used as a dual and plural def.); 8, ugoro, (apparently a flexion of the kol of 1) ; 4, wa-ran, (a reduplication of the dual postfix). Peel Riv. 1 peer; 2 pul-ar; 3 pur-la (a pho- netic flexion of 2). Wiradurei, 1 ngun-bai; 2 bula; 3 bula-ngun- bai (2, 1,); 4, bu-ngu (apparently a flexional contraction of 3, but probably the full term 3,1). Bijne-lumbo 1 war-at; 2 ngar-garhk (i. e. “one-one,” the ngar, gar being the Southern ngoro, kol, loka, which preserve the definitive postf. as in Bijne-lumbo) . 3 ngar-gark war-at (2,1) &c. Erub, 1 ne-tat; 2, naes; 3, naesa- netat (2, 1); 4 naesa naes (2, 2) &e. Ke. The prevalent Malayu-Polynesian system is the Malagasy which has strong and fundamental Semitico-African affinities, and only very remote ones with Dravirian or Chino-Tibetan. In Asonesia there are also quinary systems and remnants of binary, ternary and senary scales, but as these are in many languages more or less mixed with the Malagasy-Polynesian denary terms, and as the Sifan and Ultraindian. systems also contain quinary terms, it will be convenient to postpone the further consideration of the insular systems till the Ultraindian haye been examined. The Chino-Tibetan and Ultraindian numeral system differs from the Drayirian, although one or two elements are common to both: From the above facts we are justified in the inference that there was an archaic binary numeral system which spread from India to Asonesia, and that in later eras larger systems were built on it, generally by quinary and denary methods, but with the ancient binary elements chiefly. These later formations were entirely independent in India and Asonesia. The Australian systems aro still essentially binary. They have not become even quinary, the few terms beyond 2 being chiefty variations of the lower terms. The series is still 1; 2; 2,1; BTHNOLOGY OF THE INDO-PACIFIC ISLANDS. 69 2, 2 &c, but with some elisions, contraction and replacementa. In India and Ultraindia, only two systems remain of the ante- Chinese era, the Dravirian and the Kol-Ultraindian,—the latter however presenting two varieties, the Kol and the Mon, The Kol is quinary and denary like the South Dravirian, and it has been formed from the same elements. But while the general method ia the same, even to the preservation of possessive or qualitive postfixes, there is a deviation in the mode of forming one or two terms, and the elements in the lower numbers are in some cases differently applied. The two systems were therefore independently formed from common materials at a very ancient period and before the various elements had become concreted. The most archaic term for 1 appears to have been the labial vo, mo, mu, bo, po, ba, pa &c. It is found in 8. Dravirian, Kol, Ul- traindian, Australian and a few other Asonesian languages. It 1s the definitive and 3rd pronoun common to Draviro-Australian with Sifan-Ultraindian. Tn all the dialects of the Gondo-Tamulian branch of Dravirian, save Toda and Telugu, it has been superseded by another Dravi- rian definitive, on, un, or, which is also found in Australian as a Srd pronoun and unit. It occurs as a definitive and as the unit in Lau and in various Asonesian systems. Australian has a third term, kol, kul, &c, which appears to be preserved in the Kol 10. For 2 the 8S, Dravirian root appears to be a contraction, ir, er (en, re), euphonically vocalised into ira, era, (eno) before the con- sonant of the possessive postfix. It enters into the Kol bar, Ul- traindian bar, mar, and the Australian bul, pul, bula, bari &c. In Australian it appears in 1 under the forms bar, wara, mara, and in 3 as mur, bar, pur, mar. The variation of the vowel from u to a which appears in the Kol and in some of the Australian terms, is found also in South Dravirian higher numbers, as well as in the postfixed definitive. The Australian terms show that the Kol compound is not a comparatively recent one, and a similar inference may be drawn from the Dravirian 3, as well as from the preservation of the same compound in the plural of the Dravirian 3rd pronoun. The §. Dravirian 3 is the labial unit repeated as in other binary 70 ETHNOLOGY OF TIIE INDO-PACIFIC ISLANDS. systems, In the original crude form of the system it must have been preceded by the term for 2. The inverted Kol form op may indicate that a partial accommodation had taken place between the labial and the prefixed term for 2 (bar-op-ia, bar-p-ia, for bar-mo- ia, bar-po-ia &c), The Australian terms coincide so closely with the South Drayirian that it might be supposed they preserve the South Drayirian possessive postfix, but itis probable that the final r, ru, ra, &e. is the numeral element, as in the lower numbers and in the Kol 2, The Wiradurei bula-ngunbai (2, 1) preserves both terms. ‘The Kamilarai ngoro has rejected the term for 2 like the Dravirian words. The Peel River pur-la preserves the word for 2 (pul-ar), varied by a slight inversion,and rejects thatfor 1. The Erub like the Wiradurei is 2, 1. The South Dravirian 4 is binary, 2,2. The Kol-Ultraindian is probably also binary, In Australia some of the languages, with Erub, have 2, 2, and others 3,1, The Kamilarai ran of wa-ran is a flexional reduplication of the ara of bulo-ara, 2, as the Dravirian nal, non is of ra, no 2. The two modes of expressing 4 probably prevailed in India prior to the Asonesian migration. The Australian numeral system is identical with the basis of the Dravirian, The full Kol terms shew this identity more clear- ly than the 8S, Dravirian. The first five numerals are repetitions of the same labial-liquid root, and the Australian system explains how this arose. 2 was 1,1; 3 was 2,1; 4 was 2,2; and 5 was 1. By the dropping of some of the terms in the comjounds, and by variations in those that were retained, each numeral ultimately acquired more or less peculiarity in its form. The Kol series re- sembles the simpler Australian, such as the Peel Riy. peer 1, pular 2, purla 3. But most of the Australian have 2 elements, and thus resemble the S. Dravirian more than the Kol. The forms of the numeral roots, and their relation to the forms of the pareat definitiyes in the different dialects of Dravirian, show that there has been some displacement in most of these, An assimilative process has been in operation more than once, with relation to the 3rd pronouns as well as to the numerals, to the possessives and other particles, and to many substantive words. It may be possible to trace from what dialect va, and not ma, vo, wu &c., became the prevalent 3rd pronoun,—du, ru, tu, the most prevalent possessive and qualitive both in pronouns ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 71 and namerals, and not the Ancient Tami! and Kol iya, ia &e.,— on, the common form of 1, and not the labial or the other forms of the nasal definitive, an, yan &¢., - mu the numeral 3, and not vo as in 1, &e. &e. Such a research into the dialectic history of Dravirian would carry us beyond the scope of our present enquiry, fruitful thongh it probably would be in data illustrative of Aus- tralian and early Asonesian philology, From the accordance between the definitive and numeral sys- tems both in Dravirian and Australian, it is clear that the latter system is equally native with the former in its elements and in their combinations in the lowernumbers, Any foreign affinities not due to the spread of the Dravirian terms themselves, must hence be considered as indications not of a derivation of the numerals from another formation, but of a primary community of roots between Drayiro-Australian and certain other archaie lan- guages. Such affinities go beyond the history of Dravirian in all its later pre-Arian stages, and even beyond its erude Australian — stage. They are vestiges of a period when the mother Draviro Australian language was, in roots at least, only one of the dialects of a formation that was subsequently to be variously modified and developed in different regions and under different influences. The superimposed quinary and denary systems, with the Dravirian mode of forming 8 and 9, indicate affinities belonging to much later periods, The civilization which originated them was unknown to Draviro-Australian at the time when the early Asonesian migrations took place. It may be possible to connect their troduction with that of other words indicative of a range of ideas and of art above the Australian, and to find in them traces of a pre-historical intercourse of other civilised Asiatic peoples with the ancient Indians. The gradual departure of the Indian physical type from the Australian towards the Seythico-Semitic may also be found to synehronise with the progress of the changes in the vocabulary. The Dravirian systems have no decided affinity with the adjacent. Ivanian, Semitic or Caucasian. But several of the terms belong to ancient Asiatic formations whieh appear to have predominated prior to these. The terms in question are found in the Ugro- Koriak languages on the North East, and in the Semitico-African ou the South West. J 72 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. province its occurrence in 4 is also explained by its presence in 2 (vi, ar, li, &e). It is found in both the forms of 2, na, ni, &c. and far, fur, &e. contracting to ar as in the corresponding terms for 2, The Indo-European t-var,—in which the dental appears to be a distinct element as in 3, ¢-ri, and 2, d-wa,—contains the same root, and is a similar form tothe Seythic d-wa-ta, d-u-r-ta &e., the d-wa of 2 being the same term, with the liquid elided. From the distribution of the liquid it is probable that it was current in some diffusive Mid-Asian system before it spread as 2 and 2 dual to India, Africa and Northern Asia. The Kol labial 4, isa similar binary term to the Semitico-African fur &e. The 8S. Dravirian 5, seems to be also purely native. The Kol labial term has affinities with those Scythico-African systems in which the labial unit recurs in 5 and frequently in 10 also. The 8. Dravirian labial 10 is a common Aso-African application of the labial unit. The archaic African forms in 5 and 10, pu-na, po-na, mo-n, fu-n, bu-re, ma-r, yu-lu &c. and the forms of the same term in 1, 2, 4, 6, &c. (mal, bar, wan, mo-r, wo-ro, ke. &c.) resemble the Dravirian more closely than the Scythic in which the final element is usually the sibilant. The expression of 5 by a unit, und the formation of higher terms by using 5 as the radix (now yenerally elided or understood), appears to have preceded the denary scale in every province of the Old World save the Austra- lian. In most of the formations of Asia the quinary system is found either as the ultimate one, or with some of its terms keeping their place under a decimal system. It is still very prevalent in Africa, and many of the African systems, like some of the Asonesian, Ultraindian and N, Asiatic, have the quinary terms entire and undisguised. ?. The formation of lower numerals by subtraction from higher, is found in many systems in different parts of the Old World, (Asia, Africa, Asonesia), and also in America. That of 8 as “2 short of 10” is less common than 9 as “1 short of 10.” The fact of such a tcrm for 8 being common to Ostiak, N.E. Asiatic and to some Indo- esiun languages was remarked by Dr Peacock in his excellent treatise on arithmetic. In several of the N. Asiatic languages hoth the quinary and denary modes of expressing 8 and Dare used. In Aino-Kurilion all the numbers between 5 and 10 are denarv, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, e7 6 (4, 10), 7 (3, 10), 8 (2, 10),9 (1, 10). In the Semitico-African systems, terms for. 7, 8 and 9, formed in the same mode, occur in several languages, The combination of servile definitives with those which are used as numeral roots, is common to nearly all formations, al- though in many of the agglutinative and flexional the two elements are more or less concreted, abraded and disguised, and the accord- ance between the postfix and current possessive or qualitive parti- cles has seldom been preserved. The Dravirian postfixes -du,-ra &e, di, -ti, -ji &c, and -ia. are not prevalent in the Seythic numeral sys- tems. They are Caucaso-African. In the Semitico-Libyan systems the dental is a common postfix with numerals. In that formation it has acquired a feminine power, but it appears to have been origin- ally eommon. From these notices it appears that the Bravities system in its ultimate definitive roots, in its successive developments or acquisitions of binary, quinary and decimal modes of nume- ration, in the mode of expressing the numbers immediately below 10 with reference to it, in the recurrence of the unit to express 5 and 10, and in the use of servile definitives with the numeral roots, resembles most other decimal systems in the world. The roots are found as definitives in many other formations (Seythic, Tibeto- Uliraindian, Caucasian, Semitico-African); and in many other languages they are also used as numerals and numeral ele- ments. The Dravirian system has this peculiarity, that in Asonesian languages we have its purely binary stage preserved to thisday. Until all the Aso-African and the connected Ameri- can numeral systems have been thoroughly analysed and compar- ed, it does not appear possible to trace the later developments of the Dravirian to their historical causes. The system certainly has not been borrowed from any of the Jater dominant races of 8S. W. Asia on the one side (Iranian, Semitic, Scythic), nor from the Chinese on the other, It has elements in common with most of these systems, and it must be considered as equally archaic and independent, Its connection with them must be exceedingly remote. It belongs to an era when neither they nor Dravirian had taken their existing forms. The numeral application of the definitives probably originated in a proto-Scythic formation, like 74 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. The labial unit is found in the N, and E. Asian systems, but it is much less common as a definitive and unit than the sibilant, (varying to dental, guttural &c.), From its more general occur- rence in some higher numbers than in 1, it is probable that it was of greater importance in an archaic stage of the Scythic systems. It is sti'!l found as 1 in Japanese, Turkish, Tungusian and some Ugrian languages. As 2 it is found in remote Eastern languages, Namollo, Korian and Japanese, and as an element in some Ugrian terms. As3it is Japanese. It does not occur as 4, In 5 it is Kamschatkan, Koriak, Ugrian and Turkish. In 6 itis found in Japanese and Samoiede, and as an element in Namollo (2) and Ugrian (1); in 8, Namollo, Chinese and Ugrian; in 9, Namollo and Ugrian; in 10, Kamschatkan, Aino, Tungusian, Samoiede, Ugrian and an element in Namollo; in 100, Chinese ; in 1,000, Turkish and Mongolian, Asa definitive the labial is very archaic in the N. and E. Asian languages. As a concreted postfix it is found in Scythic vocabularies. In Yeniseian it is still current as the 3rd pronoun, bu, ba-ri. Turkish also preserves it in bu, and Samoiede in pu-da, py-da ke. [Sce the remarks on the Draviro-Australian 3rd pronoun, ante p. ] In the Scythiec languages the sibilant (or guttural) with the liquid postfix pre- dominates as the 3rd pronoun,—son, sin, kini, tha, sya &c. The history of the labial unit and definitive in, the Semitico-African systems is of asimilar tenor. In the Semitic branch it is only used as an ordinal, the cardinal being the common Seythie and Indo- European guttural, aspirate &c, In Africa several languages retain it as the cardinal, and it reappears in higher numbers. The common form wal, war, bar, bari, mal, &e. is the same as the Draviro-Australian. The Turkish bir, a variation of the Seythic bis &c. of higher numbers, is a similar form, The prominence of the labial, and the absence of the sibilant, unit is one of the chief peculiarities of the Draviro-Australian system when compared with the N. and N. E. Asian, the Caucasian, the Indo-European and the Semitico-African. In this respect it appears to preserve a more antique character than those in which the labial has given place to the sibilant &c. The Draviro-Asonesian nasal 1 is Indo-European, Mongolian, Samoiede and Koriak. It is referable in these formations, as in the Draviro-Australian, to a pronominal root. Semitico-African ’ 4 ETHNOLOGY OF TILK INDO-PACIFIC ISLANDS. 7 has la variable in higher numbers to le, ne &e., but it is very rare and may be frum the labial, ba, bal, bar, ban &c. The Australian kol, kul &e. and Kol gel, gil, gul is a unit and Srd pronoun in N. E. Asian languages, and it is also found in the Semitico-African numeral systems, The Draviro-Austrulian contracted root for 2 (ir, ar, ra &c.) is Chinese, Japanese, Ugrian, Caucasian, Indo-European (in 4), and Semitico-African. The combination with the labial as the initial element oceurs in Namollo (mal), and it is common in Africa, which it appears to have belonged to a predominant Semitico-Libyan numeral system, of which the Semitic, in its present condition, may be considered as a remnant. Semitic in its existing form has the sibilant and not the labial initial (ath-in, si-l, ta-r &e.), in this resembling Mongolian, Tungusian, Samoiede and Caucasian terms (si-ri, ds-ur, ko-ir &c.). But in 4 it appears to preserve a contracted fourm of a common African term (ba-r, ma-l, ba-ni, bi-ni, bi-ri, vi-di, fu-la), identical with the Kol- Australian. In the occurrence of the labial both tn 1 and 2, as well as in its form, the archaic Draviro-Australian system is cognate with the archaic Semitico-Libyan. In the general dual and plural force of the second element, n, 1, r &c. they also resem- ble each other and Seythic. In all the formations this generic application appears to have arisen from the use of the particle as a numeral. The Draviro-Australian 3 is peculiar. The labial does not appear to occur as a root for 3 in any of the Aso-European or African systems, save in the Turkish wise and Japanese mi (whence imu 6, i. e. 3 dual). In the other systems the sibilant unit has as much currency in 3asin1. In its double form, or with the second element as a liquid, it is common to N. and E. Asian, Indo-European and Semitico-African systems. In this numeral Draviro-Austraiian shows its primitive and_ persistent churacter more even than in its 1 and 2. TheS. Dravirian root for 4 is Ugrian and Semitico-African, and the reduplicated form is found in both of these provinces. In the Wyrian it may be referred to the Chino-Tibetan ir, il, li, ni nyi &c., as an archaic Asiatic definitive for 2, preserved in the Dravirian 2, 4 and higher numbers. In the Semitico-African, kK 76 BTHNOLOGY OF THE INDO-PACIFIC ISLANDS. the definitives themselves, and the African affinities are probably owing to derivation from a like souree. The African terms present more affinities in roots and combinations than the Seythic, but Scythic has remuants of similar forms. The African affinities conuect the Draviro-Australian quinary or basis system with the most archaic form of the Semitico-African system more closely than with any other. But the former is sim- pler and more primitive than the latter, in which the sibilant series of terms, found in all the Asiatic systems, blends with the Jabial and preponderates over it. The African systems appear to have been more influenced by the Semitic in its later gradations, and the Semitic by the Scythico-Lranian, than the Draviro-Aus- tralian by any foreign systems, The Draviro-Australian would appear to be the most faithful representative now existing of un archaic 5. W. Asian system of definitives and numerals. This system is still homogeneous, the labial being the principal current definitive as well as unit. In the Semitico-African sys- tems there are remnants of the labial pronoun, but the sibilant is now the principal one, The labial unit of these systems is heuce more closely connected with the existing Draviro-Australian, than with the existing Semilico-Lybian, pronouns. It is werthy of remark that the Euskarian, which has close af- finities with the oldest form of the Semitico-African systems, preserves a labial 1 ba-t, bo-t, 2 bi, and 5 bo-r-tz, bo-st. It is found also in 9, be-dera-tzi, and 10 ha-mar (arte, sec. 5). The Caucasian, like the Scythic, Semitic and Indo-European, is mainly sibilant, but there are some labial remnants, 2 wi-ba Abkhasian (Eusk. bi, African bi-li, bi &c.); 3 ab-al Lesg., (but this is probably a contrac- tion of chab in which the initial is sibilant asin the Georgian sa-mi &e); 4 wor-ts-teho Georg., mn:uk-ba, boo-gu, ohw-al Lesg., p-shi-ba, p-tle Cire. ; 5, wo-chu-si Georg., p-chi Mis.; 6, f-ba Abkh.; 7 s-wi-di, &e. Georg., wer-al Lesg. buor, uor-2, uosh Misj., b-le, Cire., bish-ba Awar.; 8 rwa, ruo, &c. Georg., mitl-go, mek-go, betel-na, beet!-gu, mei-ba &e. Lesy., bar, bar M'sj. ; 9, b-gu, boro Cire.; 10 wit, with Georg., wez-al Lese. p-she Cire. In some of these terms, however, the labial is probably prefixiual. The labial system would appear to have predominated in 8. W. Asia and spread thence to India and Africa before the ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 77 sibilant acquired its present prominence. Both terms may have co-existed as definitives and units in the oldest pronominal and numeral systems, although their relative importance varied in different eras, The acquired sexual application of the two defini- tives, and the proneness at one time to extend the application ofthe masculine and at another that of the feminine to inanimate sub- stances, would account for this, The later tendency to throw off the distinction of gender, and to retain only the form in most com- mon use, ends in a still greater impoverishment of the original variety of forms and terms. The Draviro-Australian, like the Tibetan and some other Asiatic systems, has no trace of gender in its labial definitive. In the Semitico-Libyan the labial and sibilant appear to have been also originally common, but at an early period the ‘former became masculine and the latter feminine. The system may be considered as of equal antiquity with a very archaic formation which was diffused on the one side as far as Africa, and on the other over Central und Eastern Asia, Although the system, both in its terms and in the principle of its forma- tion, has affinities with other languages, it cannot be derived as a whole, or even in the bulk of its materials or in the model of its construction, from any other now extant. The affini- ties, however, point distinctly toS. W. Asia more immediately, and to an epdeS. anteriour to the diffusion not only of the Semitico- Libyan and Iranian but of the Caucasian systems. It appears to be of the same archaic origin as the basis of these systems them- vessel and of the other systems which were dispersed over Asia before the former began to predominate. The Ugro-African aflinities of the Dravirian establish this. There is another test of its relative ethnic position. The remotest and least advanced Asiatie and American systems have only terms for 1 and 2, for 1 2 and 3, or for 1, 2, 3 and 4. This may be suid to be the case with that of the Australian formation, the general Dravirian affinities of which are strong. The Australian proves that the primary Ugro- Drayirian formation prevailed in 8. W. Asia, including India, at a barbarous epoch, prior to the expansion of the simple numerals 1, 2, 3, into higher binary and ternary terms by combination and acquired flexion, a process: which preceted the adoption of the quinary and denury scales in S. W. Asia, as is testified by the 78 ETHNOLOGY OF THE LND)-PACIFIO I®RLANDS. Iranian, Semitic, Caucasian and other Asiatic and African systems retaining terms so formed. The Dravirian numerals belong to the same era of S. W. Asian civilisation that gave birth to these improved systems, and they must therefore have been broughit into use in India long subsequent to that period of its history represented by Australian civilisation. The denary system was not imported by the earliest race, whether Negro or Australian, which laid the foundation of the Indian lunguages, but by a subsequent race from S. W. Asia, whose civilisation was connected with that in which the subsequent Semitic and Tranian diffusions originated. The Dravirian numerals are not derived from any of the leading Asiatic systems, and their connection with these is extremely remole, The Turkish ard Uyrian systems are nearer to the Caucasian on the one side and to the more remote N.E. Asiatic on the other, the Iranian is nearer the Semitic, and the African are nearer the Semitic, the Iranian and the Scythie, than the Dravirian is to any of them. The introduction of the denary scale into India is probably connected with the advance into it of one of those Scythoid races of partially Irano-Semitic character, the archaic influence of which on the pliysical form of the Soutli- ern Indians is so observable. The Todas may be nearly pure descendents of the very race which imported the system. * Ill. MISCELLANEOUS WORDS, For the miscellaneous glossarial comparisons of the Ultraindian and Indian division of the present enquiry, it will be convenient to take the list of sixty miscellaneous substantives originally com- * Dr Stevenson in his “Collection of words from the Toda Janguage” (Journ. Bombay As. Soc. i, 155, for 1842) gives some foreign affinities. lor 1 he adduces the Latin unus, Tungrs. mukow, Koibal unem, 2, Tungus. djuhr, Arm. yergu. 5 Chinese ing. 6 turkish alti, Yenis.ram, azam. 7 Arm. yeotn. 6 Arm. ut, Lat. octo, Eng. eight, Sansk. ashta. 9 he exp/ains as 1 less |v. 10 he compares with the Tibetan bachu, bet [the true Tib. form is 6chu in which b is prefixuai and uncon- nected historically with the Drav. labial root, save in so fur that both are ulti- mately the same EE BRS . The Rey. Bernhard Schmid, in his “ Essay on the Relationship of ruages and Nations’ (Madras Journal v, 133) had also previously (1837) given tubles in which ithe Dravirian numerals are compared with a great variety of foreign ones, but his affinities are too indiscriminating. As I had not read th vad i when my comp. voc. was printed L give his list (p. 167) of the Yoda terms, which contuins some variations not fouud in my voc. 1 odd, corresponding with Dr Stevenson's orr and a contraction of vodda, 2atu, ait. 3 muthu, mad, 4nalk, nank. 5 uy. 6 or, ol. 7 or, ud. 8 otthu, 9 unboth. 10 poithu, 11 ponnod &e, { some remarks on Dr, Miiller’s comparisons of Dravirian with Scythie nu- nivials will be jound in another place. } : RTHNOLOGY OF THE INDO-PACIFIC ISLANDS, 79 piled by Mr Brown in twenty two Ultraindian and East Hima- layan languages, and to which other Ultraindian and many Gangc- tic languages have been added by Mr Brown himself, Captain Phayre, Mr Hodgson and others. Mr Hodgson has adopted this list of ‘substantives for his series of comparative vocabularies, adding to it a large number of words of other classes. I have useil the vocabularies of the South Indian languages compiled for him by Mr Walter Elliot and others, and which have been al- ready mentioned in another place, but I have also taken words from my own smaller comparative vocabulary of above 300 words in the compilation of which all the vocabularies and dictionaries within my reach have been availed of. It will be borne in mind thatthe present paper is mainly directed to phone- tie and grammatical affinities, and that the vocabulary in question belongs to the glossarial branch of the Asonesian affinities which will be examined separately. I do not of course assume that the absolute glossarial affinities of the Indian and Ultraindian languages will be accurately represented by the results of an examination of Mr Brown’s 60 substantives, and of the pronouns, particles and numerals which have been already adverted to. A collection of whole vocabularies will probably greatly diminish the amount of agreement, because most of Mr Brown's words are of classes that are very subject to diffusion and displacement. It is totally deficient in those words exprersive of the most generic actions and attributes which appear to me to be more persistent than other. The followiny is Mr Brown's vocabulary. I have added numbers in order to save the repetition of words in some of the comparative lists, given in the next chapter.* 1 Air 5 Blood 9 Cat 2” Ant 6 Bont 10 Cow 3 Arrow 7 Bone ll Crow 4 Bird 8 Buffalo 12 Day = I have only been able to compare ahout 40 terms in the list with a large range of foreign vocables. Two of them “ Name” and ** Village” are not ine uded in my own compurative vocabulary, and several of the others, such as Ant, Buffalve, Biephant, Fiower, Goat, Hog, House, Light, Monkey. Musquito, Oil, Plantain, Koot, salt, Skin, Snuke, Tiger, Tooth, Yam, are not included i most of the shorter of those vocabniaries which have contributed to itscompilation. The omission is espe- cially to be regretted in the case of many of the Scythic vocabularies in Klaproth's great collection. 80 ETHNOLOGY OF TIE INDO-PACIVIC IBLANDS, 13. Dog 29 Horn 45 Plantain 14. Ear 30 Horse. 46 River 15 Earth 31 House 47 Road 16 Egg 32. Iron 48 Salt 17. Elephant 33. Leaf 49 Skin 18 Hye 34 Light 50 Sky 19 Father 35 =©Man 51 Snake 20. Fire 36 Monkey 52 Star 2L_ Fish 37. Moon 53 Stone 22. Flower 38 Mother o4 Sun 23. Foot 39 Mountain 55 Tiger 24 Gort 40 Mouth 56 Tooth 25. Hair 41 Musquito 57 =Tree 26 Hand 42 Name 58 Village 27 Head 43 Night 59 Water 28. Hog 44 Oil 60 Yam In estimating the per-centage of affinities IT have added 40 words of different classes to complete the hundred. A numerical mode of stating the amount of agreement has been adopted be- cause it is the most definite whatever be the extent of the vocalularies collated ; but the value of the result varies of course with the kind and number of the words compared, and all deduc- tion from purely glossarial data must be taken in combination with the evidence of other kinds as to the past and present relations of the tribes themselves. The absolute proportions obtainable from a comparison of entize vocabularies will probably differ greatly from those derived from 100 words. But the relative proportions will not be affected in an equal degree by enlarging the basis of comparison. For example the affinity of the South Indian vocabularies with the Gond may prove to be only 25 per cent. But if so that with the Kol will probably be reduced in a proportion not very dissimilar, so that the relative amount of the Sonth In- dian affinities of the Kol and the Gond will not be seriously aff cted. Iu tracing the glossarial history of any formation we must begin with the modern changes. For general ethnology also this is the best course, because the only scientific principle that can guide us in Our enquiries into pre-historic events is that nations and their ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 81 languages have always been subjeet to changes similar in kind to those whieh are now going on in the world, and have been doing so throughout historical periods. In India the Dravirian formation has ceased to be diffusive and assimilative. It has long been exposed to the influence of the Sanskrit and of the northern Indian tongues that were early assimilated in a greater or less degree to Sanskrit In the Dra- virian family we have therefore to note the mutual action of the different languages and dialects, and the action on each of the Sauskrit and of the Sanskritised or prakrit tongues of the north. The very close degree in which the Dravirian languages of South- ern India are related to each other and to the least Uliraindianised lunguages of the Vindyas, in phonology and idevlogy, has appeared from the details in chap. 1V. They are dialects of one tongue, and they appear to differ less from each other than the Philipiue languages. ‘he dialectic discordances are exactly the same in kind as those which prevail amongst the Philipine and other groups of Asonesian languages, or amongst the Asiatic members of the Semitico-Libyan formation. All the great families that have been recognized show much larger mutual deviations in their com- ponent languages, and we must include Australian to “give the iormation a comprehensiveness similar to the Scythic™ or the Semitico-Libyan. Even the Indo-European and the Malagasy- Polynesian are much more diversified than the continental or Kol-Tamulian division, All these widely disseminated families present single languages or groups that, from long and complete separation, have become alienated from each other in the greater number of their roots, in phonology and even in many details of ideology. The transitions are seldom so abrupt as from the Dravirian to the Australian, but this arises from the former beiig only the last continental and the latter the last insular remnant of a once continuous and widely expanded family, that was early disjoined, and has ever since been subjected in its two divisions to the influence of formations of opposite churacter,—the Seythico- Iranian tending in the continental division to give a more flexional development to the primary structure which it has in common with them,—and the Niha-Polynesian tending to arrest the natural flexional development and concretion of the insular division, and 82 ETHNOLOGY OF THK INDO-PACIFIC ISLANDS. to niaintain the archaic erudeness of the type while partially trans- forming it. But some of the contrasts found in other families are us great or nearly as great in degree, as, fur example, that between English or Celtic and Sanskrit,—Semitic and Egy ptian,x—Malay and Tagala—Polynesian and Mualagasy,—Manchu and Fin &e. Of such degrees of dissimilitude as that between the Kol and the proper Dravirian group inmost. large and partially mixed families present several exaniples. Ou the subject of the connection amongst the South Dravirian Janguages Mr Ellis’ observations may be cited. “ The Telugu, to which attention is here more specially directed, is formed from iis own routs, which, in general, uve no connexion with the Sanskrit, nor witli thuse of auy other language, the cognate dialects of Southern India, the Tamil, Cannadi &c. excepted, with which, allowing for the occasional variation of con-similar sounds, they generally agree: the actual difference in the three dialects here mentioned is in fact to be found only in the affixes used in the formation of words from the roots; the :oots themselves are not similar merely, but the same.” (Note to the Introduction to Campbell's Telugu Grammar, p. 3.) It must at the same time be remarked that for many ideas there is more than one native or at leust pre-Sanskritic root current, and that the different vocabularies even of the southern group otien affect different roots. ‘This feature does not militate against the assertion that the disparities are merely dialectic, for it is common to tie Druvirian with every other ancient cluster of dialects. As in other provinces, the capacity fur the cur- reney of numerous rovts was probably much greater in the earlier ages of the family, when its tribes were more barbar- ous and more divided, The progress of the great civilised nations and their mutual glossarial interpenetration and assimilation, must have been attended, as in other cuses, with the partial obliteration of the vocabularies of subdued or ubsorbed tribes. In the primary Draviro-Australian era, the number of distinct vocabularies and independent synonimous rvots was probably very great; and the difference between the Kol aud the Goudo-Tamulian yocabularies shows that in India, even to the latest period of Dravirian predo- winance, the North-Eustern dialects presented a considerable cou- ETHNOLOGY OF THE INDO-PACIFIC ISLANDS 83 trast to the southern. So long as dialects spoken by independent or separate tribes exist, the number of roots has a tendency to inerease, each dialect being a distinct inlet for foreign words, which may or may not pass by slow degrees into circulation in some or all of the other dialccts also, that depending on the nature of the relations amongst the tribes. A gradual and very great glossaiial divergency is consistent with the retention of the leading characters of the formation in phonology, ideology and even in glossary. The Indo-European, the Niha-Polynesian, the Tibeto- Uliraindian, the Scythic, the Semitico-Libyan and all other forma- tions furnish evidence of this, On the subject of the more recent interpenetration of the South Dravirian glossaries Mr Etliot remarks: “All the southern dialects become considerably inter- mixed as they approach each other's limits, Thus the three words for “egg” used indifferently by the people speaking Canarese, (matté, tetti, gadda) are evidently obtained, the first from the Tamulian, matta; the last, from the Teligt, gadda. This inter- mixture, which is of ordinary occurrence in all cognate tongues, is here promoted specially by extensive colonization of different races, as of the Teligts into Southern India under the Bijaynagar dynasty, where they still exist as distinct communities—and of the followers of Ramantja Achdrj into Mysore, where they still are to be seen as a separate class speaking Tatil in their families, and Carndtaca in public. The Reddies also, an enterprising race of agriculturists, have migrated from their original seats near Rajah. mandry, oyer the whole of Southern India, and even into the Maharashtra country, where they are considered the most thriving ryots, and are met with as far north as Poona.” (Journ. Asiatic Soe. vol. 18 p. 350). So far as the testimony of the 100 words which I haye compared can be relied on, the South Indian or purest Dravirian yocabu- laries would appear to have 30 to 40 per cent of their words in come mon with Gond, Male and Uraon; and less than 15 per cent with the Kol dialects. The specific affinities with the Middle Gange~ tic, the Himalayan and the Ultraindian languages, though considers able as a whole, are so slight for any particular language or group, that it would be unsafe to state then at even a very low number, without a comparison of much larger vocabularies. A few Dra- L 84 ETHNOLOGY OF THE INDO-PACTYIC ISLANDS. virian words are found in Dhimal, some of the Manipuri dialects and Burman. . The Gangetic vocabularies of the Tibeto-Ultraindian and Tibe- tanised class have many words that appear to he archaic Indian or Draviro-Australian, although not now extant in the South Dravirian Janguages. The most western, as Tiberkad and Milechanang, pre- sent affinities with Eastern Medo-Persian vocabularies. Some of their non-Tibetan terms are clearly, ancient Gangetic, for they are found in Asonesia. The affinities with any single Asonesian language are few, but with the Asonesian vocabularies as a whole they are perhaps more numerous than with those of any other province save the Seythic in its widest range (Caucaso-Koriak). The Australian affinities are far from being the most numerous.. Dravirian vocables are found in all the Malayu-Polynesian languages, and as several Dravisian synonyms and varieties of the same root ure extant in different vocabularies although not found in Australian, it appears that the Dravirian glossarial current not only set to the eastward in the fiist Australian era, but continued to do so while changes were taking place in the Indian languages themselves, or in the distri- bution and predominance of the tribes who spoke them. In the earlier ages of this current it must have chiefly flowed from Bengal along the western seaboard of Ultraindia, and it is to be presumed that the dominant tribes and vocabularies of the Lower Ganges were more or less changed from era to era by the intru- sion of other Dravirian tribes from the interior, and by foreign influences transmitted from Irania. In later periods they were affected not only by the ethnic current from Irania down the Gangetic basin, but by the Chino-Tibetan movement from the eastward. As soon as navigation was sufficiently improved to allow of a maritime intercourse along the coast of the Bay of Bengal, the populetion and languages of the Lower Ganges would be affected by the powerful South Indian nations and by foreign visitors from the west, while the continental and Singhalese South Dravirians themselves would then, for the first time, be enabled to carry on a direct intercourse with Ultraindia and Indo- uesia. It is probable, from glossarial evidence, that the Dravi- rians were civilised and maritime before the Arians predominated ETHNOLOGY OF THE INDO-PACIFIC IsLANDs, 85. in N. India. The influence of a Gangetic sub-formation akin to the Kol is still distinctly traceable in Indonesia, as will appear in a subsequent place. The remnants of the Dravirian formation in the other existing languages of Northern India, and especially of the Gangetic basin, are of great importance for Asonesian ethnology. It is obvious that from the first era of the Draviro-Australian move- ment towards the further east, when rude tribes like the Simangs and Australians roamed in the Sunderbunds and crept along the creeks on rafts or skins, to the period when civilised Dravirians and Ultraindo-Dravirians navigated the coasts in parayus and spread their maritime art to the remotest islands of the South Sea, the Gangetic population must Lave been the principal, and, in general, the sole, disseminators of Indian vocables in that direction. Hence a knowledge of the Gangetic tongues in every age, and under each of the great changes they have under- gone from the influence of intrusive formations or languages, is essential to a thorough investigation of Asonesian history, and whatever vestiges are recognized of their pre-Sanskritic con- dition and possessions have an immediate value for that purpose. It has already becn remarked in an earlier page, that not only the Vindyan dialects but the Marathi-Bengali or Sanskritised lan- guages of Northern India, present, in their non-Ariun clement, proportionately more numerous and direct affinities with the Indonesian languages than the South Dravirian. The glossarial and other affinities between the Asonesian formations and the Dravirian will be separately examined. It is sufficient here to indicate their existence and extent in proof of the greut antiquity of the latter in India, and of its having exercised a predominant influence in the eastern archipelago not only prior to the Papuan era but subsequent to if, for the Malayu-Poiynesian civilisation was not purely Ultraindian or Chino-Tibetan but Gangetic or Draviro-U ltraindian.* * Several examples of this class of affinities will be fuund in the annexed voca- bulary. 1 take a dove words at 4a ert peies classes. The Dravirian sorta, sariada, satin: ta,its Avian. Tinnaga Teluy, is sprend over Asonesia from Nias to Polynesia, (e.g. atula, atilu, tian, tatonu, etal &e. It is connected with the Tibeto-Himalayan thang, tong, tomlo, thunea, Naya ating, Anam thang, The Dravirian nere, nerana is probatiy also the original of the Tadoucoteat no-lor, lur-us, meruru, &e., Poly. porore. a6 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS: Unlike the Gangetic and Ultraindian vocabularies, the Dra- virian have little direct ¢onnection with Tibetan. The southern dialects do not appear to have any. The course of the Dravirian etirrent has evidently been fiom N. W. to S. E. along the south- ern foot of the Himalayas and not across them, and its immediate origin is not to be sought in the Tibetan direction. The following words in the list show Tibetan affinities, which are probably all or nearly all archaic or extra-Indian in origin, —6, 10?, 11, 19, 33, 87, 38, 892, 41, 44, 46, 51, 57, 59. Having thus glanced at the eastern affinities of the Dravirian vocabularies we may pursue the enquiry tothe N. W. The in- fluence exerted by the intrusive Sanskrit is the first to be consi- dered, The vocabularies of the existing Northern languages of India,— Konkani and Marathi, Guzarati, Hindi in its various dialects, Kashmiri, Bengali and Uria—are Sanskritic. A small proportion of the words, estimated at an eighth to a tenth, are non-Sanskritic, Croohed. The Dravirian koniya, konal, konalu, is Arian (kona &c.) Another term which has also Arian affinities but appears to be pre-Aryan is wanukara Telug., banka Bengali, bengke Uraon, leko Newar, bango Sunwar. The root is probably wang, beng &e. as it occurs in the Telugu wompy, Mulayalam valanga, Tadava mon, Chinese wan Polynesian wana, Anstralian balbal, wali &c., Nias abelu, Bawian belo, Bisayan bali-ku &c. Malayu belo (tack), Timor peno &e. But, as in numer- ous other instances, the Dravirian tfix has been imported into Indonesia a3 & substantive part of the word. The Uraon bengko is identical with the prevalint Indonesian bengko, and the Newar beko ccrrespond closely with the Celebesian peko und Polynesian biko. Round. The Dravirian urundu, urutn, &e, isa common Arian &c. root, but it is proba- bly pre-Arian-India, ‘Tib. lumpo, rirl, Burm. lung, long, Koria lung-kur, Indo- nesian limbung, ‘the Himalayan burbur, and Male bevo, Kiranti anbo, are Aso- nesian, bulu Ende, poepoe Polynesian, abola Nias, bulat indonesian, but bul, bur &v, is ulvo Indo-turopean, African&ec, ‘I'he Gon! moto may be Arian, but it has a strong jesemblance to the E, African (Makua) and Polynesian poto, Few, Smal. These words belong to a class which is nearly os persistent as the definitives and directives, and which is open to the same objections when used in ethnic com- parisons. Some of the Indonesian affinities however are so striking that I will ive them. Bengali kinchit, Bodo kitisi, tisi “few” Naga tesu, Burm. tiche, Kurnutake tusa, to Dhimal atoisa, &c3 ‘Turkish kitehi, Singlu katsi ; Malay kitehi, katehi, (‘‘small’’). Tibberkad zigit “few”, Gurung chigide, Kiranfi chichi, Indonesian sikit, sidikit, sakide, chuch, che &e, Sunwar iske, Naga ishika, Keren siko &e. Indonesian siku, sakui, usi. Most of the other numerous Asonesian terms lave also direct continental affinities, Ultraindian, Caucasian, N, and E. Asian, African or Iranian. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS a7 and amongst these Dravirian roots are found.* Most of these are probably remnants of the Dravirian basis of the North Indian languages, like the structural traits and some of the particles noticed in Section 7. The influence of Sanskrit on the southern languages has been much smaller, but it is nevertheless considerable, On this subject Mr Campbell has remarked: “The third class of words which is generally mentioned by Dr Carey as “derived from the Sungs- krita,”” I have named Sanserit corruptions; it cousists of words which have passed into Teloogoo, either directly from the Sanscrit, or through the medium of some of its corrupted dialects, such as the Pracrit, and which, in order to be assimilated to the language of the land, have undergone radical alterations, by the elision, insertion, addition, or subtraction of letters, These changes have been sometimes carried so far, that it is difficult to trace any con- nexion between the adulterated word and its original in Sanscrit.”’ “The reader will find all words denoting the different paris of the human frame, the various sorts of food or utensils in com- mon use among the natives, the several parts of their dress, the compartments of their dwellings, the degrees of affinity and con- sanguinity peculiar to them, in short all terms expressive of primi- tive ideas or of things necessarily named in the earlier stages of society, to belong to the pure Teloogoo or language of the land. It is true, (so mixed have the two linguages now become) that Sanscrit derivatives or corruptions may, without impropriety, be occasionally used to denote some of these. This, however, is not common; the great body of Sanserit words admitted into the language consists of abstract terms, and of words connected with science, religion, or law, as is the case, in a great degree, with the Greek and Latin words incorporated with our own tongue: but even such Sanserit words as are thus introduced into Teloogoo are not allowed to retain their original forms, they undergo changes, and assume terminations and inflections unknown to the Sanscri’, and, except as foreign quotations, are never admitted into Teloogoo * Dr Stevenson is still investigating this subject. Since the earlier chapters of this paper were published two portions of a comparative vocabulary of non-Sans- krit words in the Indian vernaculars have appeared in the Journal of the Bombay Asiatic Society, vol. iv p. p. 117, 319 (1852-53), and to these I may refer my read- ers for examples of yocables common to the purer and to the Sanskritised languages of India, . 83 ETHNOLOGY OF THE INDO*PACIFIC ISLANDS. until they appear in the dress peculiar to the language of the land.” (Introduction to Teluga Gram. p. p. xix, xx.) Mr Ellis, in his note to Mr Campbell’s Preface, has the follow- ing observations on this subject. ‘In the preeeding extracts, the author, supported by due authority, teaches, that, rejecting direct and indirect derivatives from the Sanscrit, and words borrowed from foreign languages, what remains is the pure native (anguage of the land: this constitutes the. great body of the tongue and is capable of expressing every mental and bodily operation, every possible relation and existent thing; for, with the exception of some religious and technical terms, no word of Sanserit derivation is necessary to the Telugu. This pure native language of the land, allowing for dialectic differences and variations of termination, is with the Telugu, common to the Tamil, Cannadi, and the other dialects of southern India: this may be demonstrated by compar- ing the Desyam terms contained in the list taken by Vencaya from the Appacaviyam, with the terms expressive of the same ideas in Tamil and Cannadi, It has been already shewn that the radicals of these languages, mutatis mutandis, are the same, and this com- parison will.shew that the native terms in general use in each, also, eorrespond.”—( p. 18). “From the preceding extracts and remarks on the composition of the Telugu language, as respects terms, it results that the language may be divided into four branches, of which the following is the natural order, Desyam or Atsu-Telugu pure terms, constituting the basis of this language and, generally, also, of the other dialects of southern India: Anyadesyam ferms borrowed from other countries, chicfly of the same derivation as the preceding: Tatsamam, pure Sanscrit terms, the Telugu affixes being substituted for those of the original language: Tadbhavam, Sanscrit derivatives, received into the Telugu, direct, or through one of the six Pracrits, and in all instances more or less corrupted. The Gramyam (literally “the rustic dialect,” from Grdémam Sans. a village) is not a constituent portion of the language, but is form- ed from the Atsu-Telugu by contraction, or by some permutation of the letters not authorised by the rules of Grammar. The pro- portion of Atsu-Telugu terms to those derived from every other source is one-half; of Anya-désyam terms one-tenth ; of Tatsamam ‘terms in general use three-twentieths; and of Tadbhavam terms one quarter. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 89 “With little variation, the compostion of the Tamil and Cannadi are the same as the Telugu, and the same distinctions, consequently, are made by their grammatical writers. The Telugu and Cannadi both admit of a freer adoption of Tatsamam terms than the Tamil ; in the two former, in fact, the discrétion of the writer is the only limit of their use; in the high dialect of the latter those only can be used, which have been admitted into the dictionaries by which the language has long been fixed, or for which classical authority can be adduced; in the low dialect the use of them is more general— by the Brahmans they are profusely employed, more sparingly by the Sudra tribes. The Cannadi has a greater and the Tamil a less proportion of Tadbhavam terms than the other dialects; but in the latter all Sanscrit.words are liable to greater variation than is produced by the mere difference of termination, for, as the alphabet of this language rejects all aspirates, expresses the first and third consonant of each regular series by the same character, and admits of no other combination of consonants than the dupli- cation of mutes or the junction of a nasal and a mute, it is obviously incapable of expressing correctly any but the simplest terms of the Sanscrit; all such, however, in this tongue are accounted Tatsa- mam when the alteration is regular and produced only by the deficiencies of the alphabet, “But, thongh the derivation and general terms may be the same in cognate dialects, a difference in idiom may exist so great, that, in the acquisition of one, no assistance in this respect can be derived from a knowledge of the other. As regards the dialects of southern India this is by no means the case,—in collocation of words, in syntaxical government, in phrase, and, indeed, in all that is comprehended under the term idiom, they are, not similar only but the same.” (p. p. 21, 22). Mr Elliot has also more recently remarked on the aptitude of the South Dravirians to substitute Prakritic words for aboriginal ones. (Journ. As. Soc, of Beng. vol. 18, p. 350). * * The form of the Sanskrit words is much parer in the Dravirian than in the Sanskritoid languages themselves, and the reason is well explained by Dr Steven- son in ove of his recent papers. ‘In reference to the Sanskrit portion of the verna- cular languages of Intlia it is a singular fact that it is purer among the inhabitants of Malabar and Mysore than among those of Bengal and Upper India, The reason of this can easily be given, though it be not at first obvious. In Upper India, Bengal. and Gujarath, nine-tenths of the language is a corrupted Sanskrit. The 90 ETINOLOOY OF THE INDO-PACIFIC ISLANDS. In the short list of familiar words contained in the appended comparative vocabulary, several examples occur of Sanskrit or Arian terms which have gained equal currency with native ones or replaced them. Parallel phenomena are found in all voeabu- laries and are very prominent in those of nations which have had much intercourse with others superiour to them in civilization, or politically paramount, Sanskrit has been received into the glos- saries and literature of Southern India as freely as it was into those of the civilised western nations of Indonesia, or as Chinese has Been into those of Korea and Japan.* The Drayiro-Australian languages have a connection with the Sanskritic which belongs to a much more ancient period of their history than that which followed the entrance of the Arians into India. This archaic connection is probably itself susceptible of reference to more than one era and condition of the Draviro-Aus- tralian formation. The glossarial affinities between Australian and Sanskritt must belong to the earliest stage of the relationship between the two formations, because the former represents the most primitive and least modified form of Dravirian. The Austra- lian form is archaic even when compared with Dravirian, and it is still more archaic when compared with Sanskrit. The glossarial affinities may be considered as carrying buck the history of the Indo-European formation to its proto-Scythic condition. The archaic affinities of the Bengali-Murathi and proper Dravirian vocabularies with those of Sanskrit and other Indo-European Brahmans and higher c’asses there more easily fall into the prevailing pronpucia- tion of Sanskrit words, whereas in the South, the Sanskrit vocables, being rare] used by any except Brahmans or well educate! persons, the primitive forms thoug with the notable exception of the dropping of the proper marks of the geniers of noons, have been most carefully preserved,” (Journ. Hombay As. Sec. vol. iy p. 121). The Sanskrit vocables that have been adopted into Indonesian languages have asimilar comparative purity. * Dr Prichard spp to have adhered to Klaproth’s belief that there is a class of words of the first necessity which are preserved long after other kinds of words are replaced, and thus form one of the tests of linguistic affihation. W. Von Huinboldt hus more accurately remarked, ‘'1t is generally believed that the affinity of two languages is undeniably proved if words that are applied to objects whieh must have beenknown to the natives ever since their existence, exhibit a degree of resemblance, and toa certain extent this is correct. But, notwithstanding this, auch a method of judging of the affinity of la ies seems to me. by no means infaliible, It often happens, that even the objects of our earliest perceptions or of the first mere , are neerenanted by words taken from foreign languages, and which belong toa t class. + Some examples of this will be found in the appended vocabulary. Others will be given in the section on Australian. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 91 languages having the same roots, probably belong in part to much later periods, and while some are doubtless of Arian origin in the trans-Indus ages of that formation, others, it is reasonable to con- ceive, must be of Dravirian origin. If, as appears to result from ethnic evidence of all kinds, the Dravirian formation preceded the Indo-European in eastern Irania, it is very improbable that no native terms were adopted by the intrusive Arian vocabularies. It is equally improbable that in Northern India, where the ancient formation has never been wholly eliminated, Sanskrit did not receive other additions from the vocabularies of the subject and partially helotised tribes.* When we compare the various forms of roots common to the eastern Indo-European languages—those of Irania and India—with the Draviro-Asonesian, we frequently find that several of the archaic insular forms, Australian, remote Papuanesian &c., are identical with Irano-Indian forms. It is sometimes erroneously assumed that roots common to Sanskrit and Zend with the spoken Medo-Persian and Indian langnages are necessarily ori- ginal in the former and derivative in the latter, and that all the variations from. the Sanskrit or Zend forms are corruptions of them. There is no reason to believe that in archaic times one Indo-European nation, speaking one dialect, was ever so civilised, populous and powerful as to oceupy all Ira. nia. The Sanskrit-speaking tribe, when it first comes into the dawning light of history, is found restricted to a petty district in N. W, India, and it never succeeded in imposing one dialect even on the basin of the Ganges. The present vocabularies prove that dialeets preserving Dravirian ingredients of different kinds have always existed in this province. The living vocabularies of Irania afford similar evidence, for they possess roots that are not Sanskrit or Zend, in common with Indian and Aso nesian languages, and varieties of Sanskrit roots which have an equally wide dissemi- nation, A large proportion of these yocables probably existed in different Iranian dialects not only contempora neously with Sanskrit * See the remarks on this subject in the Introductory Chap. of this Part (ante vol. vi. p. p. 686-8). Dr Stevenson, in a paper which lad not reached me when these remarks were written, has alluded to the additions which Sanskrit may more recently —thut is since it ceased to be a spoken language—have received froin native words introduced into the lan, usage by provincial writers, and then adopted by lexicogra-phers. Journ. Bombay As. Soc. vol. iv, p. 119. M 92 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS and Zend during the period of their predominance, but throughout the earlier ages of the formation. Those that are most widely dispersed in the Trano-Indian and Draviro-Australian languages, ani those that are found not only in Australian and other archaic Asonesian vocabularies but in Caucasian, Ugrian, western Indo- European and African, render it certain that, even in the crude proto-Arian stage of the Indo-European formation, various dialects existed. In this stage the formation approximated to the Draviro- Australian in its general character, and when it is found that dialectic varieties of a common root are also common to existing Trano-Indian and Draviro-Australian vocabularies, it results that there was a period when the external limits of the two formations were not so far sundered as Ireland and Australia, and when the line of mutual contact was further west than the busin of the Ganges. The dialectic varieties were produced not only before the Iranian formation began to spread to the shores of the German Ocean but before the Dravirian began to move eastward on ils route to the Indo-Pacific islands. If they belong to the earliest dialects of the Drayirian formation, they must have existed before the Iranian formation took its distinctive shape. It is probable that they belong to the proto-Scythie basis of both formations. They establish an early and close connection between them, and render it probable that they were at one time contemporaneous in Trania. The further our comparative glossolory advances the more minute and accurate will be our classification of the root varicties common to the two families, But until the vocabularies have been carefully compared not only with each other but with those of all the other families of language, their full historical import will remain concealed. While many of the common Irano-Dravi- rian roots may, by the structure of the vocables in which they occur or by their distribution, be referred to Ugrian or other families, and some to more modern sources, others appear to be entirely pre-Scythic, in other words they are older than the Dravirian and Iranian formations, and older than the Seythic or proto-Scythic formational basis itself. The form of the pure root in such instances is referable to a monosyllabic condition of the family, not only because it is free from any adventitious characters derived ETHNOLOGY OF THE INDO-PACIPIC ISLANDS, 95 from the phonetic and straetural habits of other harmonic families, but becanse it is extant in these or in the monosyllabic family in a similar form, although in the former it may be concreted with a definitive. The investigation of the proper glossarial history of the formation as such, commences with the separation of this basis portion of the vocabulary from that which has been since acquired. In .the Dravirian formation this appears to be less difficult than in the Indo-European. Ls. basis is closer to the monosyllabic stage. The basis of the latter is Scythie to a large extent. The other foreign Asiatic affivities of the Dravirian vocabularies must in general be either of similar origin to the common Sanskri- tic, that is, derived from languages that intruded into India from Irania prior to the Sauskrit era, or they must belong to the pre- Indian era of the Draviro-Australian formation, and have accom- panied it in its first advance across the Indus. This does not exclude the derivation of a certain portion fiom visitors by sea, and from any alien northern and eastern tribes that may have bordered the Dravirian province before the Tibeto-Ultraindians crossed the mountains. There is no evidence of the existence ot such tribes, or of the Dravirian having been preceded in India by any other formation from which words having extra-Indian affini- ties could have been borrowed. | The affinities of the vocabularies are much more numerous with other foreign languages than with the Tibeto-Uliraindian. They are very various, and those with remote languages—as the Cauca- sian and North Asiatic—are so abundant and direct, that they afiord similar evidence of the long independence and the archuic position of the mother-formation to that which we have found in an examination of the more generic words and particles. From the time that diffusive nations of higher civilization than the original Indo-Australian existed to the west of the Indus, a flow of foreign words into the Indian vocabularies similar to the comparatively recent Arian current, must have been going on, age after age, and millenium afier millenium. Each foreign, mixed or mative tribe that spread such words by its migrations and conquests, would become the cause of further movements and diffusions. The Dravirian terms relating to arts and usag¢s apper- taining to a higher civilisation than that of the Australians, Simangs O4 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. and Andaman islanders, if compared with those of the other ling- unges of the Old World, will probably enable us to ascertain with what races the Indians were most intimately connected prior to the intrusion of the Arians. . So far as I have hitherto been able to carry such a comparison, the result is strongly in favour of a great influence having been exerted on the vocabularies of India during pre-Braliminic ages, by Iranian, Semitic, Caucasian and Seythic nations, or by nations of one or more of these races whose vocabularies had borrowed from those of the other races. It is uot intended to assert that a Semitic or even a Seythie formation prevailed over Irania as far as the Indus, prior to the Indo-Euro- pean. That must depend on other than merely glossarial consi- deration. Whether or not the formation of East Irania remained Drayirian, more or less modified by Scythie influence, until it was displaced by Arian, does not affect the conclusion that, from this province, words of a more westera and northern derivation, were transmitted by its tribes to India, during the great interval between the Australian and the Arian epochs. There is no ground to believe that the Caucasian tribes were ever tliemselves nomadic and diffi- sive—although other tribes of the same family were—or that purely Semitic tribes speaking purely Semitic languages were ever durably established as far to the eastward as the Indus. The more important modifications which the Dravirian formation has undergone since the Australian cra are not of a Caueasian or Semitie character, but of a Scythie and Scythieo-Iranian. Whatever changes the vocabularies of castern Irania underwent, and however much its tribes were modified physically and in civilisation, the linguistic basis would appear to have remained faithful to the Scythico-Dravirian type. The probability jherefore is that the Dravirian vocabularies derived those Western and Asiatic terms of art and civilisation, which are posterior in origin to the Australian era, mainly from Seythic, Scythico-Iranian and Iranian tribes, that successively dominated in the basin of the Indus. This is far from excluding Semitic influence, direct or transmitted, for most of the eastern branches of the Travian race, particularly the tribes near the Indus as the Aftghuns and Beluchis, are physically highly Semitoitl. The first class of N. W. vocabularies after the Sanskritic, with FTHNOLOGY OF THE INDO-PACIFIC ISLANDS. 95 which the Dravirian fall to be compared, are the remaining Indo- European, and particularly the various Medo-Persian. Hitherto the clossarial study of the Indo-European family has been chiefly direct- ed to the vocables and rootscommon to Sauskrit with the other lan- guages of the formation, so that materials are not yet prepared foo an ethnic comparison of the In/o-Earopean roots in the mass with those of other formations. As necessarily happens in an ancient, very widely extended, and much divided family, the roots of any one language, such as the extreme eastern—Sanskrit—form but a small portion of the variety now possessed by the family asa whole. Besides the more modern acquisitions of each vocabula- ry, there can be no donbt that, as a general rule admitting of exceptions, each large group received most of its peculiar roots from the prior languages of the province in which it prevails, or of those provinces through which the tribes which established it advanced from the original Indo-European seat to the lands where they were found at the dawn of history, and that the radical differ- euces in the glossaries are, in great measure, to be so accounted for. Thus while the Arians, moving eastward into the Dravirian province, would have their vocabularies more or less Dravirian- ised, the ancient Medo-Persian tribes moving on the Caucasian and Semitic provinees, would have their vocabularies affected by those of the native tribes amongst whom they penetrated, Those hordes which passed through the variable Seythic region or continued to oceupy portious of it, would, in many cases, receive fresh acces- sions of Scythic words. Those which moved north westward would probably reccive Fino-Ugrian accessions, while those which went westward through Asia Minor would, for a time, be subject- ed to influences similar to those which have for a longer period operated on the Medo-Persian, In Europe the pioneer migratory tribes must have come in contact not only with Scythie in the north, but with Euskarian, and probably other Scythico-Libyan languages, in the south. Hence probably it is that the glossarial divergency of the Celtic, the Skipetarian, the Russian, the Arme- nian and the Sanskrit, is greater than that which divides many languages of entirely distinct formations, The ethnology of S. W. Asia cannot be well understood until the vocabularies of all the races who oceupy it have been careful- ly compared. A comprehensive comparison of this kind must N 96 ETHNOLOGY OF THE 1INDO-PACIFIC ISLANDS, precede the attempt to trace the history of any one of these forma- tions, and no satisfactory progress can be made in the elucidation of the archaic position and movements of the Dravirian until more light is thrown on those of the Indo-European ard Seythic in particular, The Dravirian vocabularies have some special affinities with the most eastern of the Medo-Persian, those of Afghanistan and Belu- chistan, This part of Irania has received new Scythic vocables subsequent to the Arian era, and some of these may have been archaically common to Scythic and Dravirian. But the special affinities in question must be remnants of the pre-Arian era, and thus stand on a similar footing with the Drayirian roots in the Sanskritoid languages of northern India, These affinities are not confined to Brabui. I have observed several in the Pashtu and other published vocabularies which I have partially examined. These vocabularies also have this in common with Dravirian, that they possess non-Sauskrit roots and forms of roots having clear affinities with Semitic, Caucasian and Sceythic radicals. While some of the Medo-Persian affinities are exclusively with the proper Dravirian vocabularies, a much larger number include also the Guzarati- Bengali class.* These non-Sanskritic roots, and non-Sanskritic varieties of roots that are Sanskrit, common to yocabularies on the western side of the Indus with the ancient Dravirian glossaries, afford some evidence of a period when Eastern Irania was not yet Arianised, and of a connection which then existed between its languages and those of India. It does not necessarily follow that the immediately pre-Arian formation of this province was Dravirian, for even if it was not, it might have had a glossarial connection with it. But as no traces have been remarked of a distinct formation, and as several of the vocables are Dravirian in structure as well as in root, the presumption is that the aflinity indicates the former prevalence of the Dravirian formation to the west of the Indus, and this presumption becomes certainty when the affinities of Dra- virian with still more western languages are considered. It is quite possible that before the Sanskrit language itself was carried * Examples may be found in the annexed yocabulary under the terms Air, Ant, Arrow, Bird, Blood, Boat, Bone?, Buffuloe, Cat, Dog, Ear, Eye, Fire?, Foot, Horse, Stone Ke. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 97 into India, other Arian dialects or Draviro-Arian dialects may have existed in the province, and the Scythic element cannot be excluded from the East Iranian languages of any period. But however this may be, we are ultimately carried back to a Dra- virian era in the linguistic history of eastern Irania, and it is the oldest that we can recognize. By far the most numerous glossarial affinities of the Dravirian languages are with a great chain of vocabularies that appears at one period to haye extended trom the Caucasus to Kamschatka, em- bracing different formations, although it is probable that this wide dispersion of the same roots was chiefly the work ofa race to which one only of these formations was native. The affinities in question embrace Caucasian, Ugrian and Ugro-Tatar, Yeniseian, and, in a less degree, Koriak and other extreme N. E. Asian voca~ bularies. The Ugrian are the most important, but a considerable proportion are exclusively Caucasian, and a smaller proportion exclusively Yeniseian. The larger portion of these roots appears to belong to the pre-Indian era of the Draviro-Australian forma- tion, and to form an integral part of its glossarial basis. The Caucasian basis is Yeniseian, N. E. Asian and proto-Scythie more than proper Scythic, and the Semitico-Libyan formation is not remotely allied to it by several phonetic and ideologic traits, as well as by roots. The Draviro-Australian formation partly enters into the same circle by some of its ideologic traits, and as the Se- mitico-Libyan type preceded the proper Seythic in the 8, W. pro- vince of the Old World, and Draviro-Australian is the earliest of the more Scythoid formations in this part of the continent, it is proba- ble that some of the Caucasian affinities are direct. The more fandamental Ugrian roots, with the Yeniseian and N. E. Asian, render it probable that they were brought by the primary Dravi-— rian—speaking tribes from central Asia, Some are doubtless of later derivation, but the greater portion must be considered as of equal antiquity with those phonetic and ideologie characters which affiliate Draviro-Australian and Ugrian. The more remote N} E. Asian affinities, when not Scythie also, may be still older, for similar affinities are found in the Caucasian, Semitico-Libyan and Zimbian languages. They may appertain to the non-Seythic southern element of the formation, or to its partially cognate proto- - 98 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Seythic or pre-Scythie condition, when it was still located in caster or central Asia. In the latter case they would rank with the pro- nouns. While the roots are largely allied to the Scythic, and especially to the more western and central languages—Ugrian, Samoiede—they are not in general derivatives from Scythic. The structure of the vocables is proper to the Draviro-Australian system, and the forms of the roots are frequeatly euch as are found in other ancient families of language. In general they are to be considered as equally archaic in the Dravirian and Ugrian families, and most of them appear to have been current in the monosyllabic condition of the Dravirian mother tongue. Many vocables proper toa considerably advanced civilization are of this class, and we must conclude that the Dravirian-speaking race which advanced into Irania from the north east and spread over India, was one of the oldest civilised peoples of Asia, and that this family of lan- guage was probably the first of the proper Scythico-Iranian stock to become dominant on the shores of the Indian Ocean. To the same great movement from the interiour of the continent on the south western lands a portion of the Seythico-Dravirian vocables found so abundantly in the Caucasian and Semitico-African languages is probably to be ascribed. The Asonesian vocabularies contain numerous Scythie, and especially Ugro-Samoidic, roots and varieties of roots that are not now Dravirian, but a large pro- portion of these appear to belong to the Draviro-Australian era and to be referable to the same long continued movement. In the next chapter we shall find that it also affected the Tibetan langua- ges, western and eastern,—throngh them, in later ages, the Ultrain- dian and Gangetic,—and, still later, the Malayu-Polynesian through the Ultraindo-Gangetic. Hence of two Ugrian forms of the same root found in Asoncsian vocabularies, it becomes possible to trace one to the primary Draviro-Australian immigrations, an! the other to the Gangetico-Ultraindian that immediately aghaa the Arian era of India. The preceding inferences will be best illustrated by taking a few terms from different classes and examining the affinities of the prevalent Dravirian roots. ETHNOLOGY OF THE I[NDO-PACIFIC ISLANDS. OO Names of parts of the body. Tn comparing the names of the different parts of the body in any croup of languages, we find that the same root has received yvarions applications. The same vocable in different languages or dialects signifies Head, Hair, Skull, Face, Cheek and Eye. We fiud also that the same vocable bas been applied to the more pro- minent portions of the person, as the Nose, Lips, Mouth, Teeth, Ears, Arms, Hands and Fingers, Legs, Feet and Toes. We find also that the same word has been applied to the Head and to por- tions of it, as the Nose, Mouth, &c. and hence the former class of names runs into the latter, and the same term has come to signify every one of the objects we have named. We also find more limited classes, founded on more specific analogies. Thus words for the Lips, Mouth, Tongue and Teeth are often specially connected. It is probable that one name originally signified the Mouth'and all its parts, and that this name afterwards became restricted to the Mouth in one dialect, to the Lips in a second, to the Teeth ina third and to the Tongue in a fourth. The Lips and the rows of Teeth might receive the same radical name. The number and regu- lar arrangement of the Teeth appears to have early suggested a con- nection between them and the Fingers and Toes. Hence the same root has been applied to the ‘Teeth (and secondarily to the Mouth and Lips), to the Fingers (and secondarily to the Hands and Arms) and to the Toes (and secondarily to the Feet and Legs). Similar specific resemblances,—as between the lateral and double appen- dages of Ears, Hands and Feet, and the most close of all that between the two Arms, Hands and Fingers and the two Legs, Feet and Toes,—have given rise to specific glossarial applications. The Eye being the most striking and important feature in the Head the same root was transferred from Eye to Head and vice versa, Glossarial change and concretion bas been attended in all families by similar phenomena, Every vocablein the progress of a nation and ofits language receives several secondary or conyen- tional applications, some larger and some narrower than the original or etymological one, and some only connected with it metaphorical- ly. Henee asingle root, whatever its original meaning, comes to be applied to numerous analogous objects. Distinetions are primarily indicated. by the addition of segregative and qualitive words or by 100 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, double words. But when an object has acquired ideal individuali- ty its name tends to become a specific conveuiional one. Itg etymological meaning becomes first obscure and then disappears. Hence any change in the name, however induced, has the effect of giving it a conventional individuality, and of obliterating the sense of the ancient connection with other applications of the same root. For example, if the Hair, Pace, Mouth and Nose were originally designated by the single root for Head, the distinctions being indicated by variations in the accessory definitives or queli- tives, these compounds might concrete into words the connection of which was no longer felt, and in the gradual phonetic change to which vocables are liable the root itself'in each of its applications might acquire peculiarities of form. By the dropping of the defi- nitives or qualitives the bare root might come to be used as a distinct word in each of its acquired forms. The most common cause of the limitation ofa root or of particular forms of a root to one of several meanings, or to one part of the object it originally described, has been the acquisition of distinct names for some of the other meanings or parts, cither in the internal progress of the dialect or from the influence of another dialect or language. The separation of languages into several dialects has been th: chief cause of the multiplication of specific conventional applica- tions of the same root, and the matual influence of these dialects appears to have been the principal reason why we find ina single vocabulary the same root current in different forms and with a different meaning for each. The glossarial variation and complexity are greatly heightened by the circumstance of the same object having often received several names. The progress of language would tend to give to each a limited conventional application, One ofthe words for Head might be restricted to Head, a second to Hair, a third to Skull, a fourth to Face, a fifth to Eye &c. Each of the dialects formed before this change began might appropriate a different name to several of these objects, so that the same word might signify Head in one, Skull ina second, Face in a third and so on. As the roots would frequently undergo phonetic changes, subsequent intercourse between the tribes speaking these dialects might lead to a yocabu lary which had retained one of the ancient roots with a particular ETINOLOGY OF THE INDO-PACIFIC ISLANDS. 102 meaning, receiving the same rool in other forms and with different meanings from the cognate vocabularies. Another cause of the same root being found with different applications is that some words were used generically. Tlius when the Bars and the Hands were designated by the same root the former might have a word for Head annexed to it. The latter might in time acquire a pho- netic form that distinguished this application from the primary one, and the dropping of the other term of the compound might leave the same ‘root current for both Head and Ear, the phonetic varia- tion being sufficient to give to each a complete conventional inde- pendence. The acquisition of new roots from foreign vocabularies and of new forms of native roots from sister dialects not only leads to restrictions in the applications of the old roots but to secondary and even metaphorical meanings becoming the principal ones. Thos a word that at one time means Hair generally becomes res- tricted to the hair of the head or particular parts of it, of the face or parts of it, as Whiskers, Mustaches &e., or to that of the rest of the person, or distinctively to that of the lower animals, or to one kind of hair, as down, wool, bristles, &c. A word that was originally applied to feathers, hair, grass and other things of a similar growth or appearance may be appropriated to, one of them, so that in different dialects and languages the same root may signify Fea- thers, Hair, Wool, Fur, Down, Moss, Grass, Bur, Bambus, Thorns, Teeth &c. It may also be applied in different languages quali- tively to distinguish particular things or animals, and thus even- tually furnish many substantive names, most of which will in time come to be purely conventional. A word for Face or Eye may become Appearanc2,.Look, See, Glance, Watch, Beware, Guard, Show, View &c., while the brightness, sharpness or roundness of the Eye, a convex in.a concave, may give rise to numerous meta- phorical applications any of which may acquire the rank of a sub- stantive terin when the word becomes obsolete in its original meaning. Thus the point or the edge of an instrument, the sun, a gem, a concavity, a hole, a ring, a convexity, an ankle, a kunckle, the navel, a nipple, a bud, a knot, a spring, the source or origin &e may in different languages be designated by a root which origi- nally meant the Eye. We can thus sce thata very few primary 103 ETHNOLOGY OF TIE INDO-PACIFIC ISLANDS. sounds may have served not only.as the phonetic but as the glor- sarial foundation and material of all language. A root for Head may have not only become restricted to parts of the head, thence to other paris of the body and thence to objects and ideas named from an actual or fancied resemblance to any of these parts, Lut may have been transferred to Scull, to Bone, to particular bones, to things round, cavernous, hard, protuberant, to the top of anything or of particular things, to masters, chiefs, governors rulers &e. No roots have been more prolific, and there is hardly any limit to their ramifications. A single instance will suffice to show how roots that have become obsolete or been displaced in their primary or older meanings are preserved in their secondary ones. In Malay Head is Aapala, a comparatively recent acquisi- tion from Sanskrit. But the native or earlier root, wl, still eur- rent in many of the cognate Indonesian languages, is found in Malay with several meanings. Unaccompanied by any other word, it was signifies “inland” and “interior.” The history of this word isclear. From the Head, it was applied to the highest part of a stream, and as streams are the Malay highways to the interior and the cultivated and inhabited tracts are in general limited to their borders, the ww sungy, the head or upper part of the streamy was synonimous with the interior of the country or district. When ulu was by degrees supplanted by kapala in its principal meaning, it came to signify the interior even without the addition of the distinctive word for stream. Another of its secondary applications is to the hilt or handle of a weapon or in- strument, the blade being termed the eye, mata, a root which, in other languages, also signifies Eye, Face, Head &e. What is found on comparing the vocabularies of any single family, is found also on comparing those of all the known families of language. The same terms recur in them and it soon becomes evident that in their primary roots and vocables, they are all inti- mately related, and are in fact ultimately dialects of one language. The glossarial resemblance is so close and unequivoeal, and the transfer of roots from one part of the body to another is. so uni- versal a phenomenon, that we arrive at the conclusion that this fundamental portion of the vocabulary was formed, to a greater or less extent, when the different families of language had not sepa- rated far from each other, This remarkable connection has ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 163 doubtless been brought about in particular cases through the mutual influenee of vocabularies that have been brought in contact by ethnic movements, although originally widely separated from each other, But the connection is too intimate and too universal to admit of such an explanation asa general one. It is more pro- bable that the comparatively barbarous and outlying tribes of the world, as the Hottentots and the Australians, carried their cog- nate Asiatic basis vocabulary from a primitive seat in the vicinity of the parent Asiatic tribes to their present locations, than that it was brought to them there by alien tribes that spread from an Asiatic centre to the extremities of Africa and Asonesia after these were inhabited. We may indeed imagine a succession of such all-embracing movements, but the source of the common vocables must ultimately be found in one centre, and there is a considerable and fundamantal class which appears to be equally archaic in all the families and must be referred to the earliest ethnic movements. Whether there were originally one or several languages, it is evident that the mother tongues of all that are now preserved existed ‘at one period as closely connected and mutually influenced dialects, and this condition of things could only have arisen from the tribes who spoke them occupying a very circumscribed portion of the habitable world. We can clearly trace the influence of several dominant and widely diffused voca- bularies, but after allowing for the common vocables thus dissemi- nated in various directions, there is a large residuum of identical roots, forms of roots, duplicated and compound roots, and com- pounds» of definitives and roots; the presence of which in all the outlying languages of the Old World can only be explained by each haying inherited them as a portion of the primary yocabularly which its mother tongue brought from some ethnically central region. It does not seem possible to go beyond this conclusion. Whe- ther the earliest central languages were of independent or of common origin cannot be determined, because while proximity and mutual contact would’result in an inerchange and community of roots between originally different languages, a single language when isolated would separate into different dialects which would ultimately vary as much in their applications of the common 0 104 ETHNOLOGY OF THE INDO-PAGIFIC ISLANDS. roots, a8 an alliance of assimilated vocabularies, In dialects of common descent the proportion of words that preserve their identity in root and meaning gradually decreases, while the pro- portion of those roots that have acquired a peculiar conventional meaning gradually increases. But in the life of languages a root that has wholy lost its primary signifieation and gained a different one, 18 equivalent to a new word. Hence in cognate dialects that are separated, alien vocabularies are constantly growing up, and they may at last come to be as distinct from each other as it is possible for human tongues to be, So that whether speech began with one language or with many, the kind and degrce of diver- gency and resemblance between all the vocabularies of the world would, in the lapse of time, be th: same. It is probable that all existing vocabularies are etymologically identical, and even that they have all been woven fiom a few primitive roots designating the most familiar objects, qualities and sensations, but it is trae at the same time that the identity of the roots with few exceptions is not a living one even inthe same language. It is on the capacity of the same root to receive almost endless changes in meaning and form, and thus to become in reality the progenitor of a sne- cession of new generations of roots, that the growth of language depends. It hence becomes possible for the human mind and tongue to create a language from a few primary cries. These sounds, partly exclamatory and partly imitative, gradually undergo infinite variation and composition, and each modification becomes a new substantive sound or root, in the linguistic pisgrice of the family, the tribe and the circle of tribe. The following are illustrations from Semitico-Libyan. Tho sibilant, varying to the dental, is used for Head in several Zim- bian languages hitoa, hizoa, mutua, mtu &e. and in Fanti ityil (pl. ityie) ; for Hye in the same family with a different pref. disu, Zisu, itu &c. and in Berber thith; for Face in Kosah with a third pref. wbuso; for Hair in simple or duplicated forms and with the labial final in Gara shof, Mahrah shob, Saumali temo, Bishari tamo, Agau sifa, sisifa, tsabka, tsebega ; for Beard with similar forms in Zim- bian dzevu, devu, debu &c.; for Hair without the labial postf. in several East Zimbian vocabularies misisi, matiti &c. ; for Mouth with the labial final in Arabie thum, Hottentot tub, Felup datum, ETHNOLOGY OF THE 1INDO-PACITIO ISLANDS, 105 Shangalla suma, and without the final in Mandingo du; for Tongue in Bishari medabo and Hottentot tama; for Lip in Sera- koli shume. The same series is found in words for Finger isha, asabi &c. Semitic, tyaba Fanti, sat Amharic, tsat Agau, and Foot tsab, chafu, chami, chapi &e. Gafat, Gonga, Agau. The simple and duplicated root is also Far ti, tu, ta, du, &c., Hand id, ad, tot, tata &c. and Dvot ti, se, sa &e. In the corresponding Caucasian series we find for Head dudi, ti, tchum, sab; for Lye te ; for Hair toma; for Mouth suma, sumun ; for Tongue sibi, zahbi; for Pinger titi; for Hand tota; and for Foot shepe, zhape, shape. In the Scythie series we have besides the simple root the dupli- eated forms shosha Ugrian Fuce ; usu Mong., sus, ses, shosh &c. Turkish ZZair ; shus, tos Ugrian Mouth; tish &e. Turkish Tooth ; and forms with a labial final soma Hung., shem Ugrian Head; sham, shem &e. Ugr., sima, saiwa, &c. Sam. Hye; asim Turk. Hair’ shum Fin Mouth; tipe, Sam., tiwn Ost. Zooth; udam, oda, hute &e. Sam., te Jap. Hand. The Indo-European series has stoma Greck Mouth; suban Pers., shiba Afgh. Tongue; sub Selav. Tooth. The sibilant or dental with a liquid final is Zboth in Semitic sin &e,; Lar in Semitic zin, zan, zun; Lip in Fulah, Sereres tony godon ; and Har in Darf. telo and Mandingo tulu; Hand in Mala- gasy tanana; and Foot in QGalla tana, Woloff tank, Saumali adin and Bagnon guidine. Caucasian has sir JZead, taalo Hand, tul-we Finger, tle Foot, sila, zul-we &e. Tooth, tzindi Nose. Indo-European has for Head sir, Tooth zan, dant &c., Tongue zange &C. Seythic has for ZZead, dil, dul &e., Zooth til, del, Face syn, syrai, mura &e., Eye sin, sil &e.; Lar shan, shun Tung., Hand aol, tol» Finger tul, tyl, dal &e Dravirian has for Head senni, tale &c., for Zand tol, for Foot adi, orri, adu-gu. In the liquid series we find in Semitico-African for Head alo, our, or, eri, ru; for Kye ain, aire, il, iri, &e.; for Mair alu, iru, riri, li, &e. ; for Mouth lah, nua, enus for Tooth reir, hauri; for Tongue arah, for Mar ilai, iroi, ru, noa,; for Finger ala, nun &e. ; for Hand nan, nen; for Foot noa. 106 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. These simple and+ double forms correspond with the Caucasian na, la, ala Face; ena, nina, nin Zbague; ain, in, lai, lar &c. Har ; and rori Fvot ;—with the Indo-European rin Nose; ohr, ur ar ; —with the Ugrian ol, er, olo, ulu, ruh ead ; nore, nanu Face ; nun, lele, ilet &. Hye; lelu, ein &e. Beard; ul, lul, an, nal &e. Mouth; urul Lip ; orr, nyr, any &e. Nose; illa Lar ; al, ol, ola, ili, nala, &c. JZand; lyl, lal, ora, ngoi, hga &c. Foot. The liquid with a labial final is found in Darfurforyenume; in Zimbian for Mouth lumu, romo; in Galla and Kosah for Lip luf- luf, lebi; in Malagasy for Tooth nify ; for Tongue in Danakil aru- ba, Saumali arub, Galla arubni; Woloff lamin and Bagnon kaleb ; for J2ye in Bagnon gwinif, Sereres nof, Woloff nop, Serakoli ai-ndofo, Hott. #” naum. The corresponding series is almost absent in Caucasian, Indo- European and Scythic which prefer liquid finals for liquid roots. Caucasian has nap, napa Face, nem Tongue, lemba, limha, lumbha Ear. Secythic has wa-nim, ny-rim Face ; namo Mouth. The liquid with a sibilant final is ead in Semitic eresh, ras, rosh, rus; Zongue in Semitic lashim, lishin, lisan, halishi Hausa, melas Amh. Tigre, arat Galla. *, Similar forms occur in Caucasian for Hair, ras ;—in Indo-Euro- pean for Mouth, rot, ert ; and Nose nas, noss, ris;—and in Seythic for Head yvesz, arsem, nash; Face rosa, rozha, ortza ; / air yorsi, ersi, nosu &c. ; Hye anysha, elisa, ilet. In Drayirian the principal series are the labial and guttural. The labial forms for Head mudd, mande, are North Indian, mud, mun, Seythie and African, mudah Saumali, mata Saum. Galla. Those for “air mir, mayir, are Caucasian. Those for Mouth vayi, bayi &c. are Scythic, but with the meanings Head paya, fei &e. Face pai, Tvoth pai. Those for Tooth palla, palluare Ugrian, pane &e., but the same form is common in Ugrian and Oaueasian with other meanings, Ear &c. Those for Nose, muku, have tho Seythie root pu, bu, but in Scythic the guttural final is absent. Those for Finger veral, birlu, are Caucasian, palit (the slender form yer, bir, being common with other meanings in Cauc.) ; Indo- European pers¢ Sclay.; Scythie parne, borne Ugr., bar-mek Turk, (the slender form pil, bel &c. being used for Foot, Bar &e) ; and Semitico-African pirure Suah., baram Woloff, fara¢schi Hausa. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 107 (The forms bir, bar, par oceur as the root in words for Hand.) The Kol terms for Mead and Hair bu, ub &c. are Seythic (pa, bui &c. Fin. Ugr. Head; up, ob &c. Ugr. Sam. Hair); and Libyan ap Eg., amo Saum. Head, emu Avekwom, wnbo Mudjnna Hair, In the guttural series ku Head is Seythic,—og Ugr., oike Fin. ; kuzha, kuda Hair is Caucasian; kan Hye is Chinese aud Turkish ; kuli Tooth is the Seythic kul, gol, kur, &c. &c, Head, Beard (kul- ye Yenis.), mouth (kur. go Ugr.) Ear (kul, gul, kor &c.), Tongue; Hand,—in the slender form ker (also Scythic) it is Tooth in Cau- easian ; kadu, kivi, kimi &c. Ear is Scythic, ku, kwna, kyrwa &e. ; and Indo-European ugo Sclav.; kai, kayi Handis Scythic, kal, kasi, kest &c., Cauc. kuer &c. and Libyan eka, kuna &e; kazh, kal Foot (the same root) is Seythic, kasa Yenis., kul Mong &e., Caucasian kash, kassi, &c. and Libyan kula, gar &e.. | In Dravirian the liquid series is only represented by the Tamil eyiru Tooth (yir, yor, Head Ugr., yir-si Hair Ugr., yul Mouth Upgr., yel-uth Har, Kamsch.); the Teluguyelu Finger; the Telugu noru Mouth (« common Scythie form, nal Mouth Sam., onnor Tongue Yukahiri, nol Nose Ugr., nore Pace Ugr. &c) ; the Kurgi orama and Gond robong Mair (which resemble the African forms with a labial final); and the common term for Tongue naku. The root na is used for Tongue both in Seythic and Caucasian languages, but not with the guttural postf. Similar forms occur with other applications, nago Ugr., nyako Fin Face, nuyak Chukehi Hair, enku Koriak, onyok-fo, nig-sha Tungus. Nose; lege, hanka, andika Andi Ear (comp. the Telugu nadike). Semitico-African has allok Tongue Felup, uluk Zar Kensy, uilge Zar Tumali, Koldagi, iluk Zooth Saumali, Galla.* a The series, as a whole, is Scythic, with a few spécial affinities to Cancasian. Haig nis trl cont eter eh tal et access to. vad katumb; if ka be prefixual (comp. kalakh cheek) it is Georgian tchum head, toma hair. Hair pish-kou ; besh Misjejian. Beard rish ; ras ye Lesgian, Finds &e. Ugr. Eye khan; iret, mice jen i reir oe poet sie Kurgi. Lip ba; Japanese fa. Nose ba-mus 5 Drav a » Japa A phen trae gian mushush. Tongue duvi; duvo Bishari ; tub, Caos du, eout ieee ike ‘African, daveda cheek ‘Velugu. Bar, khaff; Drav. (kavi Toda &e.). Hand, du ( , some root ag in duvé fongue); tota Misjejian, tot Coptic, uda Samoiede, Foot wath ; “hi ’ i 1 the i Finger. Shis vocabulary appears to be equally archaic with t Danan Hh Ansteallan,; to have the same primitive relation to the Scythic, and to have some specific Caucasian and Caucaso-Libyan affinities. 108 ETHNOLOGY OF THE [NDO-PACIFIC ISLANDS. MEAD, HAIR. The terms for head and hair being much interchanged in general glossology I pluce them together. Eleven vocables are found in the Dravirian languages,—senni, tali, mudi, mir or mayir, kuzh or kud, orama, chuti, ventruka, ku, buho and ub, Of these, two are of Tibeto-Ultraindian origin. They are con- fined to the languages bordering on the Ganges. All the otherg are archaic, and all have Seythie, chiefly Ugrian, affinities. The chain of affinity is various,—Ugrian, Iranian, North Indian, Aus- tralian and Asonesian; Ugrian, Caucasian; Yeniscian; Ugrian ; Ugrian, Sindhi, Tatar, (Malagasy, Asonesian) ; Caucasian (Ko- riak, Sanskrit) &e, Ku, “head”, is found only in Uraon and Male, where it appears to be of comparatively modern Ultraindian derivation, (Naga). The root is Tibetan, Ugrian, &e. The Asonesian varicties, like the N. Dravirian, are Tibetan through Ultraindian. The Kol termbuho, bu, “head”, may also beof Tibeto-Ultraindian derivation, but as another variety of the root, used for “ hair ’’, is archaic and as the aspirate of buho appears to have.been transmitted as a guttural to Asonesia, where the term is very common, there is some doubt as to this. Probably the form buko or buho was an early Draviro-Gangetic variety of the Tibeto-Ultraindian pu, which was carried to the islands by the Gangetico-Polynesian current, All the other terms are archaic. Mudi “head” is Hindi and Australian on the one side, and Seythic on the other. The full form, slightly modified, remains in the nasalised Gangetie pur- ing (Dhimal). Other varieties are common to Australian, Papua- nesian and Malayu-Polynesian languages with Irano-lndian and Ugrian. The root must have prevailed in all these forms in the Iranian province, not only in its Sanskvitic but in the older Draviro- Australian and still older proto-Scythic or proto-Dravirianeras. The various Asonesian terms show that the root was the most common in the North Dravirian or Gangetic province. In the south similar forms are only found in Toda and Kurgi, and in Malayalam, the last using this vocable for “hair.” The Toda and Kurgi forms ap- pear to be equally archaic with the Australian, which exhibit a si- ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 109 milar striking: identity with Irano-Indian varieties. The Australian wadi, wari, New Caledonian mari, must, like the other Asonesian forms of the root, belong to the most archaic era of Asonesian glo:- sology, yet the form is the same as the Kashmiri wad (Zend wed). The Peel River bura corresponds with the Bengali mur, Dhimal puring, Todava and Hindavi mud, [so mun Hind., umun Lobo of New Guinea], to which the Malagasy-Polynesian vulu, fulu, balu “hair” is also allied. The original is the Seythie mari Korea, murit Tung., mui Turk., bui Ugrian. The forms in a are also Seythie,—Uegrian wari, waras, awa &c. The medial u and final i of the Scythic forms is preserved in the Dhimal puring (Korea muri) only, and the i in the North Australian and New Caledonian. The form mud, mun, mande (Hind., Tod., Kurg.) is also current in Samoiede asa term for “ beard ”, mudut, munneke, mundu. In Tungusian it is applied to the “ mouth’ amun. The forms in t, matha, mata, are allied to the Sanskrit mastaka, which appears to be com posed of two Scythic roots, mas (comp. mas “hair” Armenian, masf, “hair ” Kashmiri, bas, bash, pus &e. “head” Turkish, bus “ hair” Fazoglo, iwusa “ hair” Fin, usu “hair” Tungus.), and takai (“ head” Yeniseian, tuka “ hair” Fin &e.) An allied Tamil and Toda word for hair, mir, mayir, is pro- bably a more archaic form. It is a Caucasian variety, and belongs to a dialect that shows frequent affinities with Dravirian, the Andi. In the Seythic languages the form war is found for “ head” and “hair”. The slender forms bir, pil, wil, pin are current for “lip,” tooth”, “ear” and “foot.” In Caucasian ber, were, occurs for “ face,” “eye,” “ beard,” pil, piri for “ mouth,” and mir, mer, mar for “ nose”. The preceding term is so common along the whole glossarial band of Irania, North India and Asonesia, that it appears to be connected with the Seythic or proto-Scythic movements which gave their special Scythic character to the Draviro-Australian and Indo-European formations. The Tamil and Toda variety probably marks an older Caucaso-Dravirian current, of which this remnant was left in the south of the Indian Peninsula. The Tamil term for “ head,” senni, is probably of equal anti- quity. Itis Ugrian, Celtic, Yeniseian, &c. Another variety of the 110 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. same root is Scythic, Indo-European, Semitico-Libyan, ke. The same form is used for “ face” in Turkish syn, “eye” in Ugrian sin, “ mouth” in Fin sun, and “ ear” in Tungus, shen shun. The most common Dravirian term for “ head’’, tale—found in Male as the word for “ hair’ —is also archaic. Itis Tatar. Some rare examples also occur in Asonesia. Besides the Tungusian and Mongolian forms for “ head’’, similar forms oecur in Seythic with other meanings. In Turkish it signifies “ tooth” (til, tel, del &e), and itis an archaic and widely spread term for “ hand” udol, ton, and “ finger”, tul, tol, dal &c. It is found in Drayirian also as a term for “hand” tol and in Caucasian as a term for “ finger’, “hand”, “foot”, and “eye” tul-me, tle, taalo, toli &e. A Tamil (anc.), Karnataka (ane.) and Tulaya word for “hair” kuzh, kud, is Sanskrit, Caucasian and Koriak. The Dravirian furms resemble the Caucasian most closely. In Seythic it is ap- plied to the “eye” kus &e., “mouth” agus, kuzi, “nose” kase, &c. *“‘ ear” kus. . Another archaic and comparatively rare term for “ hair”, ram, lom, rob, ran, lang (with different augments) is common to Kurgi, Gond, Bengali, Roti and Wiradurei. The ultimate root la, ra &c. has numerous affinities. The Kol ub, up, “hair” appears to be also archaic. It is Ugrian, and an allied form is found in Egyptian. All these forms are ultimately only varieties of the root bu, pu &e. already noticed. The Uraon chuti is Sindhi. The root is Ugrian. Head (a.) senni Tamil ane. sheny Ugr. (Wolg.) shem * | ar-sem » (Ost.) soma », (Hung.) ser Ossetic, Pashtu sar Pashtu, Hindi shira Sansk. ka-sira Japan sheier Ugrian shen Celtic (Erse) (hair) shaar Arabic ~ ” ETNNOLOGY OF THE INDO-PACIFIC ISLANDA. sn shnin song chang sha, ta unsn asim chinyajan chunajan shim Egyptian Kasia Newar Sunwar Tibet Mongol Turkish Yeneseian Tobi 111 The sibilant and aspirate root, in these and various other forms— including the reduplicated sis, sus &c—is very common both for ‘thead” and “hair” in all the principal formations, Chinese, Scy- thic, Tibeto-Ultraindian, Iranian, and Semitico-Libyan. The Tamil term appears to be archaic. It resembles the Yeniseian chin, Chinese shin, and the vowel connects it more immediately with the Ugrian sem, and the Ossetic and Pashtu ser. (hair) talei tala tale tare tali mi talu dil dul del deli tolo-chai tari-gun tul-gai thilu (hair) tulah The closest to the Dravirian of the Scythic terms is the Mongol tari. The Indonesian tulah, Rotuma thilu, appear to be Tungu- sian. The root is probably the still more widely spread ta, tha, tu. ® The same root is moré frequently applied to Toxovs in Seythic (til, dil, tel tul, dela.) Head (b.) Tamil Mal., Telug., Gond Karn. Tuluy. Male Magar Tungusian*® P| aD i} Mongol ; ” a) Rotuma Meri r 112 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. mudd (hair) tala-mudi (hair) a3 4 Ph mande mud mun mur wad wedege bala bal : war wal wed madam puring pala (hair) mala (hair) mbal, angbal, ji-mara a? (hair) wuran ~ ”) ” 7 _barram wil umun balang walang wari iwadi di-vara mari balu bau bail bura ae mori morye kabara kapan Head (c.) Todava Malayalam Kurgi Hind fF Bengali Kashmiri Zend Sanskrit Hind. Sindhi Kashm. 3 Bodo Dhimal Sirawi Viti Mille “Aru Lobo Australian (Bathurst) oh (Kamilarai) ” (Trusan) » (Limb. Kar.) N. Aust. Sydney New Caledonia ) xn Toro, Vate . Malikolo Peel Riv. Kamilarai - Wiradurei Bathurst Endeavour Riy. Sydney, Liverpool Muruya ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 311 [kapala Sanskrit (Indonesian)* kobbel '. Kamschatkan] uwari Ugrian, (Sam.) awa ” ” pa = (Fin) pra i (Wolg.) pank » _(Wog.) panga ” penke 99 _ pen Eusk., Celtic (hair) waras Ugrian (Ostiak) » muri Korea merit Tongus. bui Ugr. (hair) mui Turk. emu Avekwom muru Makua (hair) vulu Malagasy » fulu Pol. » bulu Indon. (com.) matha Bengali matho Sindhi mata Galla matha _ Saumali muda — ow mati Korea bash, pash, pus &e¢ Turkish pisha Tiberkad bacha Vanikoro basa-ine Malikolo ® In the Australian kabara, kapan, the ka is probably a prefixed definitive and the root bara, pan. In the Sanskrit (and derivative Indonesian) kapala the root may be the adie bY de rp he which appears with a different postfix in cap-ut, hof-d, kuw-ud b-it ‘he Latin eapillus, villus, pilus ** hair” favours the idea that pala is a root inkapala also. The true analysis may be kap~ a, kap-pillus. The labio-liquid root occurs in Selavonic also wel-is-ok, wlas, Bias &e * fair.” Kop, kap appears itself to be one of the Ugrian torms of the ‘ttural root (e). Comp. ugom Ostiak (so coma “hair” Latin), In the Indo- Eoropead branch of the Scythico-Iranian alliance the root is generally applied to Eye (oko, og &c.) and the exceptional Latin oculus appears to preserve the postfix of another Ugrian form ugol, oklu. ; : ¢ Comp. with the meaning ace, facies, face, visage; the Eusk. bisaja, Breton wisaih, Romanic wise &c; in which the same root occurs. 114 ETNNOLOGY OF THE INDO-PACIFIC ISLANDS. pathe-na-nadi . E. Tasmania (forehead) mat Torres St. (Erub.) ’ Head (d.) [See [Hair 9.) ba jg tee ©, ee buho, -boliw a pit; %t.5 5. Mikir bong § Singfu buhu, poko, poho Indonesian, Polynesian The root is very common in Asonesia, particularly as applied to “hair.” The-k of the prevalent Asonesian bok, wok &e, “hair’’, appears to be referable to the aspirate of the Kol form. fTead (e.) kuk Uraon kupe Male go Tibet, Tungus. ta-ko Gyarung, Naga gok-ti Takpa kok Manipuri D. kui “s kho “Karen koi Nancowry, Binua koik Torres St. (Masid.) “seull” guiku | Kowrarega, “ forehead" ku Balignini The root is common in Ultraindian and Asonesian vocabularies. ka Australian (Mudgee) ka Circassian akha rr akai Berber kai Hausa ikhf Berber og Ugrian (Ost. ) uk py. ugol *” ugom » okle » Hair (a.) kuzhal Tam. anc. Tt ie Te 7 ri il ti iA i i i “5 ie sane Sia a a) ee aif A i | ie in fe A : e. By Bes i ‘ = = 7 é aes, ol 3 ee = + ie! a ‘S| ce ee ai thd 7 an a oa ag z st ‘a am | +, wee - 7 | [ (os5. Fes a tad ny ‘eee | De! ve - Hes ‘ a ; Se se gee a ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 119 - kudalu {khosu kosen kacho kitang gitang gi’an kide ketha-na kesha % kassa kazh-eresh chaz kodi ketschugué kiti-gir [ (head) koltsch gashi Karn. mod., Tulu. Karen N. Tank. Naga] Australian (Kam. Syd.) »» (Liverpool) »» (Bathurst) Tasmania 3 Sansk. Lithuanian * Misj. Cire. Lesg. Koriak bP] Kamel] Hausa . The Australian and Tasmanian variety as well the Sanskrit are connected by the vowel with the Koriak. The broad ‘Dravirian form preserves the vowel of the Ugrian, Tibeto-Ultraindian and Ultraindo-Asonesian form of the ultimate root ku, ug &ec, (“head”). It appears to be connected with the Caucasian form, Hair (b.) mir . mayir [tala-mudi wil pelilo-gueni par-ba, par-eata purari [? min mier maar pilus wlas Toda Tam. mod, Mal. ] Aru Tasmanian un 7 Chepang] Lesgian (Andi.) bi Lat. Scl. The Tamil mayir and Lesgian maar, which are evidently related, appear to show that the full form preserved the broad yowel which 4120 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. EYE. The prevalent Dravirian term is kan, (also Brahui) kanu, &c. Jt is found in Polynesian, kano. On the continent itis Chinese gan, Thochu kan. It is evidently one of the primary vocahles of the Dravirian formation, and as the form is a rare one it illustrates the special Chinese connection of the basis vocabulary, as shown by the pronouns. The root is more remotely connected with the Yukahiri angoha, Scythic kus, gus, kas-ak kar-ak, osha, &c, and Indo-European ak, ank, agn, akshi, akis, oculis, okko, oko, augo, auge, oog, eye. The corresponding Scythic kar &e although found as “eye” in Turkish only is a common Scythic root with other meanings. In Fin and Aino it is applied to “ hair,” and the original is proba- bly to be found in the Ugrian and Yeniseian kol, vol. Itisa common term. for “beard” agan Ostiak (the Chinese and Dravi- rian form for “ eye”) gar, gor, gur, ger, Tung. kul, koro Yenis.; “mouth” kur, Ugr., khan, Turk. (the Drav. form for “eye”); “ear” kor, kyr Fin, kolo, kul &c, Yenis.; “hand” kal, kol, gar,; ‘ foot” kol, kur, &c. The Australian mir, mil, corresponds in form with the Dravi- rian mir “ hair, ” and the ultimate source was probably a slender form of the Seythic labio-liquid root for “head.” Scythic no longer retains slender forms with that méaning, but it has them as words for “ lip,” “tooth,” “ear,” and “ foot” while Caucasian has them for “face,” “eye,” “mouth,” “beard,” and “ nose” (mir, mer, ber, bir, pil, wil). The form dala, dana, isnot now current in Asia as a word for “eye” save in Caucasian. It corresponds with the Dravirian tala “head tol “ hand,” with the Scythic tala, dil &c. “head,” and the similar words for “tooth,” “ hand” and “finger,” in Scythic ; for “eye” toli(Mingr. Laz.) “ finger”, “hand” and “foot” in Caucasian; and for “tooth” in Indo-European. The Kol dialects have the full Tibeto-Ultraindian met, med, the vowel being that of the Gyarung, Burman, Simangand Binua forms, and not the common a of mat, mak &e. It should be remarked that this second and very widely spread root is also Chinese (mok, ma’.) Of the two terms prevalent in the proper Dravirian vocabula- ETHNOLOGY OF THE INDO+PACIFIC ISLANDS. 121 ries, the second, which is only found im Tamil (anc.), is probably of Sanskritic derivation, nattam Tam., netram Sansk. But as the root occurs without the vy in N.E. Asian and African languages it may possibly he archaic in Drayirian. It belongs to the proto- Seythic basis of Sanskrit (comp. Caucas. na, nne, la, ala, Kams- chat. lela, eled. elath, nanin, Koriak ilet, lalat, elifa, lilagin, Hind, nain, Arab aayn, Nubian nget, enes-ik, ma-inka, Gallaitsha, Danakil enti, Malagasy-Asones, inty “see,” Malay lihat “see,” Binua nihat “eye.” The N. W. Australian ira, Mudgee lun, TorresSt.ir, il, of ir-kep Erub, il-kap Murray I. (kap being a separate root, dana~ kap “eye” Port Lihou, ii-kab “ temples” Erub) with the Nilotic il, ilan Saum., ila-tua Galla, aire Tigre, ili-kumah Shangallah, are Fazoglo, yel Agan, ¢o-lele Bishari, ile Falasha, are still more faithful to the archaic form preserved in N. E. Asia, and, like a large proportion of the yocables of the Nilotic province and Africa generally on one side and of the Draviro-Asonesian on the other, ave referable to the earlier eras of the Scythie or proto- Scythic Thovements on the southern regions of the Old World. The same root is common as a term for “ head.” The prevalent Australian term mil, mir, mi, me, ma, may be connected with the Tibeto-Ultraindian mik, mit, mid, mni, mi, (Takpa melong), but the Caucasian ber, beer, ber-ik, ber-gish, ber-g, bera-/a, suggests a different line of connection. The Austra- lian mebarai, mibare, mabara appear either to be reduplications or to postfix the same root in one of its other and more preva- lent applications * head,” “ face.’”” Comp. ga, ka, ka-bara, bura, balang, wari, iwadi, &c, “head”, So in Mille baram, New Cale- donia balu &c, New Hebrides bau, bail &c. The same root is very common as a term for “hair” [See Heap, Hair}. In Menado (Celebes) it is used for “eye,” waren, and in Tasmania we find e-verai, nu-bera, nu-bra, le-pena, el-pina, ma-meri-ha, na- murt-k, pola-to-ola. Torres St, has poni “ eye-ball,” and in some compounds “ eye”. In N. EB. Australia and Torres St. dana, daan, dala, dana-kah occur as words for “ eye.’”’ The only other Asonesian example ofa similar term which I have remarked is the Loyalty I. (Lifu) tala- mek * face,” ala-mek “eye,”’ in which mek is the common term or “eye,” “ face,” already referred to, Comp. the Binua tam- 122 ETHNOLOGY OF THE INDO-PACIFIC ISLANDY, langop, Batan dangoy, Kayan inang, Buol lan-ji, Tojo ling-kina, Ende rangia, all words for “ face,” and the: Ne than “ face, ”” Dravirian tala &c. “ head.” EAR. The root, ka, ki, ke, che, se, isso common in different formations that it is difficult to indicate any special affinities. It is evident from its taking the Dravirian postfixes du, da, mi, vi, in different dialects, that the pure root belongs to the native basis glossary. As it is found with other postfixes in Seythic &c, it appears to be proto-Scythic in Dravirian. A variety of the same root is preva- Tent in Australian, kala, kura, kure, kuru &c. This resembles Sanskrit, Hindi, Georgian Seythic and Galla forms. It was pro- bably the North Dravirian or Gangetic form, and later of importa- tion into India than the South Drayirian, the Scythic postfix appearing to be concreted and to have accompanied the vocable in all its wanderings. There is a second archaic Asonesian term, pol Binua, pil Torres St., bina, bena, bidne, &c. Australian, pel-vera-ta, ti-bera-ti Tasm., ere is N. E, Asian, wilugi, wily &c. (Koriak); Ugrian, pel, Pil, pul, bol &c; and Hindi, bol. ' The Kol luturis a rare term. It is probably Sabai and proto~ Scythic,—yeluth, ilyud Kamsch. kadw Tam. mod. . kada Mal. kavi Tod. Gong kivi Kar. kimi = kemi Kurg. kebi Tulut chevi Telug. sevi Tam. anc. ; khetway Male (double postf.) khebda Uraon (double postf. karna Sansk, agantsch Arm. ugn FF ‘kan Hindi &e, kanang Milcli. ETHNOLOGY OF THE INDO-PACIFIC ISLANDSy 123 khana Key Ultraindian. kenei Yengin (New Cal.) kueni, kowan Tasmanian. kalajan Australian. kura - fs kure 5 karusa Masid gerip Erub kowra Port Lihou skor Kasia nakor N. Tangk, machor Garo nakru Mish. kori Georgian guru Galla gura ” ukuna Dalla kulak Turk. kologan Yenis. korwa Ugr. kunya Tungus ku Ugr. ko a kui P ke. ke. Hann, (a). The Tamil and Karnataka tol is Samoiede (utol). It.is also found, slightly varied, in Yeniseian and N. E. Asian yocabu- Jaries. tol utol tolondscha tono ton son taalo Tam. anc., Karn. ane. Samoiede Yukahiri (double postf.) Kamsch. Yeniseian Korea Lesgian . ; (b.) The more common kai, kay? (yi, + being probably the def.) is an archaic variety of a root which, in its Seythic forms, is very 124 YTHNOLOGY OF THE INDO-PACIFIC ISLANDS. widely disseminated (Caucasian, Tranian, Gangetico-Ultraindian, Asonesian.) kai Tam. mod., Tuluv. kayi Karn. mod., Toda kaya "Mal. kaik Gond khekhah Uraon cheyi ; Telug. khai Bodo - akhui N. Tangk kuié es akhu Knmt The root with a consonantal final t, d, |, r, is Seythic, Caucasian franian, Gangetico-Ultraindian, and Asonesian. The Scythic forms are found in all these families. The Dravirian appear to be more archaic. (c.) The Kol thi, Gond the, is found in Kasia ka hti, and Binua thi, ti. The Mon and Anam tai, Ka ‘ei, are probably connected with it, bat they also resem ble the Dyavirian kai. The closest foreign affinity is with the Sem itico-Libyan it Gara, haiyit Mabrah, (whence the Tigre id), Hebrew iath, Berber thith, Ara- bie yad, yodan, The root is found in the Chinese siu, Japan te, Samoiede hute, huite &e. The Male sesu is probably a variety of the same term, BuPit may be from the Telugu, cheyi, a variation of kayi from the com- mon interchange of k, ch and s in Dravirian phonology. ‘The Australian biré/, mara, mana, mangal, mura, tamara, ma- rigal, ‘ma, &e, Torres St. bai, pai, New Guinea mareh, Mille ban, Loyalty I. wana of i-wana-quem, New Caledonia yam-wam (Ba- lad), Malicolo vean of vean-bruas, Celebes pale, Borneo bareng, pinang, Sumatran bungu, pungu, Sambawa ima, Sasak ema, Simang weng are Ultraindian—van Maram, a-pan Champhung, pung Luhappa, mu Lau,—and N. Gangetic,—moa Kiranti, moi Gurung, palara Newar, promji Murmi. The root is found with the same meaning in Koriak minya, minyéln, minagylgen and Latin manus. Terms for “hand,” “ finger,’ “arm,’’ “ foot,” “‘leg,’* interchange to a great extent, and in the Scythic languages the present one is chiefly found as a term for “ finger,” a meaning ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 125 it also has in several of the Gangetico-Asonesian tongues in which it is current for “ hand.” Samoiede «bai, (Torres St.*) mun, mu- non, Japan ébi, Ostiak pane, Wolga parne, Perm pelu, Turk. bar- mark &e., Caucasian palik, German vinger. To connect these with the UWltraindian and Asonesian terms fur “hand” I may instance the North Gangetic brang, prach, brumu, Mikir munso, Australian mura, Tarawa abuni-bai, Ara wawanli, New Guinea amui, all signifying “finger.” Hence also the Malay palit to smear &e. with the finger, and perhaps also pala “to beat.” The various Dravirian vocables for Foot, Mouth, Skin, Tooth and Bone have affinities of a similar character. Some are more archaic than the common Indo-European and Scythic, the closest affinities being African, Malagasy &c., but each of the terms has one or more Scythic or Yeniseian roots. Most have Australian or other Asonesian aflinities. The common Australian term for “ foot” tina, dina &e. is Indo- nesian and Gangetic-Ultraindian (Bodo a-theng, Lau tin &e.) That for “tongue” talan, dalan, tale, &e. (Tasm. tulana) is Indonesian, dila, tura, jala, jila, chila (Phil., Celeb., Born.), Gangetico-Ultrain- dian, thali Naga, cholai Bodo, and Scythic til, tel, del &e. Turkish, jolma Ugr. That for “tooth” irang, ira, yira, may either be the Dravirian eyiru (anc. Tenil) or a contraction of tira L. Macq., déar Morcton B., tirreg Erab, tirig Muruya, didara Jakun, dara Bangali L. which appear to be connected with the Iranian danta, denta, dens, adamn, dandan &e. The latter forms are also Anstralian and Indonesian, danga Cape York, dang Masid I. Pt. Lihu, tango Bisayan, dungitu Buol, Trang would be referable to dang. . The Papuan vocabularies of Torres Strait have preserved numerous Scythic—chiefly Samoiede and Ugrian—words not * In comparing names of parts of the body it ia to be observed not only that the same vocable comes to he He to different objects, but that a common generic term for “‘ man,” “ body ’’ &c. sometimes accoin nies-them, and is aptto take the place of the term to which it was originally a mere adjunct, e worils for * head oat hair,” fe scull u aé foreheatl rT ee 2 | fice und and ee eye thd for oe nose,” “tongue,” and “ear,” for ‘*hand” “ finger,” “arm,” “leg” ond ** foot” are much interchanged in the Asonesian vocabularies. One of the most prevalent of the words for ** head ” &c. in some of the precedings lists appears to have had a generic meaning in some vocabularies, In Tasmanian we find le-pena “eye, ” -vera-ta “ear,” ro-wela “elbow,” le-pera “neck,” mena “nose” (W. fasm.), mena “tongue” (EF. T., so mi-mena Brumer. I.) ka-mena “ chin,” mana-bena-na knee,” ana-mana * hand,” na-jana “* tongue, ” ma-meri-ka 126 ETHNOLOGY OF THE INDO-PACIFIC ISLANDs. found either in Australian or Dravirian languages, but evidently belonging to the Draviro-Australian era of Asonesia and Ultrain- dia, when the rude Indo-Australian tribes probably possessed some hundreds of vocabularies, more or less related to each other and to those of the early Scythic or proto-Seythic tribes of Middle and Western Asia and of Africa. In the present class of words we find bai “hand,” Samoiede u-bai; pil “ ear,’ Ugrian pil; karu- sa, ger-ip, kowra &c. “ear” (also Australian kure &c.), korw Fin, kul-ak Turkish or Kasia; mus, mush, muchi “ hair,” i-wusa Fin (bus Fazoglo); mit “lip,” Samoiede pite; taip “lip,” Somoiede tip-che; ney “tongue,” enya, inyi Sc. Tungusian, nyeme Samoiede; ‘pit, pichi “ nose,” Ugrian root pid (Chinese pi &c.) The other Asonesian languages have a basis of similar archaic Scythic and Scythico-African terms, but the names for the more common objects, such as the above, have in gene- ral been replaced by vocables derived from the later intrusive formations, Malagasy and Ultraindian. Both of these, and especially the Ultraindian, being very Scythic in their glossaries, it is often doubtful by which current Scythic words found in Aso- nesia were imported. With the Malagasy terms there is in gene- ral less room for uncertainty, although the Malagasy and the Ultraindian varieties of Scythic roots sometimes closely resemble each other. The chief difficulty is in distinguishing the archaic Dra- viro-Anstralian from the more recent Ultraindian terms of Scythic origin, and it is increased by the fact of Ultraindia having been the line by which both of these Scythic currents have flowed to the islands, There need seldom be much hesitation in referring Asonesian words with a well marked Dravirian, Malagasy, Tibe- tan or Mon-Anam form to these sources respectively, but there are ava ny tt ” th ” . iy : le ** fave.” In ‘Tasman nia'we find kure-merang © tongee” Phe tooth,” ri eames iB, en da, owe, Bort. oy” ple ae) ? g ‘nose ™ oareng *‘ hair,” mara-ma ‘eye (Fol, &p.), Brabran. breast,” kom-prian “knee”; in Toro toari-hu * hair," 3 in New Ireland pra-lenhek, -| “ear,”’ bra lima ff a Lee angel + foot,” pal-bulik cp shoulder” ; in Waigui bram-pine K o “arm,’’ enom-braem “hair” ; in Dore “bra-lima “ hand, hay Na rat ores Uy alang, ka-bara ke’ ** head,” me-barai, Rai ee seg mit ke. cya eon tiling “lips,” muru “nose” (ka-muru Celebes), ta-mara bir-il “‘ hand,” wa-para, ma-pal, nga-mura, biri, eee ke + breast,” wo, éc- ETHNOLOGY OF THE INDO-PACIFIC I8HANDS, 197- many Seythic vocables in the Gangetico-Ultraindian and Asone- sian languages which. may either be of the archaic Drayiro-Aus- tralian era or of the later East Tibetan. The Torres St. kerim, kirim “head,” Timor M. Rarain, Tana karab, New Caledonian- gar-moing (kara-mai “ face”) Simang kala (Ceram wkar “ hair”), ° have Tibeto-Ultraindian affinities (koro Bodo, mkura Mishmi, kra Gurang “head”, kara Singphu, Milch., kra Tib. “ hair”), but Malagasy has kara “scull,” and the Kashmiri kala “ head,” Latin cranium, Fin karw “hair” &c. suggest the possibility of a. still older derivation. Another term, koik “scull”” Marray Ls koik “forehead” Port Lihou, koiku, ib., Cape York, appears to, be clearly Ultraindian (koi Binna, Nancowry, kui, kok Manis. puri D., kak Uraon). Tag “hand” is probably not a derivative from the Malagasy tanga, targana like the Indonesian tangan &e., but an Ultraindian term allied tothe Sangir tak-lar, Viti thaka, Tore haka, Vate tako (“ hand,” foot”), otoho Goront. Comp. the Naga dak, Aino dek, tegi, Yenis, togan, tegon “‘ hand,” tak khyai Singphu “foot.’”' Tang and tak are evidently variations of the pame ultimate root. Names of inanimate natural ijoetts . WATER, The names of the more common inanimate natural substances have a like range of connection. ‘Thus for “ Water” there are three South Dravirian terms. Nir Kargi, Tuluva, nirn Karn.’ nillu Telug. may be of Sanskrit Eerersuon (nir Sansk). Itis a rare Scythic and Semitic term enyer “river” Wolg., nehr “river” Turk., nahr Arabic. [t is not found in Asonesian vocabularies. The Koriak inh, Ostiak eny, preserve the n form of the primitive root, which is also found slightly modified in the Lepcha ong, Anam ni, Erub nie, Madura eng. It is more probable however that the original Dravirian form is preserved in the Karnataka tiru, Brahui dir, in which case the Sanskrit must be considered as a derivative of the Dravirian mo- dification nir. The root ti, di, is very common, ti Chep., Milch., di Magar, ti, thi Karen &c. The broad forms are Seythiec, Chinese, Ultrainr dian, Asonesian, The Ultraindian (Luhappu) and Mi- cronesian (Tobi) tara (Champhung thari) resemble the Dravirian, Punal Tamil (anc.), vellam Malayalam “ water” probably 8 128 BTHNOLOGY OF THE INDO-PACIFIC ISLANDS. contain the same labial root that is the most common term for “river,” pa Toda, varw-punal Tam. (anc.), puzka Malaya., pols Karn. (anc), Kurg., hole Kar. (mod. ), aru Tam. (mod), eru Telug. [from yvaru, vers), yer “ water” Gond. The root pa, va, ve, pu, po, (ho) is Scythic and it is also found in all the other families under a very great variety of forms and combinations. In the N. E. Asian and Scythic terms the primitive root generally takes a final r or /, amar, mura, muran, muren, wire, polym, bere &c. “river”’ Scythic, mul, mel “ water” Korea, mimel &c. ‘ water” Koriak. This is preserved in the Sanskrit vari “ water," Pashai wark. The same form is common in Semitico-African, bahr “river” Arabic, mura “river’’ Makua, (Mongol) mumel Felup (Koriak), The Dravirian vara shows the same combination and it has been transmitted to Asonesia,—“ water” warari (Utanata), weari Mairasi, walar Lobo, purai Bathurst,—‘ river” brang Sambawa, umala Buton, marye Trusan, bilo Sydney &c. ~The N. Indian pani “ water” is an allied form, to which the Australian bana and Indonesian banyu are related. In the other Dravirian var-eties the root appears pure with native postfixes, As examples of the simple root with its vowel variations I may instance the Samoiede bu, bi, be, Pashtu aba, abu, abe, (Sanskrit apah, Zend apem, Persian ab.) The term tanni is confined to Tamil. It is veniaeian. tatany “river”. The Gond donda “river”, Todava tude “river” are probably variations ofthe same term, It appears to have spread into Ultraindia, dak-tani “ river’ (dak is “ water’’) Ka, tunli Khom. vital The Male am, Uraon um, Kasia ka um “ water,” is a variation of the labial root already noticed. This form is found in the Semi- tico-Libyan family, ma, ma-at Arabic, mek Gara, maim Hebrew, me Galla, mah Egyptian, (ba Malagasy).* The Nicobar mok, mak, Tasmanian moga, like the Gangetic um, am, may have more direct N. Asiatic afinities., Comp. mu, muh, make, Tangus. (waka Aino.), The Kol dah “ water"’ is a very common root,—Sceythic, Iranian, Ultraindian, Asonesian. It is probably the Sanskrit udak which appears to have been early received into the Kol or Gange- * It is also found io Brahui, taho “wind.” ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 123 tic vocabulary and thence spread to Ultraindia and Asonesia. Mon dat, Tobi tat, Ka dak, Khom. tak, tag, Nicobar rak, But it may have been pre-Sanskritic in India and Ultraindia. It is Yeniseian, dok, Fin tat-se and African, mdok “ water”, dck “river” Woloff, dogo “river” Galla, date “ river” Fazoglo, the root being da, ta, &e. Other variations are found in the Turkish elga, Ultraindian lik, lika, Suowar ri, Burman mrik (Rakhoing dialect) Khom. prek, Asonesian leko, ilug, Galla lega, where the root has the slender form li, ri, le. The Uraon cheip “water” is ba Sh pe che Miri &o. &e. The Kol garra, Uraon khar “ river,” Chepang ghor, is Kashmiti, kol, kuol, Pashai gal, Semitico-African,—khar Gara, khor Mahrah, Ar., koli Tigre, kor-ama Hausa, gar ‘water’ Saumali, kero “water'’ Darfur; Mongol gol, Samoiede kolda, Yeniseian “water” kull, Wog. “water” agel, Javan. “river” kali, Australian “ water’? kali, kaling, kalere &e. . AIR. Of the four South Dravirian terms one is New Guinea and Australian on the one side and Georgian on thé other. The root ka, ga is archaic in Dravirian, taking different native postfixes.* The North Vindyan ta is probably a variation of the same root. It is Scythico-Iranian (at Fin, ot Armenian, atma Sansk. &c). A second teri, ela, is Scythico-Iranian, Semitic, Ultraindian, Asonesian. A third, bar, is Scythico-Iranian, Ultrainiian and Asonesian. A fourth, puy, is N. Indian and Australian. The Kol yocable is found in Anam. It appears to be archaic and related to the Semitic hawa &c. STONE, The principal vocable kal, kala, &c. has spread to Weoudeet kala Polynesian, kain Australian, the latter being closer to the Pashtu variety of the root, kani. The Sindhi kod is nearer the Dravirian, which is the pure Fin kalle, Armenian khar.+ The term is also N. E. Asian (Yukahiri, Kamschatkan). * It is Brahni, khall. + Koelle’s vocabularies su ply mel, anes men, man N. W, Nigritia, omi, ams, mi, min, mmeli, amn, } Niger, Chadda & «780 BTHNOLOGY OF THE INDO-PACIFI® MSLANDS. o ; MOUNTAIN, HILL. * Five yocables are current. ‘Two, found in Gond and Male only, are of immediate Tibeto-Ultraindian derivation, Both are va- fieties of a Scythic root. The others are archaic. Of two forms of a Scythico-Iranian root, one, found in Kol only, appears to be the more ancient. It is Samoiede, Saumali-Galla, Australian, Celebesian and Philipines The second form is Fin, Iranian and Australian. The third root is also Seythico-Iranian and in form Yeniseian. In Asonesia it has only a limited Indonesian range. ‘Of the two Southern roots, one, var, par, mal—Australian wahr-ro, wariat—is Scythico-Iranian. Fin ware, Ostiak palta, Wolg. wanda, panda, Sensk. parva, Kashm. bal, wan, Aino buri, Samoiede boro, Turkish muron; Fin wuori, Latin mons. The u form is preserved in the Saumali boro, bor, Galla borga on the one side of the Indian Ocean and on the other in the Kol buru, Australian murdo, mordo, murde. It is also found in Indonesia as a word for “hill,” Philipine puru, palu, Celebes bulu. The New Guinea wera is probably a local modification of the Draviro-Australian war but the form is also Samoiede bre, Ostiak pel, Caucasian méhr, German berg. | ~ The second root kon, kun, kud, gud, is also North-Gangetic gun, kung, kang, and Indonesian, gunong (Tamil konom). The root is Seythico-Dravirian, but the Draviro-Asonesian form is Yenisiean, konony. This broad form is also found in the Fin gora, karuk, gures, ko, Persian ku, Zend kof, Latin collis, Mahrah kalun, gar-tin. Another Yeniseian form, kar, is found in Pashta gar, Galla gara, Maram kalong and Ceram ukaro. The slender Ugrian ky, keras, Turkish kir, is found in Georgian kirte and Sanskrit giri, The ultimate guttural root is Chinese and Siamese as well as Ugrian. The Karnataka and Tuluva forms gudda, gudde—preserved in the Kol gutu “a small hill’— appear to have spread into Asonesia ata mueh earlier period than the South Dravirian and North Gangetic gun, kun, if the Austra- lian kata be referable to it. | In the North two other vocables are found, dungur, dongar Giond, and toke Male. Both are Gangetico-Ultraindian, Tibetan and Scythic. Sindhi has also received the Japanese, Turkish, Tibetan and Male form (takar). The New Hebrides takuar, and tof appear to be allied to some of the preceding terms. ETHNOLOGY. GF THE INDO-PACIFIC ISLANDS, 181 Terms of the preceding classes exist in the primitive era of glossology, and the roots current for them in a family of languages may be of greater antiquity than the formation itself. The vocables we have examined are of different ages in the Dra- virian family; but most of them must be considered as at least eoeval with the formation, while many of the roots have probably existed from the monosyllabic era, first receiving their present forms when the linguistic type became Scythoid. They do not necessarily throw any light on the archaic condition of the race or on the early history of its civilisation, for such terms ave essential elements of human speech in all ages, and they are found in the vocabularies of the most barbarous as in those of the most calti- vated tribes. The forms of the yocables indicate a large measure of community with the Scythic, Caucasian, and primary Iranian races, and a less one. with the Semitic, but this community may belong solely to a yery archaic and barbarous state of society similar to the Australian, for anything these terms can teach us. I will now take a few words implying an advance beyond such a condition, and indicating the possession of certain arts and usages of a civilised character. Names of Domesticated Animals. The domestication of the dog, and that of the hog, of the cat and of the fowl were probably amongst the primeval events of human history. That of the larger quadrupeds must have been later, although it may have long preceded the Australian era. All that can in strictness be concluded from the absence of the large domesticated animals in large portions of Asonesia is that the means of carrying them to the islands did not exist in the Austra- lian and Niha-Polynesian eras. The light which this class of names can throw on the early history of the Draviro-Australian family must therefore be confined chiefly to the continental branch. The comparison of the names of domesticated animals is compli- eated by the fact that. they have been interchanged to a remarkable extent. This has arisen from tribes being apt to apply to those with which they become acquainted for the first time, the names previously current for others with which they are familiar. It is not surprising that the “cow” and tho “ buffalo” should be known by similar names, or even that a tribe which possessed the 132 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. cow, should include the horse in the same term when they first saw it. But we have modern instances of races which knew only the dog and hog, applying one of their names for these animals to the cow, and a comparison of vocabularies shows that in archaic times @ similar course was frequently followed. In fact some words have been so much pressed into service to meet such emergencies, that if all their applications were included in one vocabulary the generic meaning of “quadruped ” might be given to them. The name for the dog has been applied to the cat, the hog, the cow, the horse &e. Many of these new applications become valuable guides in tracing the spread of particular roots and varieties.” Another source of difficalty and error in comparing the names of domesticated animals in different families of language is that these names are liable to change repeatedly, long after the first acquisition of the species. Such terms, and particularly those for the horse, are apt to be spread with the breed into foreign coun- tries. In many groups of languages, owing to this and other eauses,+ there are various terms for the horse, having distinct ranges of foreign affinities. Thus in English we have horse, Semitic, (also African and ultimately Scythic), mare Scythic (and African), colt probably a Celto-Seythic term (gormydd Welsh, hunde Samoiede, &c.), the Irano-Celtic equus, each, in the equine terms derived from Latin, foal Ugrian &e. These terms had probably separate origins, and belong to different eras of Moglish and of Teutonic or of Indo-European history. CAT, 1. The most common term, pusei, pusi, puchcha &c. is N. Indian and Indonesian. It is also African under the form mus, musa, &c. and English puss. The Pashtu slender form pishik, pishee, Brahui pishi, Milchanang pish?, found also in Ultraindian pishik (Kapwi), and Rotuma pitsa, is Caucasian pishih (Chari), Semitic bis, African topisa, fisona and Ostiak misah, (also matska). Tt is probable that the Brahui, Pashtu, Milchanang and Kapwi * Some illustrations of this have been given in the glossarial Appendix to the tico—African sub-section. + The chief of prea is “the: redundancy of terms to denote varieties of familier pe sort is which most yocabularies appear to luxariate in certain stages of t ETHNOLOGY OF TUE INDO-PACIFIC I6LANDs. ~ 133 are derivatives from a western term now represented by the Chari and that the dissemination of the broad form was a later event. Rut mushik, mushak is also Scythie. The Chari term with its postfix is referable to the Ostiak and the Semitico-African terms to Chari. The m form was probably the original, as it is found in Ostiak with the slender vowel, mis, in Africa with the broad vowel, mus, and it may be added in a widely spread word for mouse (mush Sansk., mus Lat., &c). Piss, pess, poss, pass is “dog” in Sclav. and pisse is “ mouse” in Samoiede. 2. The Toda kotti is a common Scythico-[ranian and Cau- casian terta. Kuti, kata, kato, kiti, kotshwm &c Ugrian, kisa, kazhi Fin, kot Sclav., katze German, kat Dutch, English (cat), gato Spanish, gadu, kit Armenian, kito keto, geto, koto, kata, katu, gado, gedu, gadi, cheto Caucasian. The ultimate root is’ found in Korean koi, kuiné “dog”. The form kot, kok, kit &e. is also widely spread as a term for “dog.” [See Doa.] 3. The Karnataka biku, beku, may be a contraction of birku, berku, from the analogy of the Uraon birkha, Male berge. But as Gond has. bokal, bhongal for the male, and Marathi boka (com.), and the Karn. form is found in Batta (Sumatra) asa term for “tiger”, biku, and in Buton for “ cat”, beku, it is probably distinct. Comp. popoki Polynesian, paka, mpaka Suahili &c. and a widely spread word for “dog” mog Tarawa, &c and “oat,” bok Dutch, bakra N. Indian &c. ‘The ultimate root is probably bi and identical with bi, mi of 1. Mongolian has mi. 4. The bir, ber of Uraon and Male is found in Gond bilal, Bengali biral, Kol bilai, and is*a common Hindi and Tibetan term. Serpa and Sunwar have the Male form bermo (Murmi tawar, Gurung nawar). In the South Dravirian and Gond dialects it is the prevalent term for “tiger,” pili, puli, huli. The Maldivian bulau “cat”, has the broad vowel of puli. The Kashmiri brair &c. resembles the Bengali and Gond. The root does not appear to be common, unless it ia prevalent as a word for “tiger”, Hind. palang, Pers. palank, Arabic babir, &c. [ Batta babet] Korea pon. But the Latin felis shows that it is not confined to southern Asia. The same root is probably contained in some terms for “ dog" balu Maldiv., balla Singhalese, a-val Champhung, perro Spanish, wuri New Guinea (Utanata), wurie¢ 134 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. “cat” Gafat, (the same form with the Semitico-Libyan fem. postf,) Bil, bi-r, bi-s (bi-t) and bi-k appear to be the same root with the ordinary Scythic range of commutable finals. 5. The Kurgi nari, Malayalam niri “tiger,” Burman nira, is Korean, nal-bi (Amharie nahar). ) bDoG. 1, In the most common Dravirian term the root appears to be na, la, ra, (nayi, naya, nai, noi, alay, ala, era). It i3 related to the Gangetic nangi, nagi, nagyu, neko, and to the Savo ngaka and Australian nagi, nago, these Asonesian terms being evidently of Gangetic derivation. In the vocabulary I have considered it doubtful whether na be a root or a prefix in these terms, and leant to the opinion that nagi &c. was a softening or contraction of nangi, nagi. From the analogy of other Dravirian terns I now consider it clear that yi, ¢, ya is merely a definitive postfix, and na, la, ra, the root. It is Australian alai¢ (allay Male), alli. [The Poly- nesian uli is a contraction of kuli, similar elisions of the conso- nantal initial of a syllable being common in that language]. The Drayiro-Asonesian root is Circasian lah, Georgian laki [=na-gi Gangetico-Austral.] and N. S. Asian iau,-Aino, Japanese. The Bisayan fra is probable of modern Japanese, and not of archaic Draviro-Australian origin. The Tungusian nyin, nenaki, nenakin, Mongol nokoi, nogai [Samoiede weneku, bu-nike, kanak) appear to be related both tothe Aino, Japanese and Dravirian, and to the Caucasian and Gangetico-Australian varieties. The term is not a common one, and it appears in the Dravirian-Australian family to be older than the Scythico-Iranian era of its-glossology, when other vocables for “ dog" were widely spread over middle and western Asia. It may either belong to the primary glossarial basis, of a N.E. Asian character, or to the allied Semitico-African for it is found in both, Hottentot arieb mase. aries fem., Serakoli uley, Galla luru-tai, lurul-tai. The close resemblance between the archaic African ari orarie and uley and the Male-Australian alay, alai, ali, renders it probable that the latter is of the Semitico-Afri- can era of Scythie or proto-Scythie like so many other archaic Asonesian vocables. | 2. The Telugu kukka is exceptional in the South and probably ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. "135 of later acquisition. It is North Indian,* whence it has also spread to Asonesia both in the Bengali-Telugu form (Beng. kukkur) and in the Hindi (kutta, kutto). Bajo koko, Mangkasar kokang, Kagayan kito, Kissa gida, Endeavour Riv. kota, It is E. African kutta, kutti Danakil, N. E. Asian, gottun Koriak, kossa Kamsch., and as applied to the “ cat” Korean, Scythic, Caucasian, Iranian, Dravirian...The guttural root under various forms, and the same root with other postfixes, l, r, s, is very common in Scythie and all the families of language that have a large glossarial element of Scythic (Tibetan, Ultraindian, Caucasian, Semitico-Alrican, Ira- nian, Asonesian of different eras.) For some examples of this wide diffusion see the Africo-Semitic subsection and the Tibeto-Ultra- indian and Mon-Anam vocabulary in Chap. vi. As it is also applied to the“ cow” and the “horse” it was proba- bly one of those words that were early used for the first domestica- ted quadruped.. As examples of these applications I may instance for “horse” the Yeniseian kut, kus, .E. Iranian and Dravirian gud, ghota, ghora, kudra &e, for “cow” the Ugrian kusa, kas Tungusian kukur [in Bengali “dog”], and for both “horse” and “cow” in the same language, the Yeniseian kus, kut, and the Kamschatkan kousha “ cow”, kasa “ horse.” 3. The sibilant root of the Kol seta is equally prevalent with with that of the preceding term. It is Gangetico-Ultraindian and Indonesian, Iranian, African (the sibilant sometimes changing to the aspirate). The Kol variety appears to be an archaic Dra- virian term, It differs considerably from the prevalent forms both on the Irano-Caucasian and on the Gangetico-Ultraindian sides. It resembles the Aino sheda (Kamsch, hetan) more than any of these and as usual some analogous forms are found in the upper Nilotic vocabularies Agau gezena &c. The Caucasian he (Chari) appears also to preserve the N. KE. Asian form, unless it be a contraction of hue (Awar) which has the broad Ultraindo- Asonesian form (asue, asu, su &c.) analogous to the Sanskrit shoa. The Kasia kasen is probably a derivative from sa Kol. [For other applications of the root, see Hoa.] , 4. The Singhalese and Meldivian balla, balu have been men- tioned under ‘ Cat.” * Brahui has a variation of the same root, kuchak. " 136 ETHNOLOGY CF THE INDO-PACIFIC ISLANDS. HOG. There are two terms, both archaic. 1. The form of the common Dravirian term panri, pandi, panji, panni, poti, padi is peculiar. The ultimate root pa has numerous foreign affinities, being found by itself, and with other postfixes in Seythic, Tibeto-Ultraindian, Malagasy-Polynesian, and African. (See Tibeto-Ultraindian Voc. App. to chap. vi.) If the » be part of the base the closest affinities are with the Scythico-Iranian base, par, por &c of por-os, por-k, par-s &c. » 2. The anc. Tamil kezhal, Male and Uraon kis is Circassian kasha. In Asonesia it occurs in the Batan kuis. The root is an archaic one, widely applied to “ dog” (also to the “cat”, “cow”, *‘horve.”) The Kol sukri is Sanskritic. GOAT, SHEEP. Many of the vocabularies to which I have access do not contain these terms. All the Dravirian vocables for “ goat” are also used for “cow”, “buffalo” &c in other families of language. The most common is Caucaso-African, and Iranian.* : BUFFALO. The ancient Tamil and Vindyan term is exclusively Indian, Ultraindian and Indonesian, and the probability therefore is that the native wild buffaloe of India was originally domesticated by the Dravirians themselves and then diffused to the eastward. The same root however occurs in other languages applied to the “cow,” and it would thus appear to be of Seythic or at least S. W. Asian derivation. The current Dravirian term has Chinese, Ultraindian and, as applied to the “cow”, Scythic, Sclavonic, and African edinities: It is probable that both terms were used for “ cow ” before the Dravirians applied them to the buffalo. 1, karaz Tamil ane. [karavai “a milch cow’’], karu Karn , Tod. “a young buffalo,” haliya Gond, kara, kera Kol. The term, like many others, has spread from the Kol (or an ancient Lower Gangetic language) to Ultraindia. It is found in the Kambojan family joined with a common root for “Cow”, Kar- bu Kambojan, kar-pu Ka, ka-pao Chong. From Ultraindia it has spread to Western Indonesian (karabau, karbau, kabu, ka- rambau, karbu, kapa, kawa, &c. &c.) thus indicating the country * Math, “ goat,” Brahui, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 137 from which the buffulo was first imported by a civilised insular nation. The term is a common Scythico-Indian and Tailo.European one for “cow” gorn Hind., wkur, hokor, kukur Tungusian, karo-wa, koro-wa, kar-we, kra-wa Sclavonic, kur Icelandic. In the other Scythic languages it is generally combined with the sibilant root, sa-gar, is-kal, ush-kal, sy-gir &c. Ugrian. Tt has a wide currency as applied to “ horse”, “ dog’’, “ cat.” 2. erumai iramai Tam., eruma Mal., enumu Telug. erme Tuluv, yerme, emmz Karn. ira, ir, Toda [“ cow” ,—uri Mandala, udu Uraon, of Male; “ bullock:”—yerutw Tam. yeltw Karn., yelta Tod.] As applied to the buffalo it has no foreign affinities. But it is a Scythico-Caucasian, Iranian, Semitic and African 1 oot for cow.” [See Cow 6.] 3. mankha Uraon, mange Male. This term is of Tibeto-Ul- traindian derivation, man “cow” Naga (root ma, ba, pa, &c). [See Cow 1]. In the Ultraindian languages the root is also applied to the buffalo. 4. bhitkil Mandala, budkil Gond (Gawil,) bode “a female buffalo”, Gond ‘Saonie Chapara’. Bhit isa North Gangetic form of an Ugro-Dravirian term for “ Cow” [See Cow 1.] The root is also applied to the buffalo in Gangetico-Ultraindian lan- guages. cow. In some of Klaproth’s Middle and North Asiatic vocabularies this important word does not occur, and I am thus without the means of fully tracing the relation between the Dravirian terms and those of Upper. Asia. 1. The most common ultimate root is pa, pe, which is Tibeto- Ultraindian, but it appears to have been acquired by the Dravi- rian family with a sibilant or dental postfix, pas, has, pet, Singha- lese ves. The North Gangetic and Ultraindian forms resemble these, pit, bik, &e Lepcha, Limbu, Kiranti [ga-bhi Bengali], mashu, masi, &c, Chepang, Mishmi, Bodo, Naga ke. The roo; in this form is Scythico-Iranian, mes, mis, mus, Perm., misye Wog., mes, neng-mes Ostiak. In these Ugrian languages there are distinet terms for “cow” which are also found in most of the other Ugrian and in the Tartar languages. Mas &c is absent 138 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. in all the Caucasian languages. In the Indo-European family it appears in the Latin bos, which in the oblique cases discards the postfixed definitive and restores the root bov or bou of the mono- syllabic group (bou Anam, woa Lau, mwa Burm. ba, pha Tib.) In Semitic the Scythic term appears to enter into words for the “‘buffaloe,”” gam-bus, ja-mus. In Africa it is rare. The Gonga miza, mia &e, Kosah maas “ cow” appear to be Ugrian. Other Ugrian terms for “cow” are common in Africa, (See Appendix to Sec. 6). Asa term for buffalo it occurs in Milchanang mosh, moesh, and somewhat further modified in the Hindi bhains, Himalayo-Ultraindian meshi, mesye, moisho &c. The original of all these terms appears to be the Ugrian mes, mis, mus, which in its turn, is probably the monosyllabic mo, bo, ba, pa, &c with a consonantal or final or a definitive postfix “father,” “ female, ” “male”. The ultimate root is one of the common primary terms for “ mother.”’ The Dravirian terms cannot be considered as archaic, or prima- ry. The root has evidently passed through the Ugrian glossary. The Malayalam paya, Kurgi payx may be exceptions. 2. The Telugu and Karuataka, avu, and anc. Tamil a appears to be a form of the primitive root, It is identical with terms for “mother’’ Drav. ava, apa, Manip. avu, aphu &c. (See also “Father” 6). The Egyptian ah, ha, aua, Emghedesie haui are modifications of the sibilant root for “cow,” and distinct from the Dravirian a. 3. The Karnataka akalu is Caucasian, aka, ata (Lesgi). The dental form of Lesgian is current in Pashai, ada, “bull”. In the form ta it is a Drayirian term for “ mother,’ tayi, tali, also with a different postfix “ father,’ tandei. The Caucasian form is Tur- kish (ata, aga), Ugrian and Chukchi (ata) “ father.” The same form is )ravirian with a feminine application, “ aunt,” “ mother- in-law”’ &c so also Sanskrit tata ‘‘ mother,” Nias eta &c &c. 4. The Karnataka dana, danam, Tuda tanma contains one of the varieties of the same root as that of the preceding term.* The root in both these terms is Seythic and Seythico-Iranian and the application of this particular term to the “ cow ” is Caucasian. * Dr Stevenson refers it to the Sanskrit dhen ‘* a mileh-cow,” but it seems more probable that the term is archaic and merely a variety of the root i ke. Th Sanskrit term I may remarks is Ugrian, thea ie vie. Cae rerts
lung ” * Koelle’s vocabularies give o!ungu, nlungu, as acommon W. Zimbian term for ** canoe,” 80 ig Mose, One of the most common terms is oko, go, kokua, guro, kor, kulun &c. The idewity of lunga with a common Gangetic term is remarkable. If this form as well as the modification donie, dingi &e was current at an ancient period on the Indus, its transfer to Africa would not be surprising. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. olung olong tilong dunga donga dongo doni laung milaung plaung li thi nli tali mali mari-kho mahi malhu lhe Rhli ria. kieng Alui liing dingi ring alina tina tena in-dyn longa tidong palang oria era rai aruer Miri Lh Mikir Murmi, Magar, Sunwar, Kol Bengali, Hindi, Newar, Uraon, Male Gond Karnat, mod. Lungkhe Khumi 6 Ayo Singpho, Kapwi, Maring Maram, Rakhoing Maram Jili Koreng, Khoibu Champhung C. Tangkhal Khoibu Burm Karen Siamese Mon Khyeng [=gru Tib.} Kasia Bengali, Sindhi Garo (Brown’s voc. ) Tilanjang Sambawa Sulu Murare ((New Caledonia) _ Tagalo Kissa [Mikir,] Jav. (7a variation of the Draviro-Sansk. term palava Kc.) Dorei [Siamese form] Mairasi Onin Gebe ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 151 ora Toro (Salomon Is, ) lo Eromango rarua Vate [= Gebe] biri Lobo piari Utanata mari-nawai Australian mari-gau ” mara-gau 3 war-gal ” gul ~ Erub, Masid, Pt.Lihu kuere Vanikoro donie Saumali donah Z deuniké Danakil raul Andi (ship) adawle Arabic ,, dau ” ” worga Samoiede oddu A antu Fs anu ry al Yakati The form in ris a common Samoiede one, and the Indo-Euro- pean nau, navis &c. appears to show that itis an archaic variety. The Indo-European word is more immediately connected with the Andi and Arabic form rau, dau. The Bruner Island daow, raow “a catamaran” is the Timor benau, wenau “boat”, Vanikora naue, Tuanlu (New Cal.) nayu, “ boat.” The Australian nawai may either be the same term, or the Waigiu wai, Port Dore ua or wa, New Caledonian uang. Dau, rau, nau may be remnants of a once common Indonesian term of archaic Semitic origin like the Vanikoro baito, “a house,” and many other Asonesian voca- bles, but the Ultraindian plaung and the Arian nau igi aoe to render the etymology doubtful. The New Guinea and Australian terms may be archaie, but they have every appearance of being derivatives from one branch of the- Ultraindian languages, the Manipuri and Yuma. If this be the case,they form a remarkable record of the period when this Ww 152 ETHNOLOGY OF THE INDU-PACIFIC ISLANDS. branch furnished the maritime tribes of the Lower Irawadi and Arracan, aud would tend to show that the Uliraindian navigators of this era were the first who had sufficient intercourse with the races of Torres Strait, and the adjacent shores of Australia and New Guinea, to communicate to them the namies of their vessels. Another Oceanic term, nearly displaced like the preceding one, is also referrible to an Ultraindian source. koa Tibetan khuonkho Naga khoa 9 khung 5 khombe Limbu ~ kupok Nicobar mari-kho Champh ung mari-khong Luhuppa- ra-kong N. Tangkhal kowa Savu rai-koi-koi Onin kinung Kissa jong-kong Bima, Tidori, Bajo ja-kong Bali, Sasak, Buol koina Maori [koi Onin] kuan - New Ireland kucre Vanikoro ku Utanata nel-kou Aneiteum (New Heb.) kalau, kui —— Loyalty Is. kwa Nikete (New Caledonia) It will be remarked that the Sayu kowa, New Ireland kua-z, and New Caledonian kwa are faithful to the Tibetan and Naga koa. To complete the history of the Dravirian maritime connection with Asonesia it may be added that the Indonesian term for a square rigged vessel, kapal, is also Dravirian, but of much later origin. It belongs to the Telugu or Kalinga era of Indonesian civiliantion: like the Sanskrit element in the languages of the civi- lised western races. The Dravirian term appears to be a Seythico- Caucasian root witha native postfix, kapal. Circassian kap, kaf, kuafa, Turkish kuafah, Ugrian kap (Wolga), chap, (Ostiak) Tur- ETHNOLOGY OF THE INDO-PACTFIC I&LAN DS. 153 kish kebe [also keme, kema, kama, gome, gemi &e.], Misjejian kema, Lesgian gyami “ship”, Sanskrit kepa. Some Asonesian terms for “boat” appear to be rather referrible to the prevalence of an archaic form of this root than to the modern kapual. kakabei . Bawian kabbi Erub kabi-nar* yy Murray Teo [kupok Car Nicobar kopapa Maori] In’ the last term the root is probably papa (Nias bubu) Tonga bopau, ko being a common Muori pref. Chinese terms are not found in India save in Sindhi, although they have become current in Tibet on the continental side and In- donesia on the insular. The Tibetan syen is Japanese sen, Chinese, ch’hiang, chun, siau, chiu, thiang &e, The form. chun is the original of the Indonesian jong, jong-kong, whence the European « junk.” The Chinese sampan is also current in Indonesia. The Sindhi jhamti appears to be Chinese ch’hiang-toi Macao. The Tibetan form has descended to Northern Ultraindia ihseng, yesang Naga. The Anam chuyen, Burman song-pua &e. are of direct Chinese derivationt. Lal Houst, The most prevalent term is South and North Dravirian, Kol and Himalayan, in different forms. ‘illam Tam, anc., Mal. illu Telug. illa Taluy. erpa ~Uraon — arra Tou I, oarra ss Gond. ad oura a ron - ora Kol * Nar is also current as a separate term in the Torres Strait dialects, [t may be a variation of the Australian [ Ultraindian] mari, mar, or directly Gangetic nat Bodo, nawar Dhimal, navar Lepeha (See nav, nau &e. supra. + The Anam ding of Mr Brown's vocabulary given in App. to chap. vi. jad not appear to be a genunie Anam vocahle, an: tthe Chinese ting isa forin ofa differ- ent root, 154 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. au’ Kol oa a uraa i ura Brahui li Lepeha le ” * ara Champhung riang Siamese lehu Aru hale Simang, Pol. sarin Lobo hari Polynesian ri Rotuma [= ha-ri, ha-le, sa-rin ] alaya Sansk. aula Latin arre Danakil ille Yoruba ire Japan . rat Koriak 2 dih Persian, “ village ” The slender South Dravirian il, er, is the Himalayan and Asone- sian form, and as it is also Japanese and African, it is probably one of the primary Dravirian vocables. The broad Tuda, Gond and Kol form has a stronger resemblance to the Indo-European, but is doubtless of equal antiquity with the other as a Drayirian term. The Tamil vudu, uidu, Malayalam vida, Male aya and Changlo phai, is found in the Asonesian bahi Sulu. It is probably con- nected with the Semitic but Gara, ut Curia Muria, bet Arab., be:h Hebrew, baiti Mahrah, abaita Egyptian, mitse Shangalia, the Gara having the Tamil, and the Himyaritic (Mahrah, Egyptian) the Changlo forms of the ultimate root (bu, be, bai, vi). The Viti mbeto, Vanikoro baito are Semitic inform, The labial is com- mon in Asonesia under different forms uba, emu, ima, im &c. but these may all be derivatives of uma, umo, &¢., a contraction of ruma, huma, ramo &e, The Sunda ima, Mille im, Sydney mya, have some claim to be considered as an independent root. The Semitic vocable is also Samoiede, mat, matsch, matn, met (Arabic ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 155 bet) in which ¢ is evidently the definitive or consonantal augment, the forms me, mye, ma, men also occurring. The root.is also found reduplieated in Yukahiri, meme, and the Chukehi mantaak appears to be connected with the broad Samoiede forms, The Koriak wal-charat has the same term in its first element. These broad N. E. Asian forms appear to be also related to the third Dravirian term manei Tam., mane Karn. Like the other Dravirian terms they are also found in the Semitico-African family, mana Galla, The Sanskrit balai “a hall’, which is found in Asonesia with the same meaning, and also with that of “house”, which it has even in Polynesian,—fale, fae, mare, vale—and Mi- cronesia—playe (Telew),—is a cognate term. To the Semitico-African form bayith, bait, baiti, mitse and the Dravirian vida, &c., the Iranian vish Sansk., bati, basa-stlan, nibas Bengali, basti ‘ yillage’’ Hindi, are allied. IRON, SILVER. The African affinities of the Dravirian word for “iron” are the closest, aud as the common Dravirian word for “silver” vili, bili, is also a Semitico-African term both for “silver” and “iron,” the Dravirian words for the latter may safely be placed in the same class of relations. The eastern prevalence of the Himyaritic forin filat is evinced by the Indonesian pilak, perak ‘‘silver’’ (ber is a prevalent Caucaso-African form of the root), But the Dravirian term does not appear to be of similar recent derivation. It has not the Semitic postfix, and in some of the northern languages of India the root occurs in other forms amel Abor, Miri, mil, mul Milchanang, mul Tiberkad. The same root, primarily meaning “white,” “light,” “bright” &e., has been applied to “silver” “moon,” “sun,” “stars,” “fire,” and to “iron,” “gold” and other metals. The direct application of the qualitive “white” to silver has been twofold. The primary one was to call the moon by the name “white.” When that name had become a generic substantive for metal, the same root ora different one was, in some languages, again attached to the primary form as mére qualitive, “white-metal.” Hence the various forms and applications in which the root is found in the Dravirian languages do not necessarily Lelong to the same era. The Dravirian term for “white” is velliya, velluta, bile, bilige, 156 ETNNOLO@Y OF THE INDO-PACIFIC ISLANDS. bollane &c. The root is veli, ‘bili, bile, velu, bola &c. If with these terms for “white” those for “silver” be compared, no doubt will remain that the qualitive root was thé original. Drav. vili, bi'i, Semitic filat, filati &e., bir, berur. In many of the African lang- uages, as in Kol, it is applied to “iron,” and as the root appears with the same meaning in the Hebrew barzel, ag well us in the Georgian and Latin terms, in Kamschatka waratsh (the base apparently of the Hebrew bar-z-el), Koriak waland, this appéars to have been a very archaic application. But it must have been a secondary one. The etymology is conclusive as to this. Silver is: found naturally in the metallic’ state, and must have attracted the attention of man long before the art of converting ores into malleable iron was discovered. When other metals came into use, the word) in accordance with the most archaic plan of naming, would ‘probably be applied to them with some distinctive epithet, and it would in some languages come’ in time to have a generic meaning equivalent to metal.” When, for example, “gold’” be- came “‘yellow-metal,” and iron “black-metal,” silver would become “‘white-metal.”’ The basis vocable might come to be applied to gold, silver or any other’ metal exclusively, in the ordinary course of glossarial conversion and displacement. In Malagasy the same root we have been examining is found with the generic meaning. Gold is “ vula-mena’”’, * metal-red”’, and “silver” is vala-futsy” “‘ metal-white”’ or simply vula’’, thus showing that the earlier ap- plication of the ferm was to “silver.” The Malagasy word for “ moon” fula-na, vula, involves the same root, and carries us back to its primitive meaning “ white”, The current term for ‘ white,” fitsy, is the Agau fuchi. In African languages the root occurs in terms for “gold,” as well as for “silver” and “iron,” warka, wirka Agau, Woratta &e., wark Tigre, baru-bera Shankala, wura Yoruba (the Malagasy-Asonesian form). In other languages also the moon has derived its name, or one of its names, from its being “‘ white,” “bright” &ce, The anc. Tamil name is piret, one of the Telugu names is za-billiand the Male is bilpe, all following the Karnataka form of the root (bile white’ Karn,). The Male name for “san”, ber, and the Kol names for “ star” epil, ipil have the same root, while the Uraon binka, Male bindeke, Tamil redu- plicated yvin-min, van-min, Toda pone-min Malayalam minganna, ETHNOLOGY OF THI; INDO-PACIFIC ISLANDS. 157 Karn. mine are but further variations, similar to those which the root has in the North Gangetic languages, mil. “ silver’ Milch. The Sanskrit chand and Malayu-Polynesian sina are applied in the same way, the former being “silver” and “ moon”, and the latter “ white”, “ radiant”, “ moon.” From the preservation of the root with its primary meaning and its reappearance in the Drayiro-Australian family in various names and in various forms, it must be considered as belonging to the glossarial basis of the family and having its closest archaic affinity with the Ugrian wal, wel which in Dravirian are best represented by the forms val, bal; vel, pel &c. Some of the rarer Semitico-African metalic terms have the same broad archaic forms and they are also Kamschatkan. The i forms are variations of those in e, and their prevalence both in Dravirian and Semitic appears to show a secondary and direct connection. The Dravi- rian terms for “silver” are applications of the native root for “white”. Were the Semitic derivatives from the Dravirian ? White veliya Tam. ane. velutéa » mod., Mal, pelpam Tuda. bile Karn. bilige * valtu‘ta Mal. (double postf.) baliad Kurgi bollane Tuluv. phulwm Gond pundi Kol punda iF punia » pandru Uraon panguro — Male balih Kasia mabulau Pagai babilim Kahayan buran Solor fuluk Roti wilban Sydney (comp. Toda pelpam) pila Hindi 8 RTHNOLOGY OF THE INDO-PACIFIC I8LANDS. walgi welkes abyalk fari “ght velichcham oli pelch veluturu . belaku berachi billi aveli naval war Moon bilpe pirei banai Star epil épil binka bindeke Sun ber Fire benki Moon berra werke fylein vulana vula bula, bulan Sky vin vanam manam minnu banu ban pone wan thang-wan Fin HF] Arabic Hausa Tam. mod., Mal. ss aHlOGs Tod. Telug. Karn. Gond Karn. Danak. Tigre, Galla Felup Malag. Ason ’ Ason. a7 ‘am. anc. »» mod Mal. Telug. Karn. * | Tod. Lungke Khoib. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 159 tang-ban Kapwi fa Lau wono Australian awan Sabimba, Sambawa “cloud” Malay &e. wang Madara banua Nias wang-hi Naga wan Lau ban .* . With these terms compare Dravirian and Semitico-African words for “silver”, “ iron” and “ gold”. Silver vili Tamil billi Mal. billili Karn. mil Milchanang mul ts Tiberkad amel Abor Iron merhad Kol merhad a medh ” . marhan 99 panna Uraon phalam Kiranti, Magar, Chepang [Gond ** white’’] per Changlo mpri Singpho maru Tangkhul mari : puruti Dtanata wuru-sesi Lobo " -wur-sasi ‘Mairasi mumu-mur Lobo qmaumu-moira Aru m Sydney Silver filat Gara filati y Mahrah filthla Arabic 160 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, berur - Tigre bir Amharic biri Galla biroa - Agan aimira = ber bs bersh Gafat birisk Pr biro Gonga bira Woratta Iron barzal Hebrew bir - Sanmali sibila Galla birta Danak. birto Gonga beretish Gafat bertt < beroa Agau ba Egyptian vi Malagasy thaty ” Gold wirka Agau warka werka wark Tigre barubera Shangalla. Iron beresh » Georgian. ferrus Latin waratsch _ Kamschatkan waland Koriak This close connection between words for & white’, “ silver” and ** iro raises the question, whether the South Dravirian term for “iron” may not come under the same eategory. The full form karumbon appears to be compound, and if bon be the root for “ white” and silver” (comp. the forms bol Tuluva, pun Kol. white’) karwm is the Dravirian word for ‘ black”, i. e. the com- pound is black-silver” or,“ black-metal.” The Karnataka kahina has the slender form or the word for white and silver, with ETHNOLOGY OF THE INDO-PACIFIC ISLAND#. 161 kar contracted to ka. Kar-ba, irum-ba, irum-bu, inu-mu are all contracted variations, the root for “silver” preserving the broad northern forms ba, bu, mu. The allied Hausa term is formed in precisely the same mode. Ka-rufa, ka-rifa iron, aze-rufa, azu- rufa “silver.”* These terms are Scythico-Iranian,; “silver” rupe rupia Sanskrit, Bengali, Hindi &e; “ white”, accho Sindhi, as Ugrian, hais Saumali, hathi Galla, ht Egyptian, saisa, sai, sa Khomen, howse Pelew. Pashtu has the Hausa combination ash- repe silver.” In Turkish it is found applied to “gold,” asherafi. Suaheli has a similar @ form in rapia “silver.” The guttural root for “black” is Scythic, Iranian and African as well as Dravirian, e. g. kara Turkish, kala Sansk., kam Egyptian, ako Gonga. If the Hausa term be of eastern origin, the full form was probably kara-rufa or kar-rufa. Se. The Egyptian ht—of which the phonetic form is probably pre- served in the Galla hathi and the original in the Saumali hais— was the term for “silver” as well as for “white,” and the parent Ugrian root for “white” is also applied to “silver”, shie Wolga, osys, esys Perm, ezst Magyar (whence se Karen, hen Khyeng, son Mon). A common variation in the final consonant or definitive of the full form of the root, sys, brings us to our own English term which is an ancient Ugrian combination of the sibilant and of the almost universally diffused labial root, first examined above. ‘Silver’? is a similar compound to the Pashtu and Hausa term; ver is the N. E. Asian, Ugrian, Caucasian, Iranian, Semitico-African and Dravirian ver, ber, vel, wel &c; sil is one of the forms of the Ugrian sibilant term for “white,” sirr, siri Samoiede, siro Japan, sairan, sorny &c. Wog., shora Turkish, asido Abor, sudu Sifighalese, sed Hindi. The combination itself is Ugrian, serembire Samoide. (srebro Sclav., silber German, zilyer Dutch, silyer English, silba Fin). The Japanese siro-kane has the same word for white, prefixed to the Chinese word for silver (gan, gin &c). The secondary application of the sibilo-aspirate root to “iron "’ » ak “ oe +oK\ (OW as werent he es SE na Spkead beers, canresnonaiog. With, (ha Agen War. | Another common teria i Uitar, zinaria, dsinalia &e. lt appears to have been received from the Spaniards on the Lower Niger and thence spread to Mid-Africa (Howsa, Bornuike.) ©. 162 _ ETHNOLOGY OF THE INDO-PACIFIO ISLANDS. is now more common than the primary one to “ silver,”"—Chinese het, thiat, thi, Anam hat, Arabic hadid, (all close to the Egyptian and Galla hathi, ht, hais), Sanskrit ayas, Latin es, German eisen, Dutch ijzer, Samoiede yese, Korea soi, suy, Tibeto-Ultraindian ‘chya, sei, si, chur, sur &c. &c.; Asonesian sua, isu, hao; Tungusian sello, selle, zhilla &c. The Dravirian terms for “ silver’ and “iron” appear to be. at least equally archaic with the Semitico-African, and even with the Ugrian or proto-Scythic. That for “iron,” while Ugrian in the ideologic basis and in the separate roots, is native as a compound . With most of the preceding words it claims for the Dravirian- speaking race a civilisation of equal antiquity with the Semitico- African and Iranian, and one which in its earlier form was probably brought from Middle Asia with the language itself. GOLD. Of the Dravirian terms for gold, suvarnam Telug. is Sanskrit (Ugrian shiortno Wolga,*—the shorter form shor, sor, son, sir, ser &c. is much more common in Ugrian, Iranian &c. Euskarian urna, Latin aurum &c) Chinna Karnataka is an example of the shorter form of the root, Silong sin, It is connected more closely with the Sanskrit and Pali hirania, hirna, Pashtu sira-zar, Brahui zar, Ugrian sirne, &c. than with the Indian sona. The Tamil pun, Malayalam punnu, is the root for “white”, “ silver” again, in the Kol form. In Semitico-African and Malagasy-Asonesian the root is also applied to “gold” ‘as wellas “silyer”. To the Semi- tico-African and Drayirian terms previously given may now be added the common Indonesian term for “gold” vulanw, fulaan bulava, bulawa, bulau, bulana, all similar to Indonesian forms of the same root applied to “moon”, “silver”, “ white.” As the most simple and methodical mode of exhibiting the short glossary and its affinities as a whole, I have thrown the numerals and the 60 miscellaneous words into two comparative yocabularies;+ giving under each word a separate place to every * The root sor, sol, son &e, is applied in Scythico-Iranian vocabularies to the aarp Tope al V(A and B). I have ventured to inficate the postfix by italicising it in most cases, but it is probable that I have sometimes confounded a particle with the root and more frequently marked as a postfix what is really « portion of a dissyllabic or polysyllabic root. To the roots accurately ETHNOLOGY OF TRE INDO-PACIFICO LANDS. 163 root current for it in the Dravirian languages, and under each root the various forms it assumes, followed by an indication of the foreign affinities, so far as the vocabularies accessible to me have enabled me to trace them. In most cases these indications must be considered as suggestions rather than conclusions, mate- rials to aid research into the history of the several vocables and aot such history itself. Where the same root is widely spread in foreign vocabularies the affinity pointed out may be safely adopted as a real historical one, although the complex relationship may remain obscure, and in many cases may belong to the primary monosyllabic stage of the language. Where the particular form of the Dravirian vocable—in root elements, or in these and the attached definitive ov definitives,—is found in another vocabulary, a comparatively elese and direct connection is indicated. A double identity in root and definitive, in the structure of the compound, and in that phonetic form which is so liable to chance, —can only be accounted for, in general, by referring the vocables to one diffusive vocabulary, or to a common mother formation. Many of the affinitiesnoted must—from the absence of cumulative or corroborative indications—be considered, for the present, as merely phonetic. Further research will raise them to a glossarial or historical rank, or discard them as fortuitous. Although they possess no récoguteable value in the present paper, I have not considered it right to omit them, because data accessible to others, or which may be hereafter published, may proye them to be real. In the more exact consideration of the historical affinities — which forms the text of this chapter I have used the Appendix as a body of suggestions only, and have not thought it worth while to point out. ia detail where my present inferences differ from those indicated in the vocabularies, which were compiled long ago.* in all eae Bde Ey an intimate knowledge of the Dravirian ages an a full comparison of their vocabularies with uli the others in the Old World at lisa -c ns\derable time has also elupsed since they were printed, and the additional materials nuw wvailable, the imereared tucilit y and certainty with which serie intimate acquaintance with the structure of different families and a longer practi in the collation of vocabularies enables me to di-tinguish roots omy peices, pos meats fixes and infixes, andthe numerous imped tcotions I new find, w dispose me to cancel the whole heel if Mere was nny prospect, of finding time to compile them &iew, Y 164 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. is found. in the more widely prevalent mur, mar &e. (See JZead (c). | Hair (¢.) orama Kurgi* robong Gond Jom Bengali [uran Goold I. 4 kiaram Wiradurei (See Head c.) | é-nom-braem Waigiu nahm Vate niem Tana (head) langan Roti nit ge lungga Buol » lunggongo - Goront. _{rambut Indonesian (Mal. &c), but ram is probably a def. pref.; bu’, buk is a common root (See Head (d.) The liquid alone in various forms, la, lu, ulu, ra, ruh, ira &c is a common archaic root, N, E. Asian, Scythic, Indo-European, Asonesian. ; Hair (d.) yentruka Telugu Hair (e.) chutti Uraon choti Sindhi The root is common chu, su &e. Hair (f. ub Y) Kol uP » pu Tibet, Horpa, Takpa mui Manyak upat Ostiak opta w up, Ip Fin bo, mo Chinese - pipe _, Tasmania — ; * Comp. oluwa “head”, Singhal. lobu “forehead”, Fin, lob “forehead”, Sclavonic and the common Sclavonic term for “ head” golowa, glawa &c, Latin calva, Celtic gal. The guttural is probably the common Asiatic go, ka &o, head,” and go-lowa may be a compound of this root und foba or lowa, in which the root is the widely spread lu, ula, olo &c, and ba, wa the common Ugrian post- fix. But the root may be golo (Ugr. u- gol). ETMNOLOGY OF THE INDO-PACIFIC ISLANDS, 165 (bead) ap Egyptian n° api ” The ultimate root is the same as bu, pu &c, “ head,” and the Kol ub may be merely a variation of bu [See Head (d)]. With reference to the yocables given under Head c and d and Hair 7, it should be remarked that the forms in mt, ms, bs, &c, are more. commonly used for face, and eye than for head, The labial root by itself and with the same range of finals -f, -d, -n, -r, -l, -th, -8, «hy -g, is also applied to Face, Cheek, Mouth, Lip, Nose and Ear. Examples of the simple labial root, —Head, pa, awa Ugrian, ap Egypt, bu Kol, phu Mikir; Hair, bo, mo Chin., up, ip Ugrian, pu Tib, &e., ub, up Kol; Cheek, ma Anam, pa Burm., fi-fi Malag. (pi-pi, pa-pa &e., Ason.); Mouth, af Tigre, ma Shan- galla, va-va Malag. (fa-fa, vi-vi Xe. Ason.); Zip, mui Anam (bi-bi, wi-wi, ke. Asons) ; Nose, pi, phi &c. Chin, pui, pue Sam., uf Galla, mu Besisi; Zar, mi-mi Japan; ye, mey Japan, ma’ Chin., wa Dalla, me Tounghlu, mhe Bongju, mi Singtu. For Fuce Chinese has mien, min, men, bin, ben, (comp. Head, Turkish benys, Ost. wanim, Singfu man, Hind. mauh, Ugr. Celt. Eask. pen, Hind. Beng &e. man, mud, mur; Hair, min Chepang; Cheek ping Milchanang, mingmo Abor, minong Hailam, pinga Kayan, ping ping Iloko, weng, Kissa, fi-fi Malag., pipi Indon.; Mouth, mieng Anam, minoe Nicobar; Lip, bir Japan, pin-yaing Yenis., minu Nicob. &c.); Fin muoto, Japan omote, Ugr. wonda, Kashm. buth, Anam mat, Pol. mata (comp. Head mata, muda, &e. &c. common; Mouth, mhutu, musu, mocha &e. ; Nose, mondu Turk., munta Ho) ; Turkish bit, pit, mes &c., Simang mid, Anam mat, Celtic wis &c., Eusk. bisaja, (comp. Nose pi’ &c, Chin., pit Torres St., petyné Aino ; Lip pite Sam., mit Torres St., pedivi Telin- ga; Head bash, pus &c. Tark., apt Maram; Hair obit, upat Ugr. ipt, apt, opt Sam., mas Arm. &e.) ; Dravirian mukhu, muka, mokam &c. Indon.muka(comp. Chee bucca, baga, bhog &c.; Mouth mukya Pali, mocha Kol, bocca, foco, bouche &c. I[ndo-Eur., pak Siam ; Nose,muku Dray., muk Simang, bokan Woloff,; Zip, meka Bongju, amga Tungus., makub Tib.; Head, abak Pont. wokbok N. Aust. ; Hair bok &c, Ultraindo-Ason). For i outh Scythic has aman, amun-yak, Indo-Eur. mund, muat, mouth, mutte, munnur, &c.; Kol mocha, Sindhi wat, Kambojan 166 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, mat; maru Limbu, wullao Gond, mar Manipuri D., mamun Kap- wi, mieng Anam, abong Lepcha, pan Mon, ban Simang, peng Be- sisi, bango Lamp., Komr., abana Sumba, manga, mangai Pol. ; mothong Chepang, mlmta Newar, mathu Maram ; Asone-ian, mo- tong Meri, montong Banj. The Jabio-guttural form is also Scy- thic—amga, Tung.; Indo-Eur.,—Sanskrit, mukh, mukya, Bengali bak-tra, Latin faux (faucis), “the jaws,” bueca “ the hollow part of the cheek", Span. foco, Port. bocea, Fr. bouche ; and U)train- dian,—pak Khamti, pak-obu Kar., m’kha Kyau, awkang Silong. The Malagasy mulu, muluts, mulu-bnra &e. (and Asonesian mu- Int) may either have the labial root orthe liquid. In the lafter ease it is probably a derivative from the Zimbian mlumu, umlumo, mulumo &c, The! root is Seythic ul, lu) &e. The Zimbian lumu resembles the Asonesian lawe, lama lida, Kumi Ibaung, Limbu Ieba, Abor nepang, Sansk. lapanam, The same root is used for Tip,—Galla lufluf, Kosah lebi, Hind, lab, Lat. labium, labrum, Germ. lippe, Eng. lip &e, The ultimate labial root is used for Mouth in Seythic,—am, im, um, &c. Ugr., ama Mong. ; Semi- tico-Libyan,—ma_ Shangalla, afa Dankali, af Tigre, of Saumali; Ultraindian,—ba Kayan. The duplicated labial root is common, ——mefo Samoiede, momo Sunhe li, vava, vave, Malagasy (in Aso- nesia fafa, haba, bibi, vivi, bafa, wuwa, buwah, bua &e. &e.) The Dravirian bayi, vaya, bai, appears to be connected bi the Ugrian radical. forms. Most of the other preceding terms are also applied to Lip, Cheek, Nose, Eye in different vocabularies, Thus for Lip Seythic has amun, mon, emga, amga, pite; Dravirian pedivi, Ernb mit, Anam, mui, dapan bir, Indonesian bibir, bibi, wiwi &e., Australian mundn, mudol, wiling, Malagasy wat Talis &e,,, Nieobar ; minu, manoey; and for Cheek, Malagasy has fi-fi (Asoues. pipi, papa Ke.) ; Burman pa, Anam ma, Latin bucea, Galla bako, Kaili baga, Erub bag, Latin mula, Indones pili, plis, paling, banga, bongi, pingi &c. For Nose phi, piti, pit &e. and ma, mui, burn, mura mondu, marh, muku, &¢c, are common. The Chinese phi, pi, &c. corresponds with the Samoiede pile lip, Aino prtyni nose, and the Torres St. pite, piti, pichi nose, and mit lip. The Samoiede pui, pue, puiya, piva &c. corresponds with the Anam mni, Nancowry moi, nose. The whole series is reproduced in the vocabulary for Hye, 1, ETHNOLOGY OF THE INDO-PACIFIO ISLANDS. 159 Sec, 12. RECAPITULATION AND 'NPERENCEs.® Ix our present enquiries we cannot go back to the period when there were no languages in India and the adjacent countries, or when some of the present great formations had not yet come into existence. We must reason on the phenomena which Southern Asia has presented to human observation since any permanent records of it began to be kept. As far as observation can carry us into the past, this region has always presented several races and formations as at present, and tribes and languages belonging to different races and formations have always been more or less intermixed and subject to change from mutual influence. In those human eras into which ethnic research has hitherto extended, South West Asia and Asonesia, considered as one continuous province, have been contemporaneously occupied by, Ist, archaic Indo-Australian, 2nd, Papuan, 3rd, Tibeto-Chinese or Ultrain- dian, 4th, Dravirian, 5th, Scythic, 6th, Iranian, and 7th, Semi- tic races and formations. In all historical times we find several of these intermixed in the same territory and influencing each other. We also find that at differeut historical eras each of the three last has become expansive or migratory. Irania from very remote ante- historic ages appears to have been occupied by these three races, at an earlier period by the 4th also, and probably ata still earlier by a race akin to the Ist. Hence in later eras each of the three last must always have been more or less subject to mutual in- fluence. In the same manner the peoples and languages of India must have been exposed, throughout these eras, to the influence, in different degrees, of the three races of Irania or of the predomi- nant one. In great periods of archaic time the language and race of the most dominant or diffusive people of Irania and India pro- bably varied, as it has done in historic eras. Nor, in our endea- vours to obtain some firm footing in the archaic world, must we overlook the mere possibilities arising out of the distribution and character of the great races. Scythic, Semitic, Iranian, Semitico- Iranian, Seythico-Semitic, Scythico-[ranian or other mixed forma- tions like the modern Indian, may have successively prevailed in _" See Sec, 10 for summary of the eomparative structural characters of Dra- virian. 160 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Irania. There may have been Semitic or Iranian tribes speaking Scythic dialects or Scythic tribes speaking Iranian or Semitic dialects, and each influencing the ethnology of India. This penin- salar region being open on the Tranian side, it is probable that it, also, in all Jater eras, has been occupied by more than one race and linguistic formation. | So far as we know, there never was a period when any one of the great formations existed in S, W. Asia in a completely isolated position. Each, so far as we can trace it, has always been surround- ed by other formations. In every considerable ethnic revolution and movement of archaic times, as in the Brahminic, Medo-Persian, Seythic and Arabian conquests of historical times, tribes of distinct races must have come in contact, one race predominating or at least maintaining its position in the lands of others by its superior power. Wherever the nature of the country caused actual contact and in- termixiure, assimilation must have begun. Onerace might change its language sooner than its physical character, or vice versa. In mountainous countries and wide steppes, isolated or nomadic tribes under favorable cireumstances would retain their native formation, even when subject to a foreign race. Hence immediately to the north of Irania there have probably always been wandering Scythic tribes in the later eras of human history, although their territories have been embraced in Semitié or Arian dominions and even been contemporaneously occupied by an Arian or Semitico- Arian people. Bat in fertile river basins inhabited by fixed indus- trial communities, an instrusive dominant people cannot remain pure, much less can the native and the introduced linguistic forma- tions be preserved unmodified. Wherever, in the ethnic revolu- tions of Irania and India, two races and formations have come permanently in contact under such circumstances, mixed tribes and dialects must have resulted. The connected province formed by the basins of the Indus and Ganges must have been the seat of settled and civilised populations from the time when agriculture and villages first existed in Trania and India, and it is probable, therefore, from the natural attractiveness of a large portion of that province, from its enervating and demoralising influence on its successive occupants, and from the permanent existence in the countries to the N, W. of more robust nations, that the formation ETHNOLOGY OF THE INDO-PACIFIO ISLANDS. 161 of hybrid races and languages has been a standing characteristic of its ethnology. The same remark is applicable to the more open and fertile tracts of Southern India. Grant that fixed industrial populatiors existed in these countries prior to the later movements of western races into India, and the gradual modification and even transformation of the principal Indian languages is a necessary con- sequence. Glossarial facts prove that the Indian tribes were set- tled and civilised prior to the Arian era, and as the pre-Arian arts were derived from different sources, and indicate the lapse ofa long period of civilisation and of intercourse with foreign races, there was room for a repeated production of hybrid formations before the Indian languages acquired the forms which they now have, and which, in their turn, will"prove the foundations of new formations, if they are not entirely replaced by foreign ones. The relation of the Dravirian physical and linguistic formations to these of the provinces around India is the first point to be considered in an attempt to ascertain their true ethnic affinities. The Chinese, Siamese and Mon-Anam nations differ essentially from’ the Dravirians in person, in language and in other respects. The North Ultraindians and the Tibetans aye very remotely con- nected. with them. Physically, both are purely Turanian and their languages, althoagh of a similar fundamental type, are ata great distance from the Dravirian both in ideologic development and in phonology. The phonctic difference is so great as of itself to prove that the Dravirian formation was not derived from the countries adjoining the Indian peninsula on the east and north while these were oceupied by the Tibeto-Ultraindian. It is also improbable that it was derived from Upper Asia through Tibet and the Himalayas, because there are no grounds for supposing that the Tibeto-Chinese race are not the oldest occupants of these countries, and any ethnic movement on so great a scale and so prolonged, as to diffuse a harmonic phonology like the Dravirian or Draviro-Australian over that barrier region and thence over India, would have. left traces of its presence distinguishable from those which mark the comparatively modern intrusion of Scythic languages. The affinities between Draviro-Australian and Tibeto- Uliraindian, considerable and fundamental as they are, appear to -be referable to a stage of the former long preceding its harmonic 162 ETHNOLOGY OF THE INDO-FACIFIC ISLANDS. development and its spread to India, and to be only less archaic than those with Chinese. The physical and mental characters of the Chino-Tibetan races who have immemorially and aboriginally— as far as that term may be applied to the human tribes of any region —oceupied the lands that bound the plains of the Indus and tha Ganges on the north and east, forbidding us to seek further in these directions for the fount of the Draviro-Australian alliance, and its various linguistic developments beiny far advanced beyond the Tibetan, Chinese and Mon-Anam, and in a direction similar to that of the great harmonic alliance of Asia, we must look for the immediate source of the formation to the basin of the Indus. This province is chiefly connected with 8. W. Asia in two direc- tions,—in a northern, through the head of the basin in Balti and the Hindu Kush, and in a western, where it is conterminous with Affehanistan and Belochistan. The Dravirian formation, accord- ing to every ethnic probability, must originally have been an exten- sion of a similar one that preyailed in this region, or at least some of its principal and distinctive elemenis must have been derived from a formation so located. There are several objections to our considering the head of the Indus as the main direction in which the Dravirian formation was spread to the south and east. It is quite possible and even probable that Balti was not Tibetanised until a comparatively recent period, and the previous population, or rather the pre-Arian, may have been an extension of the adjacent Scythic race, tethenorthward. But this race, in all its Mid-Asiatic varieties, speaks purely Scythic languages and such languagescould not have originated the Dravirian. They might certainly have sup- plied one fundamental ingredient, but some of the non-Scythic cha- racters repel us from attempting to trace the history of the forma- tion exclusively in the great Scythic field, and direct us to the western province between the Persian Gulf and India, which, in a wide sense, may be termed Irania, for there is no distinct geogra- phical or ethnic division between the eastern and western portions. In this province and that immediately to the north of it as far as Transoxiana, two races and two linguistic formations have prevailed from remote antiquity,—the Iranian and the Seythic; but a third race, the Semitic, immemorially located on the western confines of the province, has also, both in archaic and historical times, fTHNOLOGY OF THE INDO-PACTFIC ISLANDS. 163 exercised a @teat étlinic influence in it, while a fourth, with claims to at léast an equally ancient decupation of the N. W. mountain boundary of the Caucasus, has intimate linguistic affinities with all these forriations, . In later historical times the Sey thie tace as chiefly predomi- nated in the north and occasionally in some portions of Trania ilso, The present Seythic tribes appear to belong mainly to the great hordes of Tartar invaders—Turks and Mongols—who, i in comparatively recent ages, have occupied the region between China ahd the Caspian, intruding into Tibet and Irania, but théir numbers and the extent and duration of their Indian domination were not such as to produce a marked impression on the Dravirian languages. In earlier historical times the Iranian race, civilisation and linguistic formation appear to have been exclusively predomi- vant over Trania, and this supremacy must have endared for a considerable period, because it embraced an unbroken belt from: the Black Sea to the mouths of the Ganges; while its spread over Enrope is an additional evidence of its having, for the time, prevailed over the Scythic | or Turanian hordes and thrown them back on Upper Asia, To this race the present Arian and Arian- ised nations of India, the. _Affghans, the Beluchis, and the wide spread Persians or Tajiks. mainly belong, although a Semitic element is found in most. The history of the race in its Irano-Gangetic province evidently involves at least two great diffusions. Of the oldest the languages and nations of India preserve the only distinct record, with the exception of the Sia Posh. From the position and character of the latter and the general distribution of the Indo-European form- ation, it is probable that the Arian sub-formation preceded the Persian in Eastern Irania, and consequently that dialects akin to the Sanskrit prevailed there at one era contemporancously with the older languages of the land. The Arian formation partially transformed the phonotic and idealogic character of the prior Dravirian languages of northern India and displaced the greater portion of their vocabularies, pro- ducing the present hybrid tongues from Guzerathi on the west to Bengali on the east. Its influence on the Vindyan and Southern branches began later, and although it has been continued since p+ ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Sanskrit ceased to be spoken, it has only very slightly affected their phonology and ideology ; bat its glossarial action has been considerable. At a period subsequent to the advance of the Arian tribes from Irania into India, another branch’ of the same race appears to have been modified both physically and in language, i institutions, reli- gion and the general character of i its civilisation, chiefly i in remote ages through the influence of the great Semitic nations of the Euphrates but also in later ages through the eastern spread of the Arabs, This branch was, ‘altimately diffused over all Irania and the Turanian countries adjoining it on the north. As far as the Indus the Semitic physical type, ‘and the Zendic or Persian linguistic form, are strongly marked. Through the widely spoken Hindustani the latter has, in modern ages, gained a considerable glossarial and phonotic diffusion in India. In the Zend phonology the Semitico- African element is strongly marked, and this is one of its most fundamental peculiarities when compared with Sanskrit. This sub-formation does not appear to have inflaenced Dravirian, The physical character of many of the Dravirian tribes and castes, and perhaps some of the traits of the language, point, to a still more archaic diffusion of the Semitico-African element to the eastward. The modern or western Iranian idiom has also become that of some Scythic tribes of eastern Irania. But there is strong evidence that prior to the great eastern advance of the Indo-Germanic race, large portions of Trania were occupied by dominant Seythic tribes, The Tranian languages themselves, in pho- nology, ideology and glossaries shew, when they are compared with the only other formation spoken by a kindred race, the Semitic, that the Scythic formation, or formations akin to it, had been extended " into ‘Irania at a period coeval with the development of the Iraniar linguistic type itself. It is possible that some of the northern Scythic tribes of Affghanistan are pre-Iranian, and there seems no room to doubt that one of the southern, the Brahui, is a genuine representative of the pre-Arian population of 8. E. Irania or Beluchistan, as the Jats appear to be of the lower Indus. The Brahui physical type is Scythic, and the language has some strong Dravirian affinities in glossary although it is pro- bable that the grammar has become Iranised, The other voca- ETHNOLOGY OF THE INDO-PAGIFIC ISLANDS. 165 lbularies of eastern Trania and the Indus,rwhether spoken by Semitico-Iranian, Indian or Seythie tribes, have also a consi- detable number of non-Iranian vocables that are Dravirian and in many cases Seythic or North Asiatic also. The Dravirian forms sometimes resemble those ‘of the Indus and east Iranian glossaries more closely tlian the Seythic, From this it may be inferred that they were received into India through the pre-Arian languages of the ‘Indus, because the advance of the Indo-European race into Eastern Trania and India must have cut off the further diffusion of the native vocabularies to the eastward, and arrested the. regular flow of Scythic words into Trania and thence into India. We may conclude thatthe Seythic element of the ancient Indian tribes and languages was immediately received from eastern Irania at a period when it was mainly Scythie. But» the Semitico-African element both in Dravirian and Iranian re- quires us to believe either that the pre-Arian Seythicism of this province was tixed with Semitico-African ingredients, or that it was, in its turn, preceded by formations of a more archaic charac- ter, having fundamental affinities both with Scythic and Africo- Semitic. In pre-Arian India the Africo-Semitic physical element must be the most ancient, because it is chiefly marked in some of the most southerly tribes and is found also im, Australia and amongst the Papuans. The more decided or pure Seythic charac- ter of the Brahui and several of the ancient Indian tribes of the Dravirian formation must be referred to a later era when the Neythic race prevailed in eastern Irania. The partially Afiico- Semitic basis of the Dravirian race and languages and of the Australian must belong to an east Iranian formation prior to that represented by the Brahui. While therefore the latter affords one strong reason for believing that the more recent and predominant Scythic element of the Dravirian tribes and languages was derived from the west and not from the north, the commencement of the Semitico-African formation immediately ewan Beluchistan and the immemorial existence of the Iranian in Irania, justify the conclu- sion that those typical physical and linguistic traits of the Dravi- rians Which are not Scythic but rather African, Semitic or Iranian were received at a still earlier period from the same province. The Drayiro-Australian alliance, when considered in its pre- 166 ETHNOLOGY OF THE INDO-FACIFIC ISLANDS. ? Arian condition, ‘differs so much in its several developments, and there is so great a break between the Indian and the Asonesian forme, that we cari onl UR consider the Ariin as one of the latest and most partial of the intrusive elements that have modified the Indian branch. _Between the Australian condition and the proper Draviriaa, or that cle igniediate preceded the Arian, the interval, whether ay by physical, linguistic or mental and industrial ¢ha nge,-is very-great, and its Indian history must have been je ol It, probably began with negro tribes and proto- Scythic languages like . the human histories of Asonesia and Africa, while its later eras were marked by the predominance of advanced Scythic, Semitic and Semitico-Scythic races, and by the influence of Semitic and Scythie languages. The great and archaic Scythic movements that appear to have preceded the proper Se- mitic in S.W. Asia, and are so deeply impressed on the Caucasian and African languages, were felt in India also and through it in Asonesia. The early Caucaso-Semitic movements which proceeded the historical Semitic, and must have been associated with the civi- lisations out of which the Egyptian, the Babylonian and the Pho- nician grew, bave left their impress on the Dravirian languages as well as on the Nilotie and North African, and the partial approximation of the Dravirian physical type to the Semitic, with the civilisation the Indian nations had attained prior to the Arian era, need not be sought in any remoter Cause. There is no reason to suppose that the influence of the Semitic race and civilisation on the Dravirian has ever been wholly interrupted since it first began. When the Arians broke through the connection which in all probability previously existed by land, it is not likely that the maritime intercourse between the Semitic and the Dravirian ports was interfered with. The Dravirian formation is so archaic that not only all the great historical ethnic developments of 8.W. Asia, but the first rise of the Semitic power and @ivilisation, and all the later movements and revolutions of this region, including the Indo- European, mtist have taken place in its presence. Its history goes hack beyond the beginning of the civilisation of the Euphrates and the Nile, and much that distinguishes the Dravirians from the Australias may associate itself with the most archaic and as vet ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 167 undefined perivds in the gradual progress of the Semitic, Cauca- sian and Iranian tribes from a barbarism more than African. The general character of the Indo-Australian formation proves that in the most archaic era to which the positiye ethnology of 8. W. Asia can as yet ascend, the Seythic linguistic element predomi- nated, in Eastern Irania and India, But whether a Scythic or an’ A frico-Semitic race and formation: was the oldest of all, or which was the older of the two, in this region, is not clear. The early exten- sion of the Semitico-Liby an or Libyan formation over the great outlying region of Africa, its undoubted Asiatic derivation as evineed by its Caucaso-Scythic affinities, its fundamental proto- Seythie traits, the character of the purer African physical type, and the presence of a similar clement both linguistic and physical in the Draviro-Australian family, render it probable that Libyan races aud languages long preceded the Draviro-Australian in &. W. Asia, and mixed with the intruding and dominant proto- Scythians who introduced that formation. The Egyptian stage of the Semitico-Libyan formation is cruder than the Australian stage of Draviro-Australian, .It is nearer the Tibeto-Ultraindian and Chinese developments. Australian has much of the advanced proto-Seythic development which predominates in the American, the Zimbian, and the Euskarian formations, and is only less pro- minent or more modified in the Indo-European, Scythic and Cau- casian. Egyptian shows that the Semitico-Libyan mother forma- tion had separated from the great trans-Cliinese stock of Asia prior to the attainment by the latter of a highly agglomerative and harmonic phonology. It spread to the south west, took possession of Africa and long remained faithful to the archaic West Asiatic type, while in Upper Asia that type changed, and gave rise to var:ous higher phonetic formutions, including the early Indo- Australian. That formation stands in its origin at a great distance behind the Indo-Eurepean and even the Ugrian, but the early Semitico-Libyan goes back for its origin or point of divarication to an era far beyond the Indo-Australian. The latter distinctly associates itself by its phonology and structure with an archaic condition of the Scythic development, Semitico-Libyan with a condition of the Mid-Asian development between Chinese and Seythie. In this early or Seythico-Libyan stage itis probable 168 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. that languages of the oldest Libyan type were not contined to the S. W. extremity of Asia and to Africa, but extended eastward along the northern shores of the Indian Ocean, and may have preceded the Draviro-Australian on some portion of the line along which it advanced to Indin. Be this as it may, the history of the Dravirian linguistic forma- tion is far from being fully elucidated by a comparison of it with the other formations of 8. W. Asia,—Tatar, Tranian, Semitic and Caucasian. It is not closely related to any of these, and its more fundamental affinities with them, large as they are, go back for their sources to older developments, embracing a still wider range of formations. The individuality of the Dravirian formation, the impossibility of subordinating it to any of the 8S. W. Asian forma- tions, and its great antiquity, are illustrated by the fact of its archaic prevalence in a cruder condition in Asoncsia, When the characters of the present predominant formations of Ultraindia and of all Asonesia save Australia are considered, Dravirian appears to stand out from the Iranian and the Tatar as an older 8. W. Asian formation, which has survived great changes in the distribu- tion of races in Southern Asia, and which*hy the ernde form it retains in Australia, proves that the more Iranian and Seythiec character it has received in India was superinduced on a native basis of independent origin The earlier S. W. Asian history of Dravirian, when thus viewed as a prior formation to Tranian and Scythiec in Irania and India, is hardly capable of being traced, because there no longer remains any formation which can be con- sidered as the ultimate or native one and as the limit of our researches in this region. We can ascertain affinities with other and more distant formations, but these will not supply us with all the elements of the ancient linguistic history of the Irano-Indian., When the actual barrier languages on the west are removed, we no longer have any clear guide to the archaic limits or movements of Dravirian. It may have been developed in Irania or India from a type still cruder than the Australian, or, as is more proba- ble, it may bave been derived in its Australoid type from a distant land of origin. When we go beyond the Tatar and Iranian and come to the allied Ugrian languages on the north and east, and to ETHNOLOGY OF THE INDO-PACIFIC IsLANDs. 169 the Caucasian on the west, we find strong Dravirian affinities, and it has others with the N. E, Asian languages and even with American which appear to belong to a proto-Scythic development. The languages of China and Tibet on the one side and those of Eeypt and Africa generally on the other, show that the interme- diate region must have undergone great linguistie changes before even the earliest variety of Indo-Anstralian was introduced or formed. From Tibet and Egyptian—the salient members of the old formations on the two sides of the Irano-Semitic region—to Draviro-Australian, the phonetic advance alone is so great that it necessarily implies a succession of formations, although it does not follow that they were developed in this proyince. The Draviro-Australian phonology is archaic Scythico-African and not proper Seythic, Semitic, Iranian or Caucasian. The ideology is mainly Scythic of a very arehaic character, or rather proto- Seythic, for it is not merely a branch of the Tatar or even of the Ugrian. The connection is through an older and more Americo- African, Caucaso-Semitic, aud Iranian form of the inversive development, and through that form in one of its early and crude stages. Iranian im its basis is more closely akin in some respects to Ugrian than Dravirian, the pronouns for example being the same. Dravirian again has special Caucaso-Semitic and Caucaso- African affinities. The conclusion appears to be that it was a form of the proto-Scythic or harmonic and inversive development that preceded not only the Tatar but the Iranian and the allied Ugroid Scythic in Irania, and from its archaic character and early migration to the south west of Asia and thence to India and Asonesia, had independent relations with the Caucasian, the Semitic and the proto-Iranian on the one side, and with the proto- Scythic languages of Middle and Northern Asia on the other, In one point of view it is the oldest and earliest formation of the Scythic class that is now extent, its position in Asia and Asonesia combining with its general character to prove this. In another point of view it is a distinct and more ancient babi hk but of the same development. The glossaries by themselves afford considerable evidence that the Seythic or proto-Seythic formations that prevailed in S, W. Asia, spread into Africa and India, and affeeted the vocabularies 170 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. at least of the Caucasian and Iranian languages, long before the Turks and Mongols advanced from the remote east, were allied to the Ugrian, Samoiede, and Yeniseian. If the Caucasian and Ugrian vocabularies yield strong evidence of the two formations having been not only archaically connected buat in contact in periods long subsequent to their first development, the Dravirian vocabularies preserve proofs of a similar connection and contact with the Caucasian and the Ugrian, while their direct Chinese and N. E. Asiatic affinities point to a more eastern mother-land than Irania or any other portion of S. W. Asia, If the evidence of language may be trusted, the Dravirians were one of the oldest nomadic races who advanced from Upper Asia to Lrania and India. The character and position of the Semitic-Libyan formation and of the African tribes renders it probable that the Draviro-Australians found formations of this kind established in S. W. Asia, anil that by mixture with them the Semitico-Libyan traits of Draviro-Aus- tralian were acquired. The pre-historic revolutions, combinations and amalgamations amongst the nomadie hordes of Asia, probably present too complicated a subject to be unrayelled by ethnology. The languages of India have affinities not only with all the Tura- nian formations, but with the Iranian, the Afrieo-Semitic, the Tibetan and the Ultraindian. To read the ethnic history of India we must first decipher that of Asia and Africa in its leading inci- dents, for the Draviro-Australian formation strikes its roors into the Chinese even more déeply in some diveetions than the Scythic . languages. All attempts to trace the Dravirian formation to its ultimate sources must be illusory, because its antiquity is ob- viously so great that from the time it existed in its earliest development to the era when it assumed the form it now has in the principal languages of the South, there must have been a gradual extinction of many cis-Indus dialects and languages in which suceessive varieties of the formation were evolved, and of many trans-Indus ones which illustrated the formation in its pre-Indian history and development, or were instrumen- tal in producing changes in it subsequently. Between it and ill the adjacent formations there is a great break and even the chain of connection with Scythic wants many links. On the whole, we must be satisfied with the conculsion that, strong ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 171 as its foreign affinties of all kinds are, the Dravirian forma- tion cannot be considered as a branch of any existing Asia- tie or African one. It stands by itself like the adjacent Ira- nian system, and represents the most ancient form of language which can be recognized in India. Its extreme antiquity in this province is proved not only by the nature of its affinities to other languages of Asia and Africa, but in a still more striking manner by those with the most archaic formation of the Indo-Pacific ’ islands, now best preserved in Australia, although even there greatly modified. The prevalent formation of Southern India is probably that modification of the Indo-Australian type which characterised the language of the most civilised and powerful © nation of India in the era anterior to the intrusion of the Arians and also, in all likelihood, to that of the Turanian tribes of Ultraindia and Tibet. The basis of the Australian is probably one of many varieties of the same formation which were formed at a much ear- lier period when the Indo-Australian race spead over India, Ultraindia and Asonesia. It may be concluded from the facts mentioned in Chap, II that the Australians have, in a great degree, retained the physical characters of this race, and the barbarism which still distinguishes many other insular tribes, the Simang of the Malay Peninsula, the Andaman islanders and some of the more sequestered tribes and degraded castes of India (including Ceylon) can leave little room for hesitation in adopting the opinion that the Ultraindian and Indian race, whose migrations gave the earliest known population to the eastern islands, had not advanced beyond the Australian grade of culture when these migrations commenced. It may be doubted whether the Celtic or earlicr diffusive branch of the Iranian stem had itself attained a higher grade when its western movement began. Those tribes who were most remote from the later Semitico-African sources of civilisation, such as the insular Britons, continued to the age of the Roman invasion in a state of barbarism in some respects more degraded than the Australian, or the lowest Dravirian or African. If the Dravirian formation prevailed in India at a period when its tribes were similar to the Australian in character and civilisation, it must have been a widely diffasive one before the rise even of the Tatar nations. This is consistent with the relations of Dravirian to the Scythic 172 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. and other harmonie Asiatie formations. In phonology, ideology and roots it is more closely connected with Seythic than any other formation, but as it has affinities with remote N. E, Asiatic and with African languages, it is associated with a period of the Seythie development in which the proper Tatar hordes had not yet overspread middle Asia, and when the general civilisation of the world was perhaps not above the Australian level. Tn some respects the Indo-Australian formation, although more akin to the Seythie thin to any other, goes back to an era prior even to the events in which the present Scythic development origi- nated. It connects itself also by some fundamental traits, as well as glossarially, with the non-Seythic alliances of S. W. Asia, particularly with the Caucaso-Semitic. The pure phonology is more “harsh and less harmonic than the Scythic and African. The distinction of sex in the pronouns is Iranian and Semitico-Libyan, and the pronominal roots themsclves have not been derived from a distinctively Scythic source. If the immediate mother tongues both of the Indo-European and the Dravirian formations origi- nated in Irania, the latter should naturally have more intimate affinities, whatever their age may be, with the former, than the remoter Tartarian languages, in other words it should be the most Arian of the inversive languages. In examining this point it must be borne in mind that the Southern being the most distant of the Iudian languages from Irania, would probably possess fewer traits in common with the proper or later Arian tongues than those of N. W. India. ; It may be concluded that the stock from which the ancient Indo-Australian tribes were derived was not Scythic, at least in the eurrent sense of that term. All the ethnic facts favour the opinion that the race was, in its era, an influential and diffusive one of S. W. Asia, whence it extended itself into India as the Arians did at a later period in the history of the world. It is possible that the protoplastic Indo-Australian race is older in India than the linguistic formation, but there is no reason to think that this is the case. At all events it seems hopeless to attempt to grope our way back beyond the Australian era. The basis of all histo- rical inference must be the fact that the oldest race and linguistic ETHNOLOGY OF THE [NDO-PACIFIC ISLANDS. 763 formation of India were akin to the Australian. We must conclude that before the rise of the historical Seythic nations a similar formation was predominant over a large province in Asia, that it extended to the shores of the Indian Ocean, and while located in S.W. Asia was there modified by contemporaneous formations of an archaic Caucaso-Iranian and Caucaso-African character. The Scythico-Semitie character of the Indo-Australian formation is consistent with that of the race itself, which is at once Africo-Semitic and Mongoloid but the former more than the latter. The physical type became in large degree sonthern although the language retained a northern form. Tow sucha change can be produced is explained by many examples. In much later times the Turks, a pure Seythie race when they enter- ed the 8, W. province of Asia, have acquired an Irano-Semitic physical type while retaining a Scythic language. That sueces- sive families of the Draviro-Australian race were also dominant before the rise of civilised Caucasian, Semitic and Indo-European tribes in 8S. W. Asia, is proved by the very fact of their having been able to migrate over the extensive and continuous Draviro- Australian region and from a more northern land without being cut off or absorbed on the way by more power‘/ul tribes. When such a race sent forth the great swarms of men by which India and Asonesia were first people, the progenitors of the existing 8. W. Asian nations must have been inferior to them in power. The Turanian movements to the southwafd, which commenced go early as to moilify the languages of the oldest known tribes of Africa, India and Asonesia, must have continued as the general civilisa- tion of Asia increased, The Dravirian languages when compared with the Australian afford strong evidence of this in the numerous terms of Asiatic civilisation which they possess in common with Scythic tongues and which are absent in the Australian vocabula- ries. The subject is elucidated in another place, but it should be remarked here that the Dravirian affinities with the Asonesian lan- guages are not confined to the Australian, although the most archaic and fundamental are chiefly found iu the latter. Itshould also be observed that the Australian and other Asonesian affinities of the ancient Indian languages extend to all those that still remain, and are not confined to the South Indian. On the contrary, they ” 174 ETHNOLOGY OF TUR INDO-PACIFIC ISLANDS, appear to have been mainly with the ancient Gangetic languages, and even those Asonesian vocables that are now found in South India only, were probably derived from Gangetic vocabularies which have since lost these words or have themselves ceased to be spoken. The Asonesian vovabularies also contain numerous words of a similar phonetic structure to the ancient Indian, but which have no representatives in any known Indian language now existing, although they have Scythic, N. Asian, Iranian, Cauca- sian or Semitico-African affinities. Allowing for those that may have been received directly from the Malagasy and Hast African formations and from Japan, the great mass doubtless found their way to the islands through the basin of the Ganges and Ultraindia for their diffusion in the most ancient ihsular vocabularies, inclu- ding the Australian, must have long preceded the era of a direct navigation between Southern India and Ultraindia or Indonesia. Of those chiefly found in the vocabularies of the more civilised and maritime tribes of Asonesia or within their range of locomo- tion, a large nunber were probably derived from India in the era immediately preceding the Arian, when the civilisation and mari- time skill and enterprize of the leading Indian nations appear to have attained a high grade, and when their ‘boats became the models of the Ultraindian and Malayu-Polynesian. From the Australian era of Indian ethnology to that which immediately preceded the advance of the Arian race beyond the Indus, there must have been a great lapse of time. Of this we have some measure in the changes which had taken place in the Indo-Australian region. In Asonesia the Papuan race and formation. had spread oyer the islands, obliterating or modifying the ancient tribes and languages. In India the leading Dravirian tribes had probably been already improved physically by mixture with immigrants of Seythic and Semitico-Iranian race. Their civilisation and languages had certainly been deeply modified by foreign influence. Making every allowance for what the Austra- lians and other eastern tribes may have lost when they left the continent and became insular, it is probable that most of the arts for which the Dravirians have non-Sanskritic names were acquired . by the race subsequent to the Australian era. Many of these ETHNOLOGY OF THE INDO-PACIFIC IsLANDs. 173 names have Scythic, Iranian, Semitic, Caucasian and African affinities, and it may be concluded that the civilisation of the principal Dravirian nations was mainly deiived from foreign immigrant tribes, settlers and traders who entered India from the North West or visited its coasts from the northern and western ports of the Indian Ocean. The principal nations of the South are so closely connect d in person, arts and language, that we cannot refuse to recognize in them the influence of one dominant and civilised people which at a remote period raised itself above the level of the barbarous tribes of India, and then spread itself by destroying, breaking up or transforming a large number of these throughout the more open country, as the Arian race after- wards did in the basin of the Indies and Ganges. The difference in physical characters between the higher classes of these nations and someof the lower castes antl hill tribes, is so great as to indicatea larze influx of a foreign people, and it is possible that the higher civili- sution originated in a race of conquerors who were not sufficiently numerous to maintain their own language. Whatever nations,— Scythic, Iranian or Semitic—preceded the proper Brahminic Ariana in the N. W. of India and the adjacent countries beyond it, must have influenced the principal or more civilised and exposed Dravirians. Such influences operate, and must have operated in all ages, wherever human races differing in power or civilisation come in contact, and the tribes of India have necessarily been al- ways in immediate contact with tribes belonging to the races that predominated in succession to the westward of the Indas. The glossarial affinities with the Pashtu, Pashai, Brabui and other N. W. languages, although pre-Sanskritic, may thus be comparatively modern. They tend to shew that the East Iranian and North In- dian glossaries were connected with the South Indian prior to the diffusion of the Brahminte formation and Sanskritic vocables into India, and they thus help to strengthen the other reasons for suppos- ing that the grammars also were akin to the Dravirian and Scythic _before they were modified by the Arian. The next great revolu- tion in Asonesian ethnology after the Papuan, serves also to illustrate the history of the Dravyirian in the era which immediate- ly preceded that of Brahminie predominance, and was probably 176 RTHNOLOGY OF THE INDO-PACIFIC ISLANDS. marked by the influx of earlier tribes of the samerace. The great southern movement of the Chino-Tibetan race which gave a Malayu-Polynesian population to Asonesia and a Chino-Ultrain- dian and Tibeto-Ultraindian to the t:ans-Gangetic peninsula, affeet- ed not only the middle and north Gangetic race but the Vindyans also. The influx of this race from the east and of pre-Brahminic Arians or allied tribes from the west was probably contemporane- ous at an early period, originating the mixed type which predomi- nated in the ancient Niha-Polynesian branch of the Gangetico- Ultraindian Asonesians. The pre-Braliminic Arian influence wis probably sufficiently powerful and long continued to have produced an Irano-Mongolian type, prior to the proper Arian era of Noithern India. It is even probable that the Dravirian nations of the Ganges, like the more civilised ones of the south, were greutly modified by archaic Iranian influence before the Ultraindians entered the basin. - Although we have found it impossible to trace the actual histo- ry of the Dravirian formation, we have ascertained the main course of its development and vurious points of contact, at its successive stages, with other existing formations. The general conclusions may be recapitulated as follows ;* 1. The general character of its harmonic, aspirate and liquid pho- nology is Scythic, but it has peculiarities in its strong and complex sounds. Save in some of the emasculated tongues it has a more harsh and primitive character than the Scythic phonologies. 2. The structural phonology is agglomerative and harmonic, It separates the formation not only from the Chinese and Mon- Anam but from the Tibeto-Ultraindian, and allies it with all the harmonic formations. In its specific characters—a weakness of the agglutinative, elliptic and amalyamative power and consequent rarity of flexions—it is much nearer akin to Seythic than to the pre- vailing S. W. Asian and African formations and to the pre-Seythic S. European [Euskarian}. Its agglomerative power is similar to the Scythic, and is hence greater than Caucasian and Semitico- Libyan, but less than the Iranian, Zimbian and Malagasy and greatly inferior to the American. But in its archaic Australoid * See Sec. 10 for review of the ideo!ogic and phonetic affinitics. ETHNOLOGY OF THE INDO-PACIFIG ISLANDS, 177 condition the formation was much more agglomerative, and conse- quently approached closer in phonetic structure to the great agglo- merative alliance. In this respect as in the character of its elemen- tary sound, and in the absence of the regular vocalic harmony that has been developed or diffused throughout the Scythic family it appears to associate itself with a proto-Scythic phonetic type. 3. The basis cf the Dravirian vocabulary is monosyllabic. In this stage it is connected with the ultimate monosyllabic basis of all other langnayes, and by its pronominal roots, as well as many others, it specially connects itself with Chinese. . 4. The actual form of the vocables is in general that of a root with definitives attached, usually postfixually but in some cases prefixually. In this stage it connects itself generally with all the existing harmonic languages; more closely with those formations in which the Scythie postfixing of definitives prevails to a greater or less extent although combined with prefixes, as in Caucasian, Indo-European, Semitico-Libyan, N.E. Asian, American and proto- Seythie ; and spreially and most closely with the Scythic forma- tion itself in which this habit is predominant and almost excludes the prefixual. The postfixing of the pronoun possessively, and the attainment through this of the assertive form, are referable to the same idiom, and embrace a similar range of affinities. The South Dravirian group like the Indo-European formation has lost the primary universality of the habit, but, as in that formation, the postfixed pronouns and pronominal elements in assertives are a remnant of it. The Kol group in one class of words retains the idiom with substantives. The idiom is fully preserved in Seythic; in some American languages; in Semitico-Libyan with substan- tives and in most of the languages with assertives ; in the Caucasian languages with substantives and in some with assertives; in Euskarian, as in some tenses of Libyan languages and in Indo- European, with definitives used as generic or absolute assertives only; in Malayu-Polynesian with substantives and in one group with assertiyes. The pronoun is prefixed in all or in some cases in certain of the Caucasian, Semitico-Libyan, Zimbian, Yeniseian and American languages (following the Chinese and Tibeto-Ultra- 178 ETHNOLOGY OP THE INDO-PACIFIC ISLANDS. indian collocation). ‘The postposing or postfixing of words used to denote the other generic relations of substantives and assertives isa further phenomenon referable to the same idiom, for all formatives whether used with assertives or substantives are ultimately reduci- ble to definitives and substantives. In the general position of the formatives Draviro-Australian resembles Scythic, Indo-European, and Assetic generally; Semitic in its archaic directives, and Zimbian in its assertive formatives; while it differs from the Semi- tic assertive formatives which are prefixual and infixual or flexional and from the Zimbian directives and definitives which are prefixual. With the Euskarian and American systems it agrees more closcly than withthe Cancaso-African. Hence the forms of the Draviro- Australian words, whether substantival or assertive, whether simply combining a concreted definitive with the root or clothing it with pronouns, directives or formatives, normally agree with the Scythic and proto-Iranian forms more completely than with those of any other fumily. In its generally postpositional and inversive collocation, and several affinities in particles and idioms, it is Scythic, although other for- mations also possess several of these common characters. Thus the inversive tendency prevails to a large extent in the American languages, in archaic Iranian, in Euskarian, in Caucasian and in various degrees in the Semitico-Libyan and more especially in some of the Mid-African members of that alliance. The negative asser- tive is not only Seythic but Zimbian, The dual of the Kol and Australian groups and the double form of the Ist pronoun plural are very archaic andcommon idioms, But the general combina- tion of traits, positive and negative, is much more akin to Scythic than to any other formation. 5. The principal idiomatic peculiarity when compared with Scythic, is the distinction of sex in the 3d pronoun and to a certain extent in the postfixed definitives of substantives,—an Indo-Euro- pean and Semitico-Libyan trait. 6. In abstract and flexional development it has a wide range of affinities in its Australoid stage. In its Dravirian condition it takes its place with the more flexional Scythic languages. It ia much more crude than Iranian in its historical development or ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 179 than the more flexional Semitico-Libyan languages. The peculiarities of Dravirian and Draviro-Australian, even when compared with those Asiatic families that most cloxly resemble it, are conclusive against the hypothesis that it was derived from any of these. The common characters are referable to a mother formation which diverged into distinct channels and received special modifications in each, these main streams in their turn divaricating, while the different branches or some of them from time to time overflowed and came into mutual contact. Dravirian probably passed through an Australoid condition, and it is even probable that in a still older race it was more agglome- rative and pleonastic, richer in forms although eruder and lesa flexional. But it cannot have passed through an Iranian, a Semi- tico-Libyan, a Zimbian, a Caucasian or even a proper Seythic condition, All the distinctive characters of these formations are referable either to individual development and modification since they were separated from the stock common to all, or to their separation having preceded that of Draviro-Australian from proto- Scythic. It is probable that the more distinctive characters of Scythie were acquired subsequently to the migration of the Drayiro- Australian family to the southward. The full development of the vocalic harmony probably took place in a branch of the Scythic family that had not become predominant till after that migration. The connection of Draviro-Australian in the Scythie continued till the postpositional structure had been developed. Its radical con- nection with the other formations belongs to periods preceding that development, 7. Glossarially the Draviro-Australian affinities have a wide range. The pronouns, numerals and definitives are E. Asiatic and Seythic, Several particles are Scythic and several are not only Scythic but S, W. Asian and African, The pronouns are not the prevalent Scythic, and their Chinese and other affinities lead to the inference that the basis of Draviro-Australian was not.a Scythic language, buta distinct one which was placed within the range of the proto-Scythic development and took a similar form. But the vocabulary although not purely Scythic in its basis, has in common with Seythic yocabularies a large proportion of roots 180 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. and yaricties of widely scattered Asiatic roots, Amongst the Mid and Noith Asiatic affinities the Samoiede, Yeniseian and Uygrian are more numerous and often more close than the proper Tatar or any others, save the Asonesian. The Mid-Asian affinities of the latter are equally striking and very numerous and embrace a multitude of vocables not now found in Indian vocabularies. The Draviro-Asonesian languages have also a considerable number of vocables in common with the E, Iranian, Caucasian and Indo- European tongues and with the more Scythoid of the African vocabularies. he affinities with the proper Semitic voeabularivs are less numerous, The affinities in ultimate monosyllabic roots embrace Chinese and Tibeto-Ultraindian vocabularies. The aflinities with Semitic and African languages appear to be mainly indirect and referable to the common Mid-Asiatic element, but some are direct and imply an early and active commercial inter- course by the aid of the monsoons along the northern part of the Indian Ocean, The special affinities of the proper Dravirian with the Caucasian vocabularies are striking. It is probable that the most numerous classes of glossarial affinitics are connected in origin with the most striking phonetic and ideologic affinities. ‘The most positive inference that we ap- pear to be warranted in drawing is that the strongly Scythic character of Dravirian, and a large number of the Dravirian vocables, are referable to a variable Ugroid or proto-Seythie forma- tion which early prevailed in Mid-Asia, and by successive ethnie movements diffused its form or extended its influence not only to the Caucasian, Iranian and Indian but to the East and Mid-African languages, The numerous and striking resemblances of Dravirian to East Iranian, East-African, Caucasian and Mid and North Asiatic, particularly Ugrian, Samoide and Yeniseian, vocablis are best explained in this mode. As the Scythic tribes have always been the most nomadic, and the form of their langua- ges is deeply impressed on Dravirian, it is reasonable to regard their mcvements as haying been the common cause of these re- semblances. This enquiry, slight and superficial as it has been, may serveo t ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 18! show not only thatthe ethnic history of the earlier races and lun- guages of India and Asonesia is intimately connected with that of other Asiatic formations, but that larger and more exhaustive explo- rations of the affinitives of roots and vocables will certainly lead to many positive historical results. But the comparative glossology of the other languages of Asia and of the world must be prosecu- ted simultancously, for it is clear that the history of every separate vocabulary becomes more and more implicated in that of others, and embraces a wider and wider circle of relationship the further our researches penetrate into antiquity. As each successive formation of Lrania Become better defined, a clearer light will be thrown on the later stages of the Dravirian. But much of its more fundamental history will continue to depend on the progress of universal comparative glossology. Although in phonology, ideology and glossary it is distinctly connected with the Scythie, and also in a less degree with the Caucasian and Africo-Semitie alliances, it has so large a mass of peculiarities as to prove that, since the eras in which that connection arose, the languages of Western Asia and probably of all Asia have under- gone great changes. At one time Dravirian or Australian may have closely resembled languages of the Panjab, of Persia or of Upper Asia, but no ethnologist would expect to find such a resem- blance now. From all the preceding indications we are warranted in concluding that ethnic movements similar to the historical ones, sometimes rapid, and at other-times gradual, liave in all eras been going on from S. W. Asia to India and from India to Ultraindia and Asonesia. These movements have always left glossarial traces of greater or less importance, and we may therefore hope that in the progress of ethnology each will be more or less clearly fiedned, ri ly 5 dt Go. ; tt eA H fore it Py 182, ETHNOLOGY OF THE INDO-PACIFIC I8LANDS. CHAPTER VI. ENQUIRIES INTO THE FTHNIC HISTORY AND RELATIONS OF THE TIBETO-ULTRAINDIAN AND MON-ANAM FORMATIONS, [Introductory Note—The conclusion that the Mon-Anam nume- rals as well as the pronouns are of North-East Dravirian origin {chap. y. sec. 11), affects the views previously advanced in these papers as to the ethnic position of the formation, and the Si-fan vocabularies of Mr Hodgson haying now placed beyond all doubt the lines of connection between the Tibetoid languages of Ultra- india and India and the Tibetan and Scythic, it becomes necessa- ry to alter the order’in which I had treated of the Ultraindian languages in this part. Instead of having to ascertain the distine- tive characters of the Ultraindo-Gangetie group by a prior approximative determination of those of the Mon-Anam, we can now proceed much more surely by reversing the order. The form and substance of the Burma-Gangetic branch when it entered Ultraindia being traced through its affinities with the existing languages of eastern and western Tibet, a well defined basis is obtained for the investigation of the original condition of the older Ultraindian languages. The surrounding and intrusive formations —Chinese, Tibetan (Si-fan, Bhotian), Dravirian and Arian are all referable to foreign lands, and when the alien ingredients which the mixed languages of Ultraindia owe to these formations have been successively removed, we may hope to arrive at the native Mon- Anam residuum. The order I had adopted in considering the for- mations following the Dravirian was “ B. the South Ultraindian or Mon-Anam; ©. the Tibeto-Ultraindian or Burma-Himalayan ; D. the Tibetan.” (vol. vi. p. 658.) The arrangement now adopted is—A. the Tibeto-Burman formation, I, The Tibetan branch embracing 1st the Si-fan languages and 2nd the Tibetan proper which it may now be preferable to term Bhotian ; II. the Ultrain- do-Gangetic branch ; B. the Mon-Anam formation. As the Si-fan dialects have not hitherto been noticed, it becomes necessary to consider their characters so far as the materials supplied by Mr Hodgson allow. The sections relating to them are therefore to ETHNOLOGY OF THE TNDO-PACTFIC ISLANDS. 188 be received, in some of the details, as supplementary to sec, 2 of chap. iv. To show how Mr Hodgson’s Si-fan vocabularies affect the general inferences at which I had arrived, I may be allowed to refer to some of the earlier portions of this series of papers. In the 2nd section of that “ on the ethnology of South-Eastern Asia”’ (vol. iv. for 1850, p. 464) the following remarks were made on the distribution of the Tibetan tribes. “The western or inner division is chiefly oodiiled by the Tibetan tribes who possess the whole of the great trans-Himalayan depres- sion which slopes westward ‘to the margin of the Hindw+Khush, . forming the transalpine basin of the Indus, and eastward to the unknown point where the basin of the Zangbo bends south and sends its waters into the basin of the Brahmaputra or of the Trawadi. They have even extended to the 8. Hast and entered the upper part of the eastern basin of the Brahmaputra where they are in contact with the Mishmi. ‘Tibetan tribes and others allied to them have spread over the basin of the Ganges, although they are now chiefly confined to the Himalayas, the Vindyas and the basin of the Brahmaputra. In the basin of the Brahmaputra they are blended with allied tribes of the Mayama family. Rude Tibetan tribes of nomadic predacious habits, known in Tibet chiefly under the generic name of Kham and in China under that of Si-fan, are spread over all Tibet to the northward of the depression of the Indus and Zangbo, and eastward along the greater part of the eastern margin of the inner division to a considerable distance within the boundaries of the Chinese Provinces.* They probably come in contact with the inner tribes of the Brahmaputra and Trawadi basins, and are intermixed with the most westerly Chinese tribes and the Mongolian tribes who chiefty occupy the northern and N. E. portions of Tibet. “The ethnology of the E. middle division is very obscure, and will probably prove to be of extraordinary interest. In a region of which a great portion is inaccessible from’ lofty mountains and snow, many of the inhabited districts must still "be secluded. Numerous petty tribes must retain their ancient independence and * They are found to the west of the Yalong and probably in some places reach to the Yun-ling mountains. 184 ETHNOLOGY O§ THE INDO-PACIFIC ISLANDS. their aboriginal languages and manners, and it is probable that amongst the former some will be found intermediate between the Chinese, the Burmese and the Tibetan. This region promises to be the richest for ethnological discoveries of any that yet remains unexplored in Asia, or perhaps in the world. All the 8. B. Asian tribes appear to meet init. On the south the upper division of Burmah and the Chinese province of Yun-nan are known to con- tain many rude tribes akin to the Burmese and the Lau and all or most of the Turanian races who now occupy the lower basins of the rivers which descend through this region must have been derived from it. The great provinces of Sze-chuen and Kan-suh are also known to contain rude tribes, and the languages of even the more civilised communities of the latter are peculiar. In the western parts of these provinces the Kham or Si-fan of Mongolian habits, and the true Mongol tribes of the Mongfan and Kukunor Tartars meet the Chinese tribes. In the 8. the Mongfan are in contact with the most northerly tribe of the Irawadi basin, the Khanung. The civilised Chinese have pushed themselves into all the more open and fertile portions of the western Provinces. It is through the Province of Kan-suh that the great trading route lies which connects China with Western Asia, and the movements along which must in all eras have affected the distribution of the tribes of middle Asia.” In the Introductory paper (vol. iv. p. 441) and in the earlier chapters of this Part the terms Tibeto-Ultraindian and Tibeto- Indian are used as descriptive of these Ultraindian and Indian languages that are allied to Tibetan, but distinct from the deriva- tive Tibetan dialects of the Himalayas. In the Introductory paper I remarked that the languages in question had distinctive features when compared with Tibetan, and that the Tibeto-Indian tribes were directly connected not with the Tibetans but with “a proto-Tibetan era when the present widely spread Tibetan race may have only been one of several rude trans-Himalayan tribes speaking dialects of an incipient Tibetan character or even of one nearer the Chinese.” The proto-Burmans, it was remarked, “probably occupied some portion of the country on the bounda- if henaesen to Chinese writers some of the eastern Tibetan dialects approximate c ETHNOLOGY OF THY INDO-PAGIFIC ISLANDS. 185 ries of China and Tibet. Many other intermediate. languages may have existed and some are probably still preserved.” In the earlier chapters of this Part the lin® between the Ultraindo- Gangitic languages and the Tibetan was more brondly and dis- tinetly defined. In chap. I. the former was marked out in the following passage. “The next Ultraindian formation was the Tibeto-Ultraindian which is distinguished from the Mon-Anam by its Tibetan or post-positional and inversive character. It em- braces the Burman, the Karen, the Yuma dialects from Kyen to Kuki, the Manipuri, Naga, Mikir, Singpho, Mishmi and Abor- Miri. It also spread westward up the Gangetic basin and into that of the Sutlej ; the Garo, Bodo, Dhimal, the Akha, Changlo and the other Himalayan languages, as far westward as the Milchanang and Tibberkad, belonging to this formation so far as they are not Dravirian, Tibetan or Arian, and so far as they do not preserve remnants of the Mon-Anam formation, the latter being slight on the north side of the Gangetic valley compared with the south or Vindyan. This Tibeto-Ultraindian formation I conceive must have originated at a very ancient period in eastern Tibet or the adjacent territory now Chinese, because it is inter- mediate between Chinese and Tibetan and more closely connected with the latter than the former.” The Si-fan vocabularies which we owe to Mr. Hodgson have partially removed the veil which hung over eastern Tibet, and my anticipation that the ethnology of this region when explored would prove to be of extraordinary interest, has been verified. Much remains to be ascertained before we can enter ona full investigation of the relation of the Si-fan dialects to the Tibetan and Ultraindian, but enough has been published to satisfy us of the’ important fact that the Ultraindo-Gangetic languages are more closely connected with the Si-fan than with the proper Tibetan dialects. It will now be convenient to*distinguish the latter by the national name of Bhot and to use Tibetan as includ- ing both Bhotian and Si-fan tribes and languages. The term Tibeto-Ultraindian or Tibeto-Burman may be applied to the whole family—Tibetan, Ultraindian and Gangetic—and Ultraindo- Gangetic to the southern branch, excluding the southern Bhotians. 186 ETHNOLOGY OFg THE INDO-PACIFIC ISLANDS, T have not thought it necessary to rewrite the whole of this chapter. We may expect further information from Mr. Hodgson respecting the Gangetic @ind Si-fan languages, and with the pre- sent materials, it would still have remained fragmentary whatever shape had been given to it. It will be understood therefore that ~ much of the chapter remains as it stood before I received the Si- fan vocabularies, but wherever it appeared advisable in order to save repetition I have embodied the new data. In other cases the additions constitute separate sections or paragraphs. No great inconvenience can arise from the Bhotian and Si-fan branches being to some extent separately treated» There are indeed reasons in favour of such an arrangements. Bhotian is the only Tibetan dialect that has been investigated in detail and its influence on the Ultraindo-Gangetie languages is to a certain extent distinct from that of the Si-fan dialects.) 5s THE TIBETO-BURMAN FORMATION. Sec. 1. THE GENERAL ‘CHARACTERS OF BHOTIAN, AND ITS RELATION TO CHINESE AND SCYVTHIC. The phonetic and ideologic relation of the Bhotian to the Gangetico-Ultraindian languages in general, and to the Burman in particular, as that of which the grammar is best known, has been already considered. The result of our enquiries may be stated to have been that this relation is of two very different kinds and belongs to widely separated eras. A formation inter- mediate between the Chinese and the Bhotian, and, it may be added, having some Seythic affinities of its own, spread into Upper Ultraindia at a remote period, its natiye seat having been in all probability the adjacent province te the northward comprising eastern Tibet and a portion of N. W. China. Of this formation the Burman branch of the Ultraindian languages is the best known representative. But it is a comparatively recent or much modified form. The older form was less emasculated, its vowels were broader, and it used prefixes which gave it a dissyllabic rather than a monosyllabie form. The archaic formation spread down the Irawadi and is tiow best represented by the Naga, Manipuri and Yuma dialects. This form of Tibeto-Burman appears to have preceded the Burman eyen in the valley of the Irawadi; and the EYUNOLOGY OF THE INDO-PACIFIC ISLANDS. 187 ether dialects of the same group retain its phonology more tenaceously than Burman. It also spread to the westward from the Asam valley to the head of the Sutlej, all the Gangetic band of Tibeto-Ultraindian dialects from Mishmi to Milchanang adher- ing to it to a great extent. This form has itself several phases. The earliest appears to have been broad, sonant and in its finals consenantal, The later show various degrees of vocalicism, the final consonants being softened or elided. In all the groups, and in some cases even in local subdivisions of the same dialect, the broad and strong phonology still co-exists to a greater or less extent with the soft and slender. The current and the old or written Bhotian (chap. iv. See. 1.), the different Abor dialects (ib. sec. 5), Burman when compared with Karen, Karen when com- pared with Khyeng and the other Yuma dialects, Gyarung when compared with Thochu or Bhotian, all illustrate the progressive emasculation of the phonology, and in most of the dialects archaic broad voeables are current along with slender ones. In the Gyarung-Burman or Eastern Tibet and Irawadi band the atten- uation is most marked. In the Burman phonology the propensity i ellipsis, slender vowels and consonants,—as 7 for a, e for i, ¢ for , y for r—has received a peculiar development. This atest form is found most strongly marked in Burman itself which has become _ highly monosyllabic and attenuated. In Karen and some of the other members of the proper Irawadi group the older form is more persistent, The history of the direct and exclusive Bhotian influence to the southward of the snows is quite distinct. It began by the migra- tion of Bhotians across the Himalayan passes, the occupation of Bhutan, the partial occupation of more western districts, and the diffusion of Bhotian political and ethnic influence not only over the prior Himalayan tribes but partially also over those of the Gangetic plain and North Ultraindia. The Bhotian language was transported to this side of the snows. It partially communi- cated its forms to the Himalayan languages from Milehanang to Abor-Miri, and in a slighter degree to the Middle Gangetic (Dhimal, Bodo) and some of the North Ultraindian (Garo, Mikir, Naga &¢.) It thus appears that the proper Bhotian influence on le2 ETHNOLOGY OF THE INDO-PACI¥YIC ISLANDS. the Indian and Ultraindian phonologies and ideologies was incon- siderable. It remains to enquire into the extent of its glossarial influence. The connection between the tribes and languages of Tibet and those of India, Ultraindia and Asdnesia, appears also to render a brief enquiry into the trans-Himalayan relations of the Tibetan necessary for a satisfactory view of the ethnology of the Indo-Pacific islands. I shall proceed to this, in the first place. The cis-Himalayan Tibetoid languages have distinct affinities with those of the Tatar and more northern hordes of Asia, ‘There has evidently been more than one southern movement of the Tibetans in different eras. Tibet has always been exposed to the incursions of the nomadic Tatars, who have, in turn, spread them- selves*over the steppes between southern Tibet and the great Desert. The relations of Bhotian, in its present form, to the more northern languages, may therefore throw some light on the pre- historic changes which it suffered, and connect the Scythie revo- lutions in which they originated, with the ethnology of the pro- vinees to the south of the Himalayas. — In preceding chapters it was remarked that Bhotian was so highly Scythie in its ideology that it might be considered as ag non-harmofic member of the Scythic family. “The phonology preserves a crude or Chinese character almost to the same extent as the Burman. The earlier form of Burman appears to have been harsh and sonant like the purer Bhotian and both are essentially monosyllabic and non-harmonic. In this respget they depart greatly from the Seythic phonology and especially from its more agglutinative varieties. But the basis of even the Ugro-Japanese languages is monosyllabic with very little disguise, and many of them preserve a strong sonant and aspirate tendency. It is probable therefore that at the remote period when the Ugrian formation first modifiedethe earlier and more Chinese form of Tibeto-Burman, the former was equally sonant with the purer Tibetan. In the Ostiak and even in the Turkish vocabularies words frequently occur entirely Bhotian in character. Some of these are found little changed in Bhotian, For example the Ostiak log-ol, “hand”, is evidently the parent of the Bhotian lag, the Turkish haying the slender form é-lik. The Turkish syod ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 189 * light,” preserves the original of the Bhotian hold in the same sonant form. Ugrian and Turkish retain sonant forms of an ancient root for “river,” which has become softened in the pre- valent Tatar, Tibetan, Ultraindian and Asonesian glossologies (Comp. Ugr. ja-an, Turk., sug, Turk., Mong. wsun, chun, Tibeto-Indonesian chang, sung &c.) The Ugrian log, “ horse,” (also lo) is more sonant than the derivative Himalayo-Burman and Indonesian forms rang, ra &e. In the less emasculated Indo- European vocabularies, the sonant forms of the ancient Turanian roots are frequently retain. There can be no doubt that the Chinese and Chino-Ultraindian or Mon-Anam formation was also originally highly sonant, but. the strong glossarial affinity of Bhotian to the Ugrian alliance renders it clear that the sonant character of Bhotian was immediately related not only to the archaic »Chinese but to the Scythic, and through it, to the archaic Indo-European. It has a greater range of final consonants even than the most consonantal and sonant of the known Chinese dialects, the central and southern. At the period when the Tonic Dictionaries were compiled—the 6th or 7th centuries of the Christian era—the plionology of the Kiang provinces was more emasculated than the written Bhotian. The latter probably preserved an example of very archaic Chinese phonologies, anterior it may be to the development of theyharmonie phonology and when the mother dialects of Scythic, Indo-European and all of other formations consisted of crude, monosyllabic and tonic roots*. The Bhotian phonology is much cruder and more archaic than the Scythic or that of any of the other harmonic formations. When the formation separated from the common stock the latter was little in advance of the Chinese, monosyllables and homophons abounded, agglutination was feeble or only beginning to affect the form of vocables, the definitives and other particles were not concreted with substantial words or with each other. The Bhotian phono- logy contrasts so strongly with tne highly harmonie Scythic that » ™ Since chap. III. was published the Rev. Mr Edgkin in his Grammar of the Shanghai dialect has shown that the sonant tendencies of some of the middle and southern languages are more decided than previous Grammars had led us to be- lieve. In a subsequent section the results of Mr Edgkin’s original and important enquiries into the phonologies of the Chinese dialects will be noticed. 190 ETHNOLOGY OF THE INDO+PAOIFIC I8LAND3. it may even be considered doubtful whether the harmonic deyelop- ment had commenced when the mother-dialect of the former was first separated from the proper Chinese. Much of its slight agglutinative and harmonic power has probably been acquired since, and Scythic may haye had much. influence on its progress. In another place the conclusion was arrived at that the collocation of Seythic was older than its harmonic phonology, and in Bhotian we have a partial example of its pre-harmonic condition. Tis. general structure although Scythic when compared with Chinese, Mon-Anam, Asonesian, Semitico-African and Cancasian, is not purely Seythie. The use of postfixed definitives is an archaic Turanian, Cauca- sian, Semitico-Libyan and Indo-European trait. The most com- mon Tibetan postfix* ma, pa, va, ba, &e oceurs frequently in Ugrian vocabularies, and it is also Semitico-Libyan, Caucasian, Indo-European and Dravirian. In Chinese it isa 8rd pron. The postfixes distinguish Bhotian strongly from Chinese and there ean be no hesitation in considering them as of Ugrian affinity. The other Turanian, postfixed definitives are na, ni, n, Sc ; 1a, la, ol, el, er, Se; ha, ga, h, Je; s, %, t, d, ch, j; which with the labials comprise the whole range of the Tibetan. The prefixed consonants of Tibetan b or v, m; h, 3, z; 1, 7, d; g are not prevalent in the Turanian languages, but Hungarian has az, as a separate preposed definitive, and in others vocalic prefixes occur which are probably in many cases contractions. Turkish appears to have prefixual ¢, d, ch; l, s concreted. The Yeniseiau languages will probably prove to be the chief existing link between the proper Seythic and the N. B. Asian’and Ameri- can. In many respects they may be considered as entering with the Samoiedean group into the Ugrian family. Bat with strong Ugrian affinities they combine independent traits, and others that are N. E. Asian and American. Amongst the latter is the retention of prefixed, along with postfixed definitives, embracing the entire * I give a few examples in which both the root and the postfix are the same in ibe two families, ar Bhot, loma, Mordv. ope sormo, Fin so Rain, A, Bhot. charba, Sam. sero. = ETHNULOGY OP THE INDO-PACIFIC ISLANDS. 191 Seythic and Bhotian range,—ma, pa, pi, bi, &e; ta, da, di, d &e ; si, hi, chi, cho &e; al, il, ol, &e; ki, ke, ku, gi, yi, &e ; a, u, 0, i, e. These definitives are more common as prefixes than as postfixes, and when the habit of the formations which sueceed Seythice on the N, E.* and 8. W. and have fundamental affinities with it, is considered, no doubt can remain that the distinctively postfixual idiom of Scythie was exceptional in its origin, and was preceded by a condition of the mother-language in which the definitives were vurrent ag separate particles, and capable of being preposed as well as postposed according to dialectic taste and fashion. To this proto-Scythie stage of the Mid-Asian formations Bhotian, like Yeniseian, partially adheres. In this respect their form is older than the proper Seythic aud more akin to the basis-form of the Caucasian, Semitieo-African .and other formations that separated fram the common stock before the dialect in whieh Scythie origi- nated had acquired its peculiar postpositional structure. In the use of prefixed definitives as in many other traits the Tibeto-Ultra- indian and N. E, Asian families have departed less than the Seythie from the archaic type preserved by Chinese. In Chinese the true definitives precede the words they detinet. The full range is also preserved in Chinese, although the definitives are now rarely used save emphatically or as demonstratives. It has ki, ke, chi, che, ti, i, ku, tsze, hi, ho &c; na; and pe, wa, Chinese also. uses double demonstratives, or rather the demonstratives fol- lowed by the generic definitive or segregative ko, ku,—na ko, che ko, ti ku,iku, kuku. In the firat stage of an adhesive phonology these would become nako, cheko, tiku, iku,kuku, They are thus the prototypes of the double definitives, prefixes and postfixes found in most of the harmonic formations. It is obvious that the full forms of the definitives, as in Chinese, must have preceded that in which they lose the vowel and coalesce with the root inte one monosyllable. The Bhotian initial consonants were originally separate preposed definitives and they are preserved in the full form as prefixes in other dialects of the Pap ge ho * ocle-ap up has prefixes aa well-as postfixes— aha, Se ae Ra: Be Hs Yulhitl tine alay pretines bat it its 60, 8a, general se Beak er hy G arang prefix ki nel definitive ki, ke, chi, che. Hence we * ba such as ee Gare chi tun Gyami, "egg. Kwan-hwa has the G yarung vowel tan, 192 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Tibeto-Ultraindian family. In the N. B, Asian, Caucasian, Semi- tico-African and Asonesian provinces both forms of the prefixes are also found. . ’ In Tibetan the labial definitives are still current in their prima- ry character of substantive words “ father”, “mother”. As a definitive postfix —pa, —po has acquired a generic masculine appli- cation, and —ma, —mo a feminine, and they are even extended to neuter names. In Scythic both the primary and the sexual significations have been lost. In Draviro-Australian, Indo-Euro- pean and Semitico-Libyan agglutinated definitives are found retain- ing a sexual foree but with the primary substantial meaning lost. Tibetan here also stands between Chinese and the more aggluti- nated and conereted formations. In Chinese there are several classes of postposed sexual particles, as in Tibeto-Ultraindian and Dravirian. Thus for human beings Kwan-hwa has nan masc., neu Jfem.; for the lower animals generally kung m,, mu f.; for birds heang m., tsze f As in Bhotian, Indo-European and Semitieo- Libyan the idea of gender has been transferred to inanimate things, for which keén m., kwan f. and yin m., yang fare used. In some of the Scythie languages there are traces of a similar attribution of a distinction of sex, enerzy &c to inanimate objects. A marked departure not only from the Seythico-Dravirian but from -the Chinese collocation occurs in the position of the quali- tive, which follows the substantive. This idiom connects Tibeto- Ultraindian with the adjacent Mon-Anam. - It is clearly abnormal, because the primary relation of possession and attribution, of which the qualitive is but a variety, is denoted in the Tibeto-Ultra- indian languages, as in Chinese and Seythie, by preposing the pos- sessive. Consistently also with the normal structure the adverb precedes the qualitive or verb, and the subject the predicate. The Bhotian glossary is highly Scythie but in its basis it is independent to a considerable extent and with strong Chinese affi- nities. The Seythic glossarial basis, in pronouns and many parti- cles and formatives, is so uniform that it may be referred to one mother-dialect. The Bhotian basis is not a modification of this dialect like that of all the Seythie languages, It is a distinct Chino-Seythie sub-formation, and Chinese more than Scythie. ETHNOLOGY OF THE INDO-PACIFIC ISLAND#. 193 Sec. 2. THE GENERAL CHARACTERS OF THE 8I-FAN LANGUAGES AND TPEIR RELATION TO BHOTIAN. Since this paper was written Mr Hodgson has published a series of vocabularies spoken by the tribes occupying the mountainous country between the land of the proper Tibetans’ or Bhot and that of the proper Chinese. These vocabularies are of remarkable in- terest, They prove that the Tibeto-Ultraindian formation extends ‘northward, from the most northerly dialects previously included ' in it [Singpho, Jili] toa point in N. E. Tibet which has not yet been ascertained, but where they appear to be succeeded by Sok or Mongolian tribes identified by Mr Hodgson as the Olet and Kal- mak of Remusat and Klaproth. These Mongolians occupy the eastern portion of northern Tibet, the western being in like manner the southern extremity in this quarter of the Turkish province and traversed. by tribes called by the Tibetans Hor and considered by Mr Hodgson to be Turkish. These Tatars chiefly roam on the north of the Nyenchhen-thangla range but there are also numer- ous scattered Horpa and Sokpa in southern Tibet. The new series of Tibeto-Ultraindian vocabularies comprises, Ist the Takpa (of the so-called Towang-Raj west of Kwombo), 2nd the Manyak,* Gyarung+ aad Thochu spoken by tribes which oceur in this order; between Yunan and Amdo, the latter diyision of Tibet being occu- pied by a Si-fan tribe who for the most part speak Bhotian. To these are added the Gyami, a dialect of Chinese, and the Sokpa and Horpa. The last is considered by Mr Hodgson as Turkish, but it appears to be Tibeto-Ultraindian in phonology and glossary. It is a very archaic dialect of Chino-Tibetan, preserving some evidently archaic varieties. of the common root now obsolete in Chinese, in its forms intermediate between Bhotian and the East Tibetan dialects but leaning more to the latter than the former, and possessing special affinities with current Chinese and Tatar, from which it may be inferred that Horpa has not only been long conterminous with Scythic languages, but that it was in contact with Si-fan. dialects and like the southern Takpa directly acted on by Chinese before the modern expansion of Bhotian to the east- ie Hodgson describes the physical characters of a Manyak, a native of Rakha, siz south t Mr Hodgenn dactives sO} & uel from Tazar, north of Tachindo, 194 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. The most important conclusion to be drawn from these vocabu- laries is that three at least of the Tibeto-Ultraindian ones, the Manyak, Gyarung and Takpa, are allied to the TIrawadi or Ultraindian branch of the family more than to the Bhotian, The general structure and phonetic form of the vocables resemble those of the Ultraindian vocabularies that have been least modified and emasculated by the influence of yocalic Chinese, and especially those of the Naga-Manipuri group. Manyak and Gyarung how- ’ ever have also Burman forms. Gyarung, Horpa and Thochu have a considerable portion of final consonants. Manyak is vocalie in this respect, in its slender vowels and in the forms of several of its words resembling the emasculated Burman sub-formation, Slender vowels (i, ¢) are more common in all the dialects than in Bhotian. They abound in Manyak and Gyarung, especially in the latter, which in more slender, but less elliptic, than the form- er. Thochu and Horpa, especially the latter, have more frequent- ly broad yowels as in Bhotian. But it is to be remarked thata special connection exists between Horpa and Thochu and between both and Manyak. The glossarial affinily between Thochu and Manyak in particular is often very decided. The common varie- ties have often @ as the vowel where the other Tibetan or Chino- Tibetan varieties of the same root have 0, u, i, Ke. I give a few examples of the great attenuation some of the Tibe- to-Ultraindian yocables undergo in Manyak “ I,” a (nga Gyarung) ; “ arrow” m-a (m-la Takpa); “bird” ha, (bhya Lhop.); “ boat” g-u (g-ru Takpa, Bhot.) ; “ village” hu (khyu Gyar.); ‘ ripe” de-mi (min Thochu, ka-s-man Gyar.); “ black” da-na (ka-nak Gyar., nya-nya Horp.). The vowel generally retains an archaic broad form. The broad phonology appears to have predominated in archaic eras. Some of the remote Scythic and N. E, Asian languages still affect broad vowels. The Scythic languages vary amongst themsclyes in this respect, but in many there is now a disposition to slender vowels. The Turkish dialects frequently affect them. In the modern Chinese they are common, and the strong development of this tendency and of gencral emasculation in the Tibeto-Ultraindian languages, and especially in the eastern ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 195 or Gyarung-Burman band, is chiefly ascribable to the predomi- nance and diffusion of the modern North Chinese or Kwan-hwa phonology. But the slender forms of the Sifan-Burman vocabu- laries are not always to be considered as the result of a native development of a soft phonology under the influence of Kwan- hwa. On the contrary, many slender varieties are of the highest antiquity in all the East Asian formations,—Chinese, Scythic and Tibeto-Ultraindian. They are even preserved with the archaic final consonants in many words. The co-existence of broad and slender forms, e. g. log, lik “hand,” is in accordance with the unstable character of the vowel in the Scythic phonology. This vocable affords an illustration of the independent development of slender forms in the Scythic and in the Tibeto-Ultraindian provinces. The broad archaic form was common to both, e. g. log-ol Ostiak, lag-pa Bhot., e-lag Abor, luch-led “ finger,” _Kams- chatkan. The Abor e-lag, a-lak has the Turkish prefix, but in Tur- -kish the slender phonology has produced the forms é-lik, a-li, e-li, while the archaic broad vowels are preserved in a-la,a-lo. In Bhotian &c the g has also become k, lak-pa, and the emasculated Burman has not only evolved a slender vowel but changed & to ¢, lat, let. In this case the Burman let and Turkish lik are obvious- ly independent. But there are other cases in which archaic slender Scythic forms have spread notonly through the Tibeto- Ultraindian but through the Asonesian provinces. For example, the Gyarung-Burman li “air,” Asonesian iri &e is clearly archaic and Seythic, lil, il &e (as well as Caucasian, Semitic &c) and not a modern variation of the Bhotian lung, lhak &c. Even Thochu and Horpa have several vocables with slender vowels where Bhotian has broad ones. Ex, Hog, phak Bhot., pi Thochu; Earth, sa B., zi-p T.; Road, lam B., rah Manyak, g-rih T., tri Gyarung; Salt, chha B., cheh T.; Snake, sbrul B., brigi T. ; Bone, ruspa, ruka B., ripat T., rera Horpa. Gyarung has often ¢ where Bhotian has 0, %, a, or 7; i where Bhotian has u ore; and sometimes o where Bhotian hasa. Ex. Bird, byu B., pye G. ; Day, nyi B., nye G.; Har, na B., neG. ; Moon, la B., lheh G.; Water, chhu B., chi G.; Zooth, so B., syo Horpa, swe G., Thochu; Afouth, kha B., khe G, Gyarung is equally prefixual with the more prefixual vocabularies of N. A 196 RTHNOLOGY OF THE INDO-PACIFIC ISLANDS. Ultraindia, the common prefixed definitive being ta-, varied to to-, ti-, tir-, tar-, [as in some Ultraindian and Himalayan vocables|, da-, na-, and also passing into ka-, as in many Ultraindian vocabu- laries. Qualitives take ka-, corresponding with the Bodo ga- Manyak has fewer prefixes than Gyarung and they are more mixed. The labial which is rare in Gyarung occurs frequently under the forms ma-, m-, ba-, mer-, wo-, and postfixually in the forms -bi, -mi (the Bodo prefix b in be). De-, da- is common with qualitives, corresponding with the Chinese possessive ti, di. Thochu words have much more frequently a Bhotian form. The prefixes which are comparatively infrequent occur both in the Gyarung and Ultraindian voealised form and in the Bhotian consonantal one, mo-; ki-, cha-, ra-, da-; cha-, ki-, r-, 3-, g-, h-. Some words have also the Bhotian postfixed labial definitive (-mo, -pa, wo, &e.) The numerals take -ri, -re, the Scythico- Tibetan poss. Horpa has also prefixes and they are generally in the conso- nantal Bhotian form s-, 7-, r-, k-, v-, Qualitives have frequently ka- &c, and assertives ta-, tan-; tam-, ta-r, kha-, gu-, gu-r-, na-, na-ha-, na-p-, ya-, rha-, rhang-, zu-, zu-r-, wa-, wa-n-, Ke asin Gyarung, Thochu and Bhotian. The use of prefixes in languages so far north as Gyarung, Thochu and Horpa renders it probably that this habit also charac- terised the eastern and southern branch of Tibeto-Ultraindian in its primary form, thus confirming the opinion expressed in chap. iv. (Journ. Ind. Arch. vii, 126) that the system of prefixed and pre- posed definitive: was the original one of the whole Chino-Tibetan linguistic province, as of a much wider area, and probably also the earliest in the world. In harmonic power the Gyarung appears to be somewhat in advance of Bhotian, but this may arise from the curt and conso- nantal phonology of the latter having obscured the vowel changes. In agglutination they are probably nearly alike. In Gyarung the vowel of the definitive appears to be affected by that of the principal word as much as in the closely connected Dhimal and Bodo (see chap. iy. see, 3). Hence the superiority in this respect of these Gangetic dialects over most of the Ultraindian can no ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 197 longer be exclusively ascribed to the influence of the Dravirian phonology. In Gyarung it must be cousidered as an acquired Scythic trait and in Bodo and Dhimal it must be Scythic through Tibetan so far as it is Tibetan. Mr Hodgson gives we-pe “ his father,’ wo-mo “his mother’ &c. In like manner the vowel of the root is modified by that of a postfix. Mr Hodgson gives a few examples of the formative system of Gyarung. The formatives are prefixual as well as postfixual and they are to some extent combinable, as in Bhotian on the one side and Burman on the other. From these examples it may be gather- ed that the common definitive prefix ka, ta, da, na, or yais, when the sense requires it, assertive (present) or generic. In the sonant Bhotian the definitives g- [=A-], d- [=t-], which I have consi- dered as identical with the localitive na, la, ra, &e (Journ. Ind. Arch. vii, 113), m-, h- &e, are all assertive, with a variable tense power. In Gyarung the repetition or addition of ta (ka-,ta-, ta-,ta- &c,) distinguishes the past from the present. In the nega- tive assertive hka-, ta- &c is replaced by ma-, corresponding with the Bhotian mi-; Chinese m &e. Sa, postfixed to the common assertive definitive, renders it causative. It is the Bhotian instru- mental, active, intensive and causative particle s, which in that language is postfixed to the reot. But it is also postfixed to the definitive Ja, na &c to form the ex-transitive. In Gyarung -si, -té is personative and participial. The use of double and even triple definitives is common to Gyarung with most languages which retain such particles, The power of combining them and of using both prefixes and postfixes with the same root is Tibetan-Ultraindian, N. E. Asian, American, Caucasian, Euskarian, Semitico-African, Asonesian and archaic Indo-European, that is, it is common to all the formative alliances, _ From the proximity of Gyarung to the Chinese and Chino- Ultraindian province it will probably prove to be more prefixual or less Scythic than Bhotian. But without even excepting the prefixual position of the qualitive (possessive) definitive ka- (in Bhotian -kyi &e, Changlo -ga), the examples hitherto given have parallels in Bhotian. In Chinese itself the poss. and qualitive particle is postposed, aud although Gyarung generally dispenses 198 ETHNOLOGY OF THE INDO-PACIFIC IsLANDS. with declensional signs, as Chinese does when they can be avoided, it preserves the Chinese and Bhotian idiom when it uses a posses- sive particle, asin Lama um boroh, “ the Lama's horse.” That Bhotian also used the qualitive and poss. definitive prefixually is proved by several examples amongst the namerals and qualitives. Thus gchig 1, gnyis 2, gsum’ 3, correspond with the Gyarung huti 1, hanes 2, kasam 3. When the qualitive prefixes of Bho- tian do not agree with the Gyarung ka- they are sometimes similar to the Manyak and Dhimal dé &c, or to other Tibeto- Ultraindian forms. The Gyarung verbs like the Chinese ani Bhotian are simply substantives or ecrudes and the particles of tense, mood &e stand ideologically on the same footing as the definitive and directive particles. In Chinese some of these are preposed and some postposed. In Bhotian ‘the definitive d- or da- is used as a generic assertive, while with some words it is future or aorist (past and future). B- or ba- is generic, past or aotist. M- or ma- is commonly generic, but with some words it is aorist exclusively. H- or jia- is generally present, but sometimes present and future. In Gyarung the prefixual definitives are more fully preserved and freely used than in the old or written Bhotian. But their redundant cumulation is not peculiar to the verb, as substan- tives and qualitives occur with double and triple prefixes (Hodgson, 134). In the ordinary possessive use of the pronouns they are preposed, in accordance with the regular idiom of Chinese, Bhot- ian, Scythie aud Dravirian, and not postfixed as in the abnormal or secondary and euphonic pronominal habit of most of the Sev- thic and Dravirian languages. Ex, nga-pe “my father”; na-pe, “thy father”; .wa-pe, “his father”. The same idiom is followed with assertives. Nanre na-syo, thou ridest. It does not appear that the pronoun is always thus preposed in its separate form as well as prefixed in its radical form. The asser- tive idiom is obviously the simple possessive na-syo, my-riding. In the first person the assertive or attributive root takes a postfix ~ang. Mr Hodgson appears to consider it as representing the Ist pronoun, and generally indicating a reflexive character. To this he attributes its employment in the poss, case and ils so frequently designating the first person when appended to verbs and their ETHNOLOGY OF THI INDO-PACIFIC ISLANDS. 199 participals (p. 66). In some of the other languages to which Mr Hodgson refers, it appears to me to be nota distinct reflexive ’ particle representing the Ist pronoun, but merely a variety of the 1st pronoan itself, which is the same nasal root in Chinese, Tibcto- Ultraindian and Draviro-Australian, In the Naga thien-ang, thien-o, thien-a, I put, thow puttest, he puts, ang is as evidently the Ist pronoun nga in a postfixed eaphonie form, as o is the 2nd and a the 8rd. In Bodo ang is the separate form (sce other examples ante, p.p. 37,38). It would therefore seem that in such Gyarung uses as kazang [=ha zo+ ang] “I eat” (radically “ the-eating- my’’), -ang is identical with the Naga -ang, that is, the Ist pronoun itself. Gyarung may have lost the postfixed pronoun in the 2nd and 3rd persons. That it once possessed them and that they were emphatic repetitions of the preposed pronouns is rendered highly probuble by the existing usage of the closely related Dhimal which retains them in the 1st and 2nd persons, but wants them in the 3rd. Bodo again wants them in all the persons while Namsangya Naga preserves them in all.* The usage in Gyarung must be of Scythice origin like the other Scythie traits which the Tibetan formation acquired from its contact with Scythic in its native province and retained in yariable degrees in its Ultraindian dialects. A few examples will serve to illustrate the Gyarung system of composition. From the crude root zo, eat, (or rather eating) are formed with the prefixed definitives ta-, da-, ka-, ya-, na- the sub- stantive or participial ¢a-zo, sa-zo Kc. From ka-zo, by the post- fixing of nga in its eughonic form is obtained haz-ang, my eating, [primarily donbtless nga ka-zo nga, like the Dhimal ka hade khi- * At the conclusion of chap. iv. I remarked with reference to the emphatic and euphonie postfixing of the pronoun. ‘It is not a trait that we should expect to find epeeeeneaeety shewing itself in many languages, and it is more likely to have been derived by the Gangetico-Ultraindian tongues from a highly harmonic group like the Dravirian or Fino Japanese, than to have originated close to the mono syllabic boundaries in such a language as the Naga, and been thence transmitted to nore remote and harmonic members of the postpositional alliance.” I added that if the trait were a native Indian one it probably arose in the Dravirian family and was communicated by it to the Gangetico-Ultraindian. It may now be considered that this habit, witli much of the harmonic and agglatinative tendency which I had attributed to Dravirian influences, was imported by the Gangetico-Ultrain- pir iga from its native location in Easteru Tibet, where it was adopted from Scy thie, ; 200 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, ka and the analogous Scythic forms]. With -dior -si, which Mr Hodgson wrms “the participial attributive suffix” and which is evidently the common definitive demonstrative &e, found in the ~ same form as a prefix and occurring in chi-di “this,” ha-di that” * (Bhotia wr. ha-de, de, sp. di, phi-di), the compound becomes definitive or substantival ka-z-ang-ti “ the I eating,” “I who eat,” “TI the eater.” With the causative particle sa prefixed to zo it becomes sa-zo or definitively ta-sa-zo “ feeding,’ ta-sa-z-ang-ti “ I who feed ;” and with the negative definitive ma-sa-z-ang-ti * I who “feed not.” Zu-sa-ze-st according to the context is “ he (or thou) who feedest,” the pronoun not being postfixed. From the root ma or man, sleep (nan in Thoehu) are formed ‘a-r-man “ sleep,” ha-r-tna-ng “ I sleep,” ma-r-ma-ng “I sleep not.” The repetition of the definitive in the form ta forming the past or completive we obtain ka-ta-r-ma-ng or ta-ta-r-ma-ng “T slept,” ma-ta-r-ma-ng “I slept not,” ta-ta-r-ma-ng-ti “ I who slept,” ma-ta-r-ma-ng-ti, “J who slept not,” ¢a-ta-r-me-ti, ma-ta-r-me-ti “ thou (or he) who slept” or “ slept not.” As an example of the comulative definitive prefix I may give da-na-rt-sa-gyu-ng-ti “ I who cause to run,” l. e-gyuk run, or running, sa-gyuk, make-running, da-na-ra-sa- gyuk (triple def.) emphatie “the,” “ this,’ “am,” making-run- ning, ng-ti, “ T-who” or “ I-the.” The following additional example shews that in Gyarung as in Bhotian the object precedes the assertive. Ngare nga-pe boroh dovo-ng, “I my-father horse give-I,” The pronotn when used objectively in the imperative has the eame form as when used agentively in the indicative, davo-ng “T give” or “give me.” (Hodgson 66). Horpa has also the postfixed Ist pron. in the same objective form tu-khye “ give” tu-kh-ong “ give-me.” Thochu prefixes the pronoun da-goh “ give,” kwu-goh “ give me.” Save in those points in which the Scythie inversive collocation departs from the Chinese there are few traits in Gyarang or in Bhotian which may not be considered as fundamentally Chinese. * See Sec. 3 for the various forms and uses of this definitive. It varies from ti, thi, st, di, de to ri, re, ra &e, As the relative it has the following forms shu, shui, si sui, chi chui, ei tang &e, Chinese; ¢ii-nda Bhotia ; awi-n Serpa ; ka-di lint su-re Lepelia; a-ti Limbu; sa Kiranti; te-kwe Suuwar ; su Gurung, Sowo, ETHNOLOGY OF THE TNDO-PACTFIC ISLANDS, 201 The habit of treating words as erudes, of placing a series of crudes together and indicating the common relation by a single postposed particle is Chinese as well as Seythic. Even the compounding of particles is but a remnant of the crude Chinese stage when the formatives and flexions did not exist, and when complex relations were indicated by several unconnected crudes or particles. Some of the ordinary Chinese definitives and prepositions are double, and repetition and cumulation are much used in the general structure of the language. The Tibetan languages generally may be des- cribed as sister dialects of the Chinese, in some traits standing between Chinese and the Mon-Anam family, but in their general structure Seythico-Chinese, the distinctive Scythic traits being probably secondary or acquired, If the pronouns and particles had been Seythic more than Chinese we might have recognised in Tibetan the genealogical link between the former and the latter. But as the reverse is the case, the glossarial basis of Scythic must be considered as representing that archaic dialect—allied to the Chino-Tibetan but distinet from it—in which the inversive structure was developed, and from which it was transmitted to the western or outlying branch of the Chino-Tibetan family. Sec. 3. PRONOUNS. 1 Bhotian. The lst. pronoun of Bhotian, nga, na is Chinese, ngo, ngai &e, and although not now a prevalent Turanian form its wide diffusion in archaic eras is proyed by onr finding it in the Draviro-Anustra- lian, Caucasian and Semitico-Libyan formations, and in N. E. Asian, and American formations. Other formations are also used. The honorifie nged, is distinguished by the slender vowel and the dental postfix found also in the 2d pron. The form nge occurs in the Lhopa oblique forms nge-yi, in the adjacent Takpa as the regular form nge, also softened to nye, and in the oblique form of Singpho, nge-na. It is not probable that in the Tibeto-Ultrain- dian province the e form originated in Bhotian and in Tibet was confined to that dialect. It appears to have been an archaic Tibetan form current with nga and ngo. Ngo itself, the current Chinese form, is no longer found in Tibet, but its former existence there and its antiquity are attested by the Abor-Miri ngo, Lepcha > 22 ETHNOLOGY OF THD INDO-PACIFIC ISLANDS. and Sunwar go, Tiberkhad geo, Milchanang and Sumehu gu, which cannot have been directly derived from the Chinese ngo, ngu, nzoi, gu &e. In like manner the prevalence of e and i forms in the eastern Takpa, in Mikir, in some of the Naga dialeets, in Tunglhu, in some of the Nipal dialects and in Tiberkhad shows that they were widely diffused at an ancient period, and that they cannot be referred to the modern spread of Bhotian across the Himalayas. In a previous page, while adverting to the difficulty of distinguishing between the Dravirian and Tibetan forms in e, 2, I observed that in Dravirian the slender forms had been produced by the incorporation of a possessive i, e, with the pronoun, while the Tibeto-Ultraindian appeared to have incorporated a plural 4, and were hence regularly or most commonly found in the plural only. The Takpa, Kinawari, Tibetan and Mikir e forms [ attri- buted to a purely phonetic substitution of ¢ fora. Itis hardly possible to decide to what extent the variations may be simply phonetic as in the Chinese ngoi, ngai, ngei, but it certainly ap- pears probable that in Tibetan the variation of nga to nge, ngi was originally an assimilative or incorporative plural form analogous to the Scythic. As Tibetan has also a ni, i, possessive postfix a similar variation may have been also possessive as in Scythic and Dravirian. The Kinawari Bhotian nga singular, ne¢ pl. suggests that ne was an archaic pl. form, and its honorific use in the Tibetan Bhotian nged might be explained in accordance with this, the use of “we” for “1” being the most prevalent honorific idiom in the Ist pronoun. The Serpa nga s.ni-rang pl, Gurung nga s., ngi-mo pl., Dhimal ka s., ky-el, pos., ki-ng pl., (2d pron. na 2., ni p.), Garo ang s, ning pl., Naga nga s., ni-ma pl. and some of the other forms given in the Table (chap. y. sec. 11) are strongly in favour of the archaic Tibeto-Ultraindian having pos- sessed an incorporative or assimilative plural in ¢, @. The Lhopa nga nom, nge-yi peu (2d pron. chhu 2, chhe-gi p.), shows how possessives might be formed in the same way by the euphonic assimilation of the radical vowel to that of the postfix, Mr Ro- binson gives rang, dag, and kho as other forms of the Ist pron. The Ist is the reflexsive affix (“self”), the 2d is the plural particle, and the 3d is the 3d pronoun used for the Ist. The 2d pro}. ej ETHNOLOGY OF THE INDO-PACIFIO ISLANDS. 203 khyod wri, khe sp. is not the prevalent Chinese ni, na &c. . like the Si-fan and common Ultraindo-Gangetic terms. It appears to be an archaic Chinese or Chino-Scythic form, and a variety of the same Chino-Seythic definitive that is used in Bhotian as the 3rd pron. The Chinese ju, jo, may be a variety of the same archaic form. The Sokpa broad form of the Scythic 2d pronoun chha is identical with the Newari chha and the Kiranti kana pre- serves the same vowel,* The Bhotian khe has the common slen- der vowel of Scythic (i, ¢) and the same vowel is found in the N. E, Tibetan dialect of Thochu kwe, in Limbu khene and in Gurung ken which preserve the Scythic pronominal postfix (comp. Yakuti -ghen, -gen, -kex and the current Scythic sen, sin, sina &c.)° The Thochu has also a form in a, kwa, corresponding with the Sokpa, Newar and Kiranti. In the Bhotian khyod the root iskhyo or kho, T being a common Bhotian augment Kinawari has keof, kherung in which the root is keo, khe. Serpa has khyo, khye. The -d, -t, of Bhotian and Kinawari is a postfix not found in the 2d pron. of other languages of the alliance, and only occurring in the Ist pron. in the exceptional Bhotian honorific nyed, Ladak and Kinawari Bhotian pl. net, in both of which it may be a form of the plural dental and sibilant postfix (comp. en-eshe Tiberkhad). In khyod it may be merely a Tibetan augment or postfixed defini- tive, corresponding with the -s, -t, -d, - found in some of the Tibeto-Ultraindian varieties of the Chinese numerals, If not a mere phonetic augment, it would appear to be a Scythic trait and to correspond with the nasal postfix of the Scythic pronominal system which occurs in Limbu khe-ne, Kiranti kha-na, and Gurung ke-m. A similar postfix. was probably current in the Tibetan dialect from which this form of the 2d pron. was derived. The Bhotian d, ¢ may be merely a variation of n. In one ofthe Samoiede dialects a similar replacement of the Scythic -x by -d takes place, to-di (comp. the Ugrian to-n &c).t+ * In a later page the connection of the Sokpa pronoun with the Scythic on the one side and the Wibeto- Himalayan on the other is examine:!, + The current Chino-Tiberan numeral 2 is the same liquid def. that forms the 2nd pron., li, ni, urh &e. But in the Tibetan 8 (4, 2) a form of 2 occurs which has 4 derable resemblance to the Tibetan 2nd pron. gyud gyet, gye, in Lepela keu, It may however be merely the current 2 nyis, Lepcha nyet, Burman uhit with the g- pref. and nelided. Ifgyu, gye, heu be a Chinese def. it is similar to the unit preserved in 9 kiu, kyeu &c. The existence ofan urchuic Chinese dia- Ject in which the same def, slightly varied or dou'led entered into 2 as well as 1, would be in accordance with Seythie and N. E. Asian anology, B 204 ETHNOLOGY OF THE INDO-PACIFIC I8LaNDS. The 3d pron. ts kho, khu, in Lhopa kho, Serpa khwo, Lepcha heu, Limbu khune, Magar hos, hoch. The regular vowel is o, u and this alone distinguishes the root from the more prevalent form of the 2nd pron. which has e as its proper vowel. The only analo- gous pronouns in the adjaeent languages are the Thochu kwan, and tha-cha and the Sokpa and Gyami tha. All these varieties are Chinese. T’ha is current in Kwan-hwa ; ki in Shanghai (also gi) and Chio-hu, and in the contracted form { it is common to Kwan-hwa, Shanghai, Tie-chia and Hok- kién; ku is Kwan-hwa and khui Kwang-tung. Ke, keu, keue [comp. Lepeha peu] are other varieties. The Bhotian khu, kho and the allied Tibeto-Ultraindian forms are most closely related to khui and this is consistent with the affinities of the numerals which are also in the full archaic Chinese forms best preserved in Kwang-tung and some of the other southern and central Chinese dialects. The dental with the slender yowel is a demonstrative in Shanghai ti, “ this” and Kwang-tung, deng “ that.” “ This" is che in Kwan-hwa, chi in Tie-chiu, chit, chia in Hok-kien and koi in Kwang-tung. Burman like Bhotian has a broad form thu, and Lau has it gatturalised khon as well as in the dental form tan, corresponding with the Changlo dan. The Chinese slender forms are found in Manyak thi, Gurang thi, Murmi the, Kinawari te, Lepcha he, Naga a-te, a-ti, Singpho khi. Thi occurs in Burman also but as a demonstrative “ this.”’ The same root is the pre- valent Scythic 3rd pron. varying tos, h; ta, tam; han; son, zo; ten, teu, ze, se; sin, tida, di, kini &c. Mongolian has e-gun corresponding with khune of Limbu. The dental form is also N. E. Asian, cha-ta Yenis., tun-dal Yukahiri, tana, taan Aino-Kuri- lian, tana (Sanskrit tad) Namollo, tie, tugh Kamschatkan, tsyo, dsee Korea. Japanese has the guttural form kare. The Bhotian root of the 2nd and $rd pronouns may be consi- dered as Chinese and Chino-Scythic, Its use for the 2nd pron. is not Chinese, but Scythic. Possibly it may have displaced the common Chino-Tibetan and Ultraindian root in the 2nd through the influence of Sokpa or another Seythie dialect. Its absence in all the Tibeto-Ultraindian dialects save Bhotian and the few Himalayan dialects that have been much affected by Bhotian, is in favour of its having always been confined to that dialect and of ETHNOLOGY OF THE INDO-PACIFIC ISLAND@, 205 its not being archaie even init. The identity of the Sokpa and Newar pronouns is also consistent with its recent introduction. But the Bhotian form, its connection with the 3rd and the archaic Chinese character of the latter, make it probable that the west Tibetan system is archaically connected with Scythic as well as Chinese. The other evidences of an archaic connection with Scythic are too numerous to render the presence of a 2nd pron. analagous to the Scythic anomalous, That the connection be- tween the 2nd and 3rd pronouns is Scythic will appear by com- paring the Thochu kwa, kwe 2nd, kwan, tha-cha 3d; Bhot. khycd, khe 2nd, kho, ku 3rd; Lhopachhu 2nd, kho 3rd; Lepcha hau 2d, heu 3rd; Serpa khyo 2d, khwo3d; Limbu khe-ne 2nd, khu-ne 3rd; Kiranti kha-na (Newar chha) 2d, mo-ko 8rd ; with the Mongolian chha 2nd, tha 3rd of Sokpa; the Turkish ghen; -ken,-gen (verbal) 2nd, kini 3rd of Yakuti, sin 2nd, kin, -sin-si, -i 3rd of Osmanli &e; with the Tungusian si 2nd (i, pl.), tche 3rd of Manchu, si, sin, s, 2nd, in, (pl. tin) 3rd of Nyertshmsk ; with the Ugrian sina, sa, si, ton, tin, te, d, t, k, &c 2nd, han, nsa, sa, son, sya, tida, s, si, t, d, ka, ja 3rd; and with the Samoiede tan 2nd, tam, tap 3rd, pu-dar 2nd, pu-da 3rd, -t, -th, -d, -dh, -r, 2nd and ord. The same root is common as a demonstrative, relative, interro- gative and Jocative in the Chinese and Tibeto-Ultraindian vocabu- laries. ‘“ This,” tsz, che Kwan-hwa, ti Shanghai, chi Tie-chiu, chit, chia Hok-kien, koi Kwangtung; cha Thochnu, chi-di Gyar., thu Many., wo-chu Takpa, de, di, re Bhot. &c, kon Limbu, chun Murmi, tho Newar, chun yo Gurung. “ Z’hat,” ki Kwan. hwa, i, ku Shanghai, deng Kwang-tung, tha Thochu, wo-tho Takp. (wa. Hok-kien, pi Kwan-hwa), gua-thu Many., de, re Bhot. &c, khen Limbus ‘“ Who,” shai, shu, sa, si sui, chi chai, titiang &e Chinese. “ Which?” su Thoch., Gyar., Hor., Many., Takp., gang, khangi, ka-di Bhot, &c, kha, ko Kir., kha Murm., gu, su Newar,su Gurung, kos Magar. “ What?” thu Gyar., si Takp. achin Horp,, chi Bhot., khang, kan Bhot., shu, chhu, ta, the, di, de, tigi, hi Himalayan. The guttural forms are Bhoto- Himalayan (Bhotian, Limbu, Kiranti, Murmi, Magar). Forms in # are found in Manyak, Gyarung, and Horpa, as well as in Bhoto- Himalayan. The slender forms in e, ¢, it will beremarked, are also current in Bhotian. 20H ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. The plural postf. -chag has the form -dag with substantives. In Lhopa it is contracted to-cha, and in spoken Tibetan varied to -jo or-njo. It would probably be more correct to consider the final -g as the common Tibetan final augment, corresponding frequently with the softer -ng, n, r of other languages of the alliance, but it may be the guttural Scythic pl. def. as in the Horpa ri-gi. The root cha, da, jo is the widely prevalent Scythico-Tibetan plural particle. Comp. the Scythic forms in t, k, g, d,s, z, ch, r, n, J, all variations of t (ante vol. viii, p. 204), and corresponding with the Chinese tu, su, shu, chu, chung, chai, tang, teng, tse, with the Manyak -du-r, -ju, Bodo -chu-r, Burman -do, -to, Serpa ra-ng, Garo -ra-ng da-ng, Horpa ri-gi, Magar ri-k, Bengali di-g, Tiberkhad a-tu-ng, Kinawari ta-m, ta (in tam-she, ta-she, from the Chinese double pl. tang-tse), Ihe vowel of the spoken Tibe- tan corresponds with the Manyak du, ju, Limbu yu, Bodo chu, Mongolian od, Chinese tu, Burm. to,do. The written form may be referable to the Chinese tang, like the Kinawari and Chang- lo tam, but it is also Mongolian -da ( Buriate) and Manchu ta. . The poss. -ki, gi, kyi, hi, yi is the common Gangetic, Ultrain- dian and N. Indian guttural found also: in Chinese, tih or teik, che, te, ku, ko, kei, koi, keu, ge, e. It occurs in the adjacent Tibetan dialect of Thochu, k. 2. Horpa. Mr Hodgson informs us that the Hor-pa occupy the western half of Northern Tibet, “and also a deal of Little Bucharia and of Songaria, where they are denominated Kao-tse by the Chinese and Ighurs (as would seem) by themselves.” ‘ In southern Tibet there are numerous scattered Mor-pas and Sok-pas as there are many scattered Bod-pas in northern Tibet.’ (p,. p. 122,128). Further on he remarks that on the evidence of his vocabularies the Sokpo of the Tibetans are the Olet or Kalmak Mongolians of Remusat and Klaproth “whilst their confréres the Horpa are almost as evidently Turkish, the Turkish affinity of the latter being inferred, not only from the vocables, but from the complex struc- ture of Horpa verbs and from the quasi Arian physiognomy of the samples he has seem of the Horpa race.” Professor Muller has remarked that by its pronouns and numerals, it is Bhotiya (i. e. Tibeto-Ultraindian) and he has accordingly ranged it provisional- ETHNOLOGY OF THE INDO-PAUIFIC ISLANDS. 207 ly as the most western branch of the Trans-Himalayan dialects of that family. Both pronouns and numerals undoubtedly belong to the derivative Chino-Tibetan system, but they have some pecu- liarities when compared with the other known Tibetan languages. The Ist pron. nga is the Gyarung and Bhotian form of the Chinese. | The 2nd, ni, is not Bhotian and it differs from the Gyarung nan, na, and Manyak no in its possessing the more prevalent of the Chinese forms (ni Kwan-hwa &c, found alsoin Gyami). This form is comparatively rare in the Tibeto-Ultraindian dialects, Takpa i, Dhimal ni, Deoria Chutia a-ni.. Thee, of Namsang Naga and Burman is probably a variation of i. The 3rd pron. vja, vjya (in pl. ei) is peculiar. It appears to be a variety of the Scythic sibilant (and dental) 3rd pron. (comp. Ugr. sya, Sam. di, &c) corresponding with the Magyar ja. The plural postfixes are -ni (Thochu, Sokpa &c, supra p.) ; and -rigi or rigya, the first element of which is either a native varia- tion of ri or the Manchu -ri, while the second is the widely preva- lent -ki &ce (Chinese, Scythic &c). Manchu has a similar pl. -jer-gi. ‘The possessive is formed by an elongation of the vowel of the root, ngaa, nii, vjaa, an idiom, the same as the Newar locative (“ in,” on”) and analogous to the Bhotian and Garo repetition of the final sound of yocables when used assertively. The prefixual v- of the 3rd pronoun is an example of a usage which is found in other words and is distinctively Tibeto-Ultrain- dian of the cart Bhotian type, 3. Thochu. The Thochu pronouns are :— Ist chi, ka ; 2nd kwa, kwe ; 3rd kwan, tha-cha. Ka (1st)—probably a variation of the common Tibetan nga— is found in Dhimal and in the oblique form of Lepcha. The change from rig to k also takes place in Naga, -ak for -ang. Similar guttural forms are found in Milchanang, Tiberkhad, Naga, Khyeng, Kyan, Silong and Lau. Chi (1st)—recurring in the Newar ji—is a remarkable term as it has no direct or apparent affinity with the Bhotian nga, na, the Chinese ngo &c and is still more remote from the Scythic labial. But it is highly improbable that itis a distinct root. The ch 208 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. appears to be merely a variation of k, for in the plural and dual forms chu-h-lar, chi-ki, che-un the k is absent altogether. In the possessives there is a similar alternation of the two forms, ka-k-chi “mine” chi-k-wk “ our’s"’. These variations give us chu, chi and che, or gutturalising them and adding the current guttural form, ka, ku, ki, ke, analogous to the Ultraindian series, ka, ku, ki, ti, gi, geo, he, and to the Chinese ngai, ngoi, ngu, ngei. The vowel of the 2nd pronoun like the Ist varies from a to e in Thochu. The root occurs in the form ti in Mulung and Tablang, and the guttural forms also take i in Tiberkhad, Milch., Khyeng, Kyan and Silong. Joboko Naga has ke. Some of these forms are plural, and probably the primary i, e, form was plural* 2nd kwa, kwe, “ thou” is similar to the Bhotian khyod, khe (in Himalayan dialects khe, ke, ki, kha &e.) The first of the words given for the 3rd pronoun, kwan, appears to involve the root of the 2nd pron. with final n. In Bhotian as in several other formations the same definitive is a common element in the 2nd and 8rd pronouns, Bhot. written 2 khyod, 3rd kho spoken, 2nd khe, 3rd khu, Serpa 2nd khyo, 3rd khwo. In the last form the vowel has the amplified Thochu form of kwa, kwan. Tha-cha “ he” &c is composed of two vocables or forms of the same root. Tha is Sokpa, Gyami and other Chinese dialects and in the slender form the Chinese. In the forms ta, da, it is also Scythic, Manyak &c, thoi Dhim.,, ate he &c Naga, (thi Bur- man “this,” also Murmi, Gurung, Bhot. demonstrative cha is but another form of the same definitive). It occurs as a variation of the prefix ka, ta, in Ultraindian vocabularies. Thochu has three plural postfixes, which occur both separately and conjoined asin some Scythic pronominal systems -ni (Sokpa, Horpa -ni, Manchu -ri, Horpa ri-gi, Ostiak, Yeniseian, Yukahiri, n, Ultraindo-Gangelic ni, in, li, &c. Da, ir, n &c); ki, ko, ku, k; and -lar. Ki, ik, is Chinese and Scythiec (Chin. ki, Hungarian -ek, Turkish, N. E. Asian.) It recurs in Sunwar -ki. In the Kasia definitive ka sing., ki pl. thei by itself is plural, as in Scythic. Kol has ko, Gond k, g &c. Lar is Turkish lar, ler, Mongol nar, ner, Kol nar. Kwe-ni-ko, kwa-ni-k lar “ ye” are examples of the sin- gle, double and treble plurals. Lar has obviously been the latest © See the preceding remarks on the Bhotian Ist pron, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 909 acquisition as in the Turkish b-iz-ler “we,” s-iz-ler “you'’*. The Dravirian -kal, gal, -ngal,-kulu, Dhimal -galai; Naga -khala, kara combine two of the roots. Z,7 without the kis Dravirian as well as Scythic, and common in Ultraindo-Gangetic languages e. g. Takpa-ra, Abor -lu, Dhimal -al, -cl, Mikir -li, Garo, Miri, Serpa -rang, Bengali -era. These forms and those inn are yaria- tions of the same root. There are two possessives -chi and -k. Chi is Chinese -ti, Serpa -ti, Tengsa Naga -chi, Dravirian -di, -ti &e. The possessive hk,— which may be radically the same as that in ch, t—is Chinese ko, ku, keu &c, Bhotian -kye, -gi, -hi, &c, Takpa~-ku. It is very common in the Ultraindo-Gangetic vocabularies ko, ku, ke, ki, ka &e. Ex. of the Thochu possessives, ka-k-chi “ mine,” kwe -A-chi “thine” tha-k-chi, kwana-k-chi “his,” chi-ku-k “ ours,” kwa-ni- ku-A “ yours,” tha-ku-k “ theirs.” 4, Gyarung. _The Gyarung pronouns are— Ist nga, nga-yo. Horpa, Bhotian, Namsang Naga Kasia, Bur- man; Murmi, Gurung, Magar, Serpa. The Chinese form is ngo found in Abor-Miri and with the consonant guttural’sed in Lep- cha, Sunwar and Milchanang (go, gu). 2d, nan-re, nat. This pronoun like the Manyak no, differs from the Bhotian and Thochu. It is a variety of the Chinese ni (also Horpa) found in the Shanghai dialect na, nong and in the ancient Kwan-hwa nai, nei. It is also Ugrian in different Chinese forms, (nan, nei, num, nyn), and in the slender Chinese and Ugrian forms ni, ni-n it is Dravirian and Australian. The variedies na and nu are also found in Draviro-Australian. The Gyarung forms nan, na are the common Ultraindo-Gangetic ones, nan, (more frequently nang), ngar, nga, na. From the great and wide prevalence of the forms in a, an, they appear to have ee the dissemination of the Manyak variety no. 1 hha oan sean pa Sere Dei px nt ery root occurs twice, m and n, and each time with a different plural postfix, -i, This numeral like 5 preserves no distinct affinity with the lower numbers. In purely quinary systems 6 is a variety of the unit as 5 itself is. But in some of the Scythic systems the scale is terna- ry, 6 being 3 (for 3, 3), and 7 being 1 (for 6,1). As the only term similar to the Chinese in the connected systems is the Kams- chatkan roch, roka of ng-ro-ch, ng-ro-ha, 3, it was probably deriv- ed from a cognate N. E. Asian system. The Chinese luk is identical with this term, while it cannot be referred to the dtirrent terms for 3 or 1, nor to any def. current in the pronominal system. 7. Cur. ts’hih, chhi, ch’hit, tsat, thet, ch’het, t’sih, sit (Gyami chhi) Trs. ? skwibi Manyak, stare Thochu. | With this numeral the current Chinese system departs entirely from the proper Tibetan. The latter has a quinary term for 7, that is 7 is the numeral 2 (from 5,2). The Chinese 7 is not a variety of 2 but of 1 (comp. the full forms chit 1, ch’hit, sit 7). It adheres therefore to the ternary scale and this confirms the conclu- sion that 6 was 3, 3. Japanese, Yukahiri and in N. America Athapas, can have a similar ternary 7, and in some of the Scythic systems the same double definitive is used as the unit in7—Ugrian seitse-man, sis-im, si-m, sata &c. Turkish site, seti, siche, che- dy &e. Tipetan. ku-sh-nes Gyar. (nes 2), z-ne Horpa.. This term is quinary, 2 for 5,2. The etymology of the prefixed sk and 2 must remain for the present uncertain, In the other Tibetan terms there is much irregularity. 1 have placed the Manyak and the Thochu with the Chinese, but the con- nection is doubtful, especially in the case of Manyak. In both the initial sibilant (s-ta-re s-kwi-bi), may have the same origin as that of Gyarung and Horpa, sh-, z. This would confirm the Chinese affinity of the Thochu root, for tais an archaic Tibetan form of the Chinese 1 (tabi Manyak). The guttural in the Manyak s-kwi may be from the initial guttural in some Tibeto-Ultraindian forms off2 (gnyis Bhot., khi Karen). The Bhotian 7, bdun, dun, appears to be a Mongolian engraft- a “ ETHNOLOGY OF THE INDO-PACIFIC ISLANDs, 227 ment (Sokpa tolo, Mong. dolon, &c), . It is probable that it is later than the other Tibetan terms, and displaced a quinary one, for it is only found to the south in the Bhotian dialects of Lhopa, Serpa and Changlo, while the other Himalayan systems connect- ed with the Bhotian have quinary terms similar to the Gyarung. 8. CHin.—pat, pab, pe, boi, poi (Gyami pa). This root has no connection with any of the lower numerals. It cannot there- fore be explained as a native binary (4, 4) or quinary (5, 3) term. Besides quinary terms, several of the Mid and North Asiatic system have terms formed subtractively from 10, In these the root for 2 is frequently alone preserved. Pat however has no connection with the Chinese2. A similar root is 100, pe’, be’, pa’ and the Bhotian 100 appears to be also related to the root for 8, Such a connection would most naturally happen through a labial root for 10, since 100 is very generally expressed like 10 by the unit. In the other systems of Mid and North Asia the labial is a definitive and unit, and it appears at some archaic period to have been a very important root in expressing higher numbers also, as it still is in some of the older systems of the S,E. provinces of the Old World,—Kol, Australian, African. Inthe Scythie and N. E. Asian system as in Chinese the t, s, r, k, &c. def. isnow the chief numeral root, but most of them preserve remnants of labial numerals. Chinese has the labialasa def. (3rd pron. and demons.) under the form pi. In the Scythic and N.E. Asian systems it has still a considerable currency as 1, 5, 10 and 1000. For 1, Ugrian has vaike, va &c, Tungusian emu &c, Turkish bir, [—bis, bit] per, Japan fito ; for 5, Ugrian has vate, vis &c, Turkish bish, besh &e, Iroquois wish, wis &c; for 10, Samoiede has bet, bi, wi, bu, bun, Tungusian men; Turkish wona.. In the Ugrian languages it has been superseded by the dental &e as 10, but it is preserved as an archaic 10 in 8 (2, 10) and 9(1, 10) in the form mis. Aino has wam-bi &e 10, From the occurrence of the labial in the Chinese 100 and 10,000 (wan, ban), its presence in 8 is best explained as an archaic and obsolete unit applied to 10, and 100. The full term was probably similar to the Ugrian 8, kika-mis (2, 10), the Dravirian 9 om-bad (1, 10), and the analogous N. E. Asian and African terms. The Dravirian patta, bad &¢ 10, vodda 1 preserves a broad form of the labial unit similar to the Chinese r 228 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. pat 8, and the Ugrian vate 5, vaike 1. The common Scythiec form is the slender vit, vis, mis, bis, bir&e. [See the remarks on the labial definitive and numeral in the Draviro-Australian, Semi- tico-African and Scythic systems] In some of the Tibeto-Ultraindian dialects a labial 10 and 5 are preserved. Undoubted instances of it are the 10 of Kasia shi- pon (shi 1, Chinese), Limbu thi-bong, Murmi chi-wai, Naga pan, ban, the Chinese form. In the higher numbers of Kasia and Limbu it is 10 (App. A p. 6), and Kumi also has it in the higher numbers apong. The Limbu and Kiranti 9 phanesh, phangya are probably remnants of 1, 10, as the Chinese 8 is of 2, 10 the nume- rals for 1 and 2 having been dropped. The labial occurs in 5 in several languages but in some at least it appears to be prefixual as in 4. In the Chepang pu-ma-zho, Shindu me pa the root is clear- ly ma, pa and it favours the opinion that the Thochu wa is also an archaic labial root. The Bodo pa may perhaps be placed with them, and not with the doubtful bonga Garo, phong Mikir, manga Singpho, banga, pungu, phanga &c Naga, pan Kumi, banga Ma- gar, in some of which at least the root is the Chino-Tibetan nga (comp. Naga pha-li4, pha-nga5), The Murmi chi-wai 10 (i. e. 1, 10) is a compound similar to the Kasia and Limbu 10, but the labial has the form found in the Kambojan ma-pai 20 (i. e. 2, 0), Kumi wai-re 100, Sunwar s-wai-ka 100, The form resembles the Kasia variety of the Kol-Ultraindian 1, wei. The 5 of that system being mon, mona, mun, mo in Kol althongh not in the allied Ultraindian systems, it must be considered doubtful whether the Ultraindo-Gangetic labial 10, 5 &c, are referable to that system or to archaic Chino-Tibetan or Chino-Ultraindian labial numerals. Ultimately the Dravirian, the Scythic and the Chinese Jabial numerals are connected through an archaic Mid or N. S. Asian system. Tis. br-gyud, gye Bhot., or-yet Gyar., rh-iéé Hor., khrare Thochu; zibt Manyak. In the Appendix the presence of 2 in most of the Tibeto-Ultraindian terms for 8 is indicated and they are considered as binary. The Bhotian term is left unexplained. From the Gyarung or-yet, Takpa gyet, it appears that the root is yet, corresponding with ye of the Bhotian gye. In br-g-yud the root must also be yud. This analysis is confirmed by the Himala-. ELHNOLOGY OF THE INDO-PACIFIC ISLANDS. 229 yan and Ultraindo-Gangetic forms g-ya, g-ye, yet-sh (Limbu), yoh, sh-yit (Burm,) ri-yat Mru. &c. In all these forms the constant root is evidently yet &c, and yet itself is a variation of 2 (comp. in Limbu nyet 2, yet 8, Burman nhit 2, sh-yit 8; Horpa nge 2, rhiéé 8; Thochu nga-ri 2, kh-ra-e 8). In the Abor-Miri pu-nit- ko, pi-nye the 2 retains its full nasal form nit, nye. In Appendices A and C, I have considered this as 4 dual or 2d 4, From the analogy of the Chinese, Scythic and Dravirian terms it might be inferred that in all the Tibeto-Ultraindian words for 8 in which 2 is the root or an element, the primary form was 2,10. Prof. Miller has pointed out that the Mikir nir-kep, 8, and chir-kep, 9, are formed from fini, 2, ichi 1, and kep 10, But the initial elements br, rl, or, re, ri in some of the preceding forms and the pre, pra, pla, pi, pu &e of other Ultraindo-Gangetic languages are evidently the re Manyak, hla Horpa, pli Takpa (6zhi Bhot., di Gyar.) of 4.* Similar forms are common in the Ultraindo-Gangetic vocabu- laries (App. Four). The Bhotian br-gyud, Gyar. or-yet and all the cognate terms are thus 4, 2, (i. e. 4 the 2nd time). The Manyak zibi appears to preserve the root for 4 only, in its primary Bhotian and Chinese form zhi, si, But for the analogy of the other languages and the occurrence of zyi in 40, it might be explained as a native quinary term (5, 3), 3 being sibi while 2 is nabi. The formation of 8 from 4 is found in Yukahiri, Japanese and * If the Bhotian br stood alone it would be considered as a mere def. prefix, similar double prefixes being used with some other words. It is noticeable that it is not the current Bhotian dzhi; but that the numeral was at one time cur- rent in Tibet as bri, bre &c is evident from the Takpa and Gurung pli, Bodo, bre, Garo bri, Murmi bli, (in 8 pre) Magar buli, Newar pi, Lepcha phali, Chepang loi-zho, Kiranti la-ya (re-ya in 8), Mikir eet Li yk ec, Singpho meli aga deli, pili, phali, Kami mali, Kumi palu, hindu puli, Sak pri, and e radical hla Horpa, re Manyak, le Sunwar, lish Limbu, lika Kuki, thi Khyeng, li¢ Tunghlu pi Newar, phi Chungle, a-pi-ko Miri. It is probable from this wide prevalence o the form in Si-lan-Ultraindian yocabularies that it was current for 4 and entered into the compound for 8 in the system of one of the more dominant and dispersive Si-fan tribes, It may have been communicated by it to Bhotian, but it is quite pos- sible that both bzhi and bri forms were current as 4in Bhotian dialects. The form gyud, gyet for 2 appears to bealsoa Bhotian dialectic variation. Bhotian is very prone to liquid augments, and in the current 2 gnyis w. nyi o. the Chinesen of ni becomes ny. In the Manyak and Gyarung forms na, nes, the augment isabsent. Gyet is evidently from a diulectic variation of gnyis, contracted y the ues sion of the nasal and the conversion of the final sibilant into a dental. The Lep- cha full form nyet, probably an immediate derivative from the Bhotian dialect in question, and the Takpa gyet 8, and Gyarung oe i 8, are also referable to it and not to the native forma of 2 (nai Takpa, kanes Gyar.) The spoken Bhotian gye preserves the same form contracted, The written gyud is a secondary dialectic variation, the original vowel being i gnyis, nyi as in Chinese, e o 230 ETUNOLOGY OF THE INDO-PACIFIC ISLANDS. some of the Ugrian and Samoiede systems (e. g. Sam. sin-ilet, from side, 2, and tet, 4). 9. Cun. kyeu, Reon, kiu, kan (Gyam. chyn); Trs. dyn, guh, gu Bhot.; ung-gu Gyar., gubi Many., go Hor., rgure Thochu, ’ The root is probably the unit in the guttural form found in the Mid and N. Asian systems as a variation of ch, t,s&e. It occurs in these systems in 9 by itself or with a root for 10 (i. e. 1,10; or 1 with 10 elided). Ugrian has ok-mys (1,10), aktse (akt 1), &c, Japan ko-konoz, Koriak, Yukahiri, chona, chonai, (Kamsch. koni 1, Namollo kule 1). The first vowel of the Chinese is the same as that of chit 1, of the pronouns and demonstratives ki, ti, chi &c, and of the cognate Ugrian unit ik, it &e. But Chinese has also broad forms. The def. ku is used as a 8rd pron. in Kwan-hwa, and under the amplified form khui in Kwang-tung, in Shanghai it is “that”, in Kwantung under the form koi, this.” 10, Cun. shi’, shih, ship, sip, chap, tap, chap, zeh, (G@yami ish). Tis. sih Gyar., che-chi-bi Manvak (che, 1, a Chinese form, i. e. 1,10) bchu, chuh Bhot. The shi, si, ta, cha, chi, che, of this term is the def. used for 1, The labia! final may be a mere phonetic aug ment, but some of the Ultraindo-Gangetic forms are suggestive of its being a remnant of the labial unit used as 10. Kasia thi-pon, Limbu thi-bon, The final labial has been lost in the Tibetan terms, but it is found in Mikir kep, Kiranti kip, and Chepang gyib-zho. If the labial be neither a mere augment nor a separate root in, Chinese, it may be a remnant of the def. postf. like m in sam 3, and thus be indirectly eonnected with the Manyak chi-bi. The Horpa sga (ska in higher numbers) appears to be a broad form of cha. ‘Fhe s appears to be prefixual as in z-ne 7, (Gyar. si-nes) in 9 of Bodo ckku, and Garo shkuj and in the other Ultraindo-Gangetic higher numbers which have ta-, cha-, tha-, sa- &c. The Garo s-kang 10 has the Horpa form. ' The Thochu hadure is probably a corresponding form hadu with the pref. aspirated as in the Kami hasuh, and the root with the Bhotian vowel (chu, in the Changlo ], thu,). ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 231 The prefixes and postfixes of the Tibetan systems,—Bhot. g- (1, 2, 3), d- (9), b- (4, 7, 8,10) ; Ggprung ha-, hu-, hung-; k-, hh-, kha-, ha-, r-[=+d- Bhot.] ; Manyak -b%, Thochnu -re,-ri,— are not of Chinese origin. They belong to the Scythic and proto- Scythic (Yeniseian, N. E, Asian, Caucasian) connection of the formation, and have been added to the Chinese roots. The Manyak and Thochu in the regular use of a qualitive postfix are Tibeto-Scythic. 'The Gyami -ku is the Chinese segregative. The segregatives vary with the class of the substantives enumerated and not with the numeral. The Tibetan systems present some of those irregalarities which evince the long prevalence and partial blending of different dialects, but with the exception of the Bhotian 7, all the numerals are refer- able to the Chinese system. Close representatives are current of most of the Chinese numerals, not in the modern diffusive forms of the Kwan-hwa found in Gyami, but in the forms in which they are still preserved in the least abraded Chinese dialects as the Kwang- tung. It is probable, however, that some of the variations from these forms are not purely local, but are archaic Chino-Tibetan, and indicate the existence in China of more than one dialectic system of numerals when they were first spread westward into the Tibetan province. ‘The Tibetan 7 and 8 must have been derived from a dialect distinct from the single one which now prevails throughout all the Chinese provinces. They are pure Chinese in roots, but the one is quinary 2 (for 5, 2), and the other binary 2, 4, whereas the current Chinese is ternary in 7 (1 for G, 1), and apparently denary (10 for 1, 10) in 8, Both Chinese and Tibetan are denary in 9, As all these methods are found in the other numeral systems of Eastern Asia, and as the union ofall tribes of China into one nation is a historical event, it is probable that in archaic times several similar divergent systems existed in the Chino-Tibetan region. The firs® introduction of Chinese numerals into Tibet may be equally ancient with that of the pronouns and definitives, which also show some dialectic variations of an archaic Scythic kind. In other words, the tribes that gave a Chinese formation to Tibet may not have separated from the cognate Chinese tribes till some at least of the numerals were in use. 232 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, When we test the Chino-Tibetan numerals by their relationship amongst themselves and jy the current definitives, they are found to be less regular and homogenous than many of the other systems of Asia, Africa and Asonesia. Many of the Scythie and N. E. Asian systems are less disorganised. But in these, irregularities of the same kind occur, and the Chino-Tibetan system, if considered as only the last remnant of several dialects that existed from a very remote era and borrowed from each other, will take its place with those Seythic ones which have been most changed by a similar cause. The liability of numerals to be displaced by the roots and forms of other dialects is fully illustrated in the sections on the Draviro-Australian, Semitico-African, Indo-European and N. E. Asian numerals, and even in the limited Tibetan field we have ' found some examples. Thus inGyarung 2 has one Bhotian form, nes, in 2, and another, yet, in 8; while 4 has a native variation di, in 4, the Takpa form pli in 40, and a third variation, or, in 8. Manyak has one variation of the Chinese 4 in 4 re, but pre- serves the common Chino-Bhotian form in 8 zi, and 40, zyi; it has a peculiar form of the Chinese 1 in 1 ta, but possesses the Chino-Tibetan in 10, chi. The archaic Chinese numeral systems were evideutly closely related to the archaic Scythie or proto-Seythic. ‘They were not mere derivatives of the Scythic nor the converse. They go back to the period when the Asiatic systems were little dispersed geogra- phically, and some of the extant forms resemble those of the remoter Scythoid languages—as those of N. E. Asia,—and those found in formations of which the connection with Scythic is very archaic,—as the Caucasian and Dravirian. The roots are all or nearly all current as definitives, and both the definitive and numeral systems of Chinese proper are remarkable for the secondary rank which the labial holds. But there are strong grounds for believing that in the primary ems of the Chinese glossaries, as in those of the more advanced formations, it held at least an equal place with the dental &c, The Australo-Kol, the African, the Dravirian, the Scythic and N, E. Asian, and the Chinese, illustrate various stages in the decadence of the labial. The monosyllabic dialects that haye been transmitted in the basis ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 233 of the Indo-Australian and African glossaries probably separated from the Mid-Asiatic linguistic provinge before the dental began to predominate as a definitive and unit. It may be remarked that languages and formations that have lost the labial as a 3d pronoun preserve it as a demonstrative, and even when it iz no longer current as a demonstrative, it sometimes lingers as an interrogative, rela- tive &e. The existing Chinese has doubtless suffered great changes during the period in which the various harmonic formations have been developed and dispersed, and these changes must have been chiefly glossarial. It is” consistent with the history of all forma- tions that primary or archaic vocables and forms should sometimes be found best preserved in those languages and families that were earliest removed from the primitive ethnic location. In the con- tinued mutual linguistic influence of the East Asiatic tribes, Chinese and Scythic, changes have probably taken place in the glos- saries of all the less secluded nations, from which the Dravirian, Asonesian, African and American remain free. Sec. 5, THE MISCELLANEOUS GLOSSARIAL APFINITIES OP THE TIDETAN DIALECTS AMONGST THEMSELVES AND WITH CHINESE AND SOYTHTIC. A glance at Mr Hodgson’s tables shows that the Tibetan vocabularies are all intimately connected. Comparing the west- ern or Bhotian with the eastern or Si-fan we find that in the list of GO or rather 58 miscellaneous vocables,* Bhotian has about 24 in common with Thochu, 33 with Gyarung, and 26 with Manyak. The agreement is thus from 30 to 60 per cent. The adjacent Horpa has 36 of the 58 words Bhotian. Of 59 Bhotian vocables only 7 are not found in any of the other Tibetan vocabularies (8, 24, 30, 41, 45, 46,50). Of the remainder, 7 are found in all the other vocabularies (2, 7, 20, 26, 42, 48, 51); 3 in Horpa, Thochu and Gyarung (15, 27, 56) ; 1 in Horpa, Thochu and Manyak (8); 2 in Horpa and Thochu (1, 25); 7 in Horpa, Gyarung and Manyak (6, 14, 19, 29, 37, 38, 54); 6 in Horpa and Gyarung (12, 17, 21, 36, 40, 60); 5 in Horpa and Manyak, (22, 28, 32, 33, 52) ; 5 in Horpa (16, 23, 39, 44, 55); 2 in Thochu, Gyarung and Manyak (31, 47) ; * For the words corresponding with the numbers see Vocabulary te, p. 183. In some of the Tibetan lists two and even more words are defiatent, © fa 234 ETHNOLOGY OF THE INDO-PACIFIC IsLaAN DS, 4 in Thochu and Gyarung (11, 13, 57, 59); 1 in Thochu and Manyak (34); 3 in Thochu (9, 43, 49); 1 in Gyarung and Manyak (4); 3 in Gyarung (18, 35, 53); and 2 in Manyak (5, 10). The Chinese affinities with the Tibetan vocabularies collectively are considerable, About 31 of the Chinese vocables in the list are found in one or more of the Tibetan vocabularies (3, 4, 5, 6, 10, 12, 13, 15, 16, 17, 18, 20, 21, 22, 23, 25, 26, 33, 35, 39, 40, 44, 47, 52,55). With single vocabularies the agreement is much smaller. Bhotian has about 14 Chinese words, Horpa 10, Thochu 8, Gyarang 12 and Manyak 6. * The extent to which the same Chinese vocable has been diffused or preserved in several of the Tibetan dialects may be seen from the subjoined statement which, like those that follow it, is not to be considered as minutely accurate, the object and the value of comparisons on so limited a scale not rendering rigid precision worth the labour of attaining it. Several of the Chinese words are obviously of modern importation, a consequence of the great political and social influence the Chinese have long enjoyed in Tibet and their constant intercourse with the Tibetans. The proportion of voca- bles archaically common to the two families cannot be ascertained, without larger and more exact comparisons, but many of the common words in the list are certainly archaic. Some are found as roots with variable forms and meanings in all the 8. E. Asian vocabularies. E Of the 60 words Chinese has 1 in common with Bhotian, Horpa, Thochu, Gyarung and Manyak (20); 1 with B., G. and M. (47); 1 with B., H. and G.(17); 1 with B., H. and M. (4); 1 with B., T. and G. (13); 4 with B. and H. (23, 25, 39, 40); 2 with B. and G. (37, 59); 3 with B. (6, 33, 49); 1 with H., T. and M. (10); 1 with H., T. and G. (18); 1 with H. (5) » 3 with T. and G. (15, 18, 55); 3 with T. (22. 35, 44); 5 with G. (5, 12, 16, 26, 52); and 2 with M. (21, 55) ;—being 25 vocables in all. The Seythic ingredient is much larger and more important than the Chinese. Bhotian has at least 29 or 80 Scythic roots in 78, that is about 40 per cent (1, 5, 7, 10, 12, 14, 15, 19, 20, 21, 25,— two roots, three if one common to Chinese be included—26, 27, ETHNOLOGY OF THE INDO-PACIFIO ISLANDS. 235 29, 33, 35, 37, 38, 39—two roots,— 43, 46, 48, 50, 52, 53, 54, 56). Morpa in about 60 words has 20 of these Seythico-Bhotian voea- bles and 9 other Scythic’roots (30,034, 35, 43, 46, 50, 58, 57, 59) making about 50 per cent. Thochu is much more independent of Bhotian and Horpa in the range of its Scythic affinitives than these are of each other. It has only about 9 of the Bhoto-Scythic roots, but it has 16 others, some of which occur in other Si-fan vocabula- ries or in Horpa, although most are peculiar. Gyarung has 6 non- ,Bhotian Scythie vocables (1, 23, 39, 43, 49, 52) and most of them are Turkish. The connection thus indicated must be modern com- pared with that in which the Tibetan affinities with Samoiede, Fin and other remote Scythic languages originated. The special in- flaence of Turkish on the Gyarung is farther shown by the Turkish forms found in those Ultraindian vocabularies that are most allied to Gyarung. The Turkish words have frequently slender vowels, e. g. il mind, Gyar., Burman &c li, le air; tin, night, G. to-di; diri, tire, stin, G. ti-dri. Manyak has 4 Turkish words, 8 Mongolian and about 9 more remote and non-Bhotian Seythic. All the vocabularies have afew Turkish and Mongolian terms, the close accordance of which with forms current in these groups, indicates that they have been communicated to the Tibetan tribes by their Tartar neighbours during the latest era of Scythic history or that in which the Turks and Mongols have marched with the Tibetans and spread themselves into their province. The Turkish words are more numerous than the Mongolian and this is proba- bly to be ascribed to the fact of Turkish predominance in the northern borders of Tibet having preceded the Mongolian and endured for a muchlonger period. The numerous Turkish forms in Ultraindian and Asonesian vocabularies corroborate this infer- efice. There are a few Tungusian terms but it is doubtful if they are to be distinguished from the general mass of Scythic words, which form a large and essential ingredient in all the Tibetan voca- bularies, These Scythic roots are archaic and they are in general found in remote N. and N. E, Asian vocabularies. They are chief- ly Ugrian (Yeniseian, Samoiede, Ugrian proper, Fin), but some are also Yukahiri, Aino-Kurilian and Kamschatkan. This class of affinities may embrace eras as long as all the later @ > 236 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. ones (Chinese, Mongolian, Turkish) down to the present time, but we must in the actual state of ethnology be content to refer all these remote affinities to one nebulous archaic period which we may term the Ugro-Kurilian or simply the Ugrian. Further research will probably distinguish the Samoiede, the Yeniseian &c from more ancient affinities. A considerable portion of these archaic affini- ties embrace also Iranian, Caucasian, Semitic and African langua- ges. From their forming so high a percentage, and being the most important of all the ingredients of the Tibetan vocabularies,, they clearly connect the history of the Tibetans with that of the ancient Ugrian race, which prior to the predominance of the Tatar branch appears to have spread not only over the whole breadth of Asia and Europe from Kamschatka and Korea to Lapland, but to India, Irania, the Caucaso-Semitic province and N. Africa, for their vocables are abundantly dispersed over this wide region in languages belonging to various formations. So great must be the antiquity of this cardinal ethnic movement that the origin of the Tranian formation itself in its Scythic basis, may be referred with probability to it. The Mid-Asian affinitics of Iranian are Ugrian much more than Tatar. The large Scythic ingredient in the Tibetan vocabularies whea taken in connection with the Scythic character of the ideology, re- duces the enquiry into the more archaic history of the formation to this,—were the Tibetan languages originally Scythic or were they crude monosyllabic tongues akin to Chinese? To answer this.ques- tion we must take the position and character of the Burman branch of the alliance into account, and it leads us to the conclusion that the archaic or pre-Ugrian languages of the Tibeto-Chinese province were closely allied to the Chinese and the crude proto-Scythic ; and that they were partially transformed by Scythic nomades advancing into the province and blending with the native tribes, after Scythic had acquired its harmonic and inversive character. At the same time many of the common roots must be considered as of equal antiquity in Tibeto-Burman and Scythic. The Mon- Anam race was probably identical with the ancient Tibeto-Burman, for there was hardly room for another between them, and the languages have some non-Chinese traits in common, as the posi- tion of the qualitive after the substantive, the use of prefixed or ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 237 preposed definitives, besides possessing many common roots. It is probable that the Mon-Anam was at a comparatively early period pushed to the southward, although not before it had receiv- ed a gonsiderable portion of Scythic vocables. The Tibetan miscellaneous vocabulary, like the pronouns, and the general ideologic character of the formation, show that it is Chino-Scythic. An examination of the vocabularies separately gives the fol- lowing results. In the Bhotian list we find about 14 vocables with Chinese affinities; 6 with Turkish; 3 with Tungusian; 20 with more remote Scythic and N.#E, Asian languages which may be termed Ugro-Karilian ; and 18 which I class as peculiar, simply because I have not ascertained any foreign affinities, but many of whieh will probably prove to be Ugro-Kurilian. The Horpa vocabulary differs little from the Bhotian, at least 36 of the 58 words are Bhotian, and 4 of the others are also Bhotian in root; 4 are Chinese (besides 6 which are Bhotian also, making 10); and 18 are neither Bhotian nor Chinese, although 4 of them have Bhotian affinities. Several of the others are Scythic. The Bhotian vocables have, in general, the same form as in Bhotian, but they are softer. Thus rog ant, phag hog, metog flower, lag hand, discard the fina! g. In some eases the Horpa form is broader, e. g. ram horn, mah fire. Most of the Horpa forms are found in the Si-fan or Ultraindo-Gangetic vocabularies. S-gre star, is a slender form found in Burman kre, the Bhotian and Manyak being s-kar-ma, krah. Phiri snake is a similar slender form of the Bhotian s-brul, Manyak bru, Takpa mrai. It is also Thochu bri-g¢ and Gyar. kha-bri. Where the Horpa form differs from the Bhotian and has special Si-fan or southern affinities, these are indicated in the subjoined list. « Thochu has 24 or 26 words in common with Bhotian, and 3 with Chinese in addition to 5 Bhoto-Chinese. Of the 35 remain- ing vocables at least 13 (4, 19, 21, 23, 25, a and b, 26, 27, 30, 37, 38, 40, 46) are Scythic. They are nearly all archaic, that is they are not derivatives from the adjacent Mongolian or Turkish, but belong to the primary Scythico-Tibetan stock. Some preserve > 2:38 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. forms now found in the more remote or sequestered branches of the Seythic and N. E. Asian family, Samoiede, Yeniseian, Aino &e, When to these we add the Scythic affinities of the Bhoto- Thochu words it will be seen how slight the Chinese gijssarial ingredient is when compared with the Scythic. The Thochu forms of the common roots differ considerably from the Bhotian. They are frequently slender and curt, e. g. 7 ri, B. rus, ru; 15 zi, B. sa; 26 pi, B. phag; 31 ki’, B. khyim; 47 ri, B. lam; 48 che’, B. chha; 49 pi, B. pag; 51 bri, B. brul. In some cases the Thochu forms resemble the Manyak and _ not the intermediate Gyarung. Sky, mah-toy ma’; Stone, ghol-opi, wobi; Blood sa’, sha’; Goat, tsah, tsah; Light uik, wa’; Salt, che’, che ; ?Skin ra-pi, g-ra. The vocabulary has numerous southern affinities, but fewer than Gyarung and Manyak. The Gyarung list has 33 words ia common with Bhotian, in- cluding Bhoto-Chinese words. 5 with Chinese not found in Bhotian (besides 7 Bhoto-Chinese) 4 with Turkish, only 2 appa- rently with Ugro-Kurilian which are not Bhotian also, and 16 peculiar in the above sense. The Manyak list has 26 Bhotian words, 3 Chinese (besides 3 Bhoto-Chinese) 4 Turkish, 3 Mongolian, 9 Ugro-Kurilian and 14 peculiar. ‘ I proceed to illustrate the preceding statements by some details. The Bhotian words in the list of 60 miscellaneous terms, which as some have synonyms and others differ in the old or written and the current or spoken dialects, amount to 78, may be arranged under five classes. J%rst,—Words that are apparently peculiar to Bhotian. These amount to about 18 or 23 per cent of the whole, but as there must be many Mid and North Asiatic vocabularies, not collected or not accessible to me, and as even Klaproth’s want some of the terms in the list, it is probable that this proportion would be much reduced by a more ample collation of vocabularies. Second, ——Words having affinities with Chinese, mostly archaic, but one or two appear to have been received from it since the Chinese spread into Tibet. These amount to about 14 (18 percent). Third,—Tuar - kish words, probably derived from the Turkish hordes during their 2000 years of contiguity and partial intermixtare with the Bhotians and only amounting to 5, one being Mongolian as well as Turkish. ETHNOLOGY OF THE INDO-PACIFIC Mrisns. 239 Fourth,—3 Tungusian terms, probably archaic Scythic. Fi/th,— archaic Scythic or Upper Asian. These vocables amount to 29 or 30, without reckoning those Scythi¢ words which are Tungusian or archaic Chinese, which would give 5 more. According as we includeWer exclude the latter the percentage will be 37 or 44, in the last case about double that of the apparently peculiar Bhotian vocables. COMPARATIVE VOCABULARY OF BHOTIAN. a. Chinese. 4 Bird,—Spoken dialeet chya. (Sunwar chiva), Chin. chio &c. 6. Boat:—Spoken, syen; Naga ihseng, Ch. ch’hiang. 15. Dog :—khyi; (com. Tibeto-Ult.); Ch. khiau (Burm. chhang, shen), I7. Elephant :—glang-chen, Ch. chhiang. 20. Fire.— me, mi, ma; Ch, we. 18. Hye:—mig, mik; Ch. mok. 28. Foot:—kang; Ch. kha. 25. Hair:—pu; Ch. bo (Fin has up, but as there are two other Tibetan synonyms for “ hair,” kra and ta, pu is probably Chinese), 33. Leafi—s. hyo; Ch. hio. 389. Mountain ;—ri; Ch. lia; Tungus. alin (allied forms in Mong. and Fin) (a). 40. Afouth :—kha; Ch, khau. 47. Road:—lam, lan; (Newar lon, Sunwar la) Ch, lu, lau. 49. Skin :—pag (b); Ch. phi, phue. 57. Z'ree:—jon-shing; Ch. shi, chang &c. (also Kamsch., Yenis., Sam., Cauc.) 59. Water: chhu; Ch. chui, shui &e., (Sam., Ugr., Tatar, Afr.) b. Lurkish. 30, LHorse:—ta; T. at. 34. Light:—hod; T. syod. 389. Mountain:—West Tib, dak; T. tak, Japan dake. 48. Night: -—tshan, chen; TT. achsham, Mong, suni; allied forms in Semitic, Malagasy &c. 46, Hiver:—tsang, chang; T. usun, sug; Mong. chun, usun &c. Ugr. jugan, Pashtu sean, sin, sint. e. Tungusian. 10. Cow:—s. pha chuk; Tung: chyukun. 12. Day:—nyin; T. ininy, manyi. "54. Sun :—nyi (See “* Day’’), d. Ugro-Kurilian. 1. Atr:—lung. The Ugro-Scythic forms of this root when used for “air are slender, but the Ugrian lux ‘ day” preserves the (a) Takpa ri, Horpa ri-rhap, (b) Tekpa phyekh. ; 240 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. full Tibetan form. 5. Blood:—khrag; Ugr. wuorak ; Saumali, Galla dik, diga (the root is also current as water.) 7. Bone:— rus-pa, ru-ko, lu, lu-4, lush-am, lush-yn, by Ugro-Fin, luy, by Sam: Semitic alam, alat, Pashtu Iu, ro, ra; re, alukei, Dray. yelu &c, Lesgian. 10. Cow—lang; Finlehmu; Cauc. al; Galla, Amh. lam &e. 14. Har:—na; Caue., en, in (Face, Mouth, Nose &e in Seythic). 15. Harth:—sa; Sam. -ja; Zend sa; Horpa zi-p, Gya- rung se’, Ch. ti, Jap. zi, tsi, 19. Puther—pha; Sam. &e kc (com. ) 20. [Fire:—ma, me; Aino abe, Chin. we; (Jap., Ugr., Tumali, Malagasy &e have allied forms.)] 21. Fish ;—nga, nya; the guttural form ka of Naga, Anam, Mon &c appears to connect the root with the Fin kal, Sam. kual. [24. Goat:—ra; Semiti- co-Aftican aron, illa &e.] 25. Hair:—kra: Fin karw; Aino karnu. 25. (Hair:—pu; Fin up, Ch. bo]. 25. Hair :—ta; Fin ata. 26. Hand:—lag; Ost. lagal Turk. dik. 27. Head,— go; Ostiakog, [28, Hog:—phag; Iranian, Drav.] 33. Leap ; —loma; Fin lopa. 35. Man:—mi; Fin mis; Galla mi; Zend memio. 937. Afoon: -la, da; Koria oru, Cbhuk. iraluk, Sam. iri. 38. Mfother:—ama; Yukahiri, Yenis, Fin, Samoide and com- mon in other alliances, 40, [Mouth:—kha; Ch, khau, Yenis. ko, gou, khan, Sam. ake &¢ Semitic kho] 48, Sal¢:—tsha, chha; Ugr. sow, sal &c, Sam. si, sir, sak &c (also Iran., Semit., Afr.) 650. Sky:—nam; Sam. nom, num, nob. (Ugr., Semit. Kashm.) 52. Star:—kar; Korea kurome, Koriak agor, (Iran., Afr.) 53. Stone:—do; Korea tu; Ost. to. [55. Tiger :—tag, tak; Iran. tigrts &e] 56. Tooth:—so; Japan cha, ha. In Ugro- Fin the sibilant root is Head, Eye, Hair, Mouth, Ear. The Bhotian so is closest to the Fin su ALouth (Chinese sui &c). The root is Tooth: in:Caucaso-Semilic and Afiican vocabularies” dsa &e Cire., silay zul-we, sol-wol &e Lese., sin, sin-on; kc. Semitic (sil, sin is Hye in Seythic, sun AMou'h, shun Lar &e.) ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, APPENDIX TO CHAP. V. OF PART II, A. COMPARATIVE VOCABULARY OF THE NUMERALS OF THE DRAVIRIAD FORMATION, One. a. Onru Tam.; onna, wunnu M., onji Tul., Tad. ; undi Gond., unta,; Uraon; ort, ondong Male. (Root, on). b. Tel. vokati, Tod. vodda. ; On &c. is a very archaic Asiatic particle, and current in many lan- guages as a definitive or article and unit. It is chiefly prevalent at the western and eastern extremities of the Old World. Exrope:—en, oino, wena, one(wan), Afriea:—uan Berber, yean Shillah, wani Fetu, na Jbo, na-ya Kashna, inni-he Danakil ond Amharic, adde Tigre, inye Kalir. N. £, Asia:—Samoide viol, Koriak onen &e., Kamch koni (Air. akun &e.) {t has also male its way into America, being found in some of the Sioux and Californian languages :—wan-che, yon-kai, na-ngu &e., henai San Raph. (Kamch). Tn the Mongal ni-ge, ni is probably a pref. (int comp, Danakil inni-ke), The Tungusian and Samoide ujnon, om &¢. may be related to on, butit is more yrobable that they are connected with the labial definitive. To the east and south of the Dra- virian region it is found in the Lau nung, ning, (also a demonstrative), Nicobar eng, and in several Asonesian languages. The exceptional Telugu vokati, Tod vodda, is a rare but widely scattered combination of two common numeral and delinite elements (Comp. afok Tumbuktu; veike, veit Ugrian; fiio Japan; woto Kaili (Celebes), motu, wakal Australia.) . Twe. : Trandu Tam; randw Mal.; yeda Tod. (root elided); yeradu, randu,’ erad Tul.: erudw Kar.; rendh, yeradu, ‘Tel.; ranu, Gond; e-no-tan, Urao; (irat, Brahui). The root is evidently ra, er, tr, with or without a picfixed vowel, nd being an euphonic augment of the definitive postf. This definitive (varied by the change of the consonant tod, J, n §c) is one of the most prevalent terms for 2in Asian and Europe (Iranian), In N. Asia itisrare, but r is an element in Mongol, Tungusian and Koriak terms, Jy, identical with the Draviriar, is the North ' hinese term, In the original Iranian, d is combined with a labial definitive, and the few E. African terms are evidently of Iranian origin (Danakil, Galla, Malagasi and its numerous Asonesian derivatives). The Georgian yern, deur, ori, shiri “ two" probably presents the original of the N. Asiatic . and Dravirian terms, because it is regularly formed trom 1, to which indeed the ru, ri &e properly belongs. The form in ‘‘ one” ar, is still closer to the Vravirian, ar being an inversion of ra. Three. Murr Tam.; munnar, muna Mal.; min, mudu, Tod.; muji Tul.5> murz Kar.; muds Tel.; muse Gond; ma-no-tam Uraon; (musif Brae hui, Comp, muji Tul), (Root, mw.) This numerai (the labial det, ma, wa, ba, va &e,) is not Iranian or Se mitic, Butitis Caucasian, N, Asiatic, and African, Caue, mi (combin; . - 2 ATHNOLOGY OF THE INDO-PACIFIC ISLANDS: ed with s. of 2).; Ugrian Im, rm (combined with kof 2); Japan mi (uncombined), It is vot found in the other N. and NE. Asiatic sys- tems, and it is rare in America (Catawbas na-mun-da, Sahaptin mi-fat, Shoshoni manu-thit?, Sioux dialects na-mena, Jaw-b wah, yah-mani). It has also spread into Africa, where it is combined with other particles as in Cauc, & Ugr., but it is not prevalent. - gat?" . Nangu, nalu Tam; nal« Mal., nonk Tud., nalu Tul, nalke Kar., nalugu Aare nalu Gond, (Root probably na, and gu, ku a second def. osttix. : This term is Dot Iranian or Semitic, (but a is an element in the Sem. 4) One of the Caucasian terms approaches to it, unukwgu (Leszian), But the clearest and most vumerons affinities are Ugrian and Airican. Ugr. nila ke., Afr. na, no, ni, nani & ». (Mandingo &e.), ne-nvl ( Bullom), inan (Moko), mani (Kosah) &e, From the distribution of the term in Africa and Asia, and the mode in which thayelements are interwoven into some of the African systems, it is probable that na &c. was the radix of a bi- nary system belonging toa formation that predominated in S. W. Asia be to the epochs of the diffusion of the Caucasian, the Semitic and the anian. It doesnot appear to have made much progtess to the East- ward, agit is not found in the existing N. E. Asiatic an the adjacent N. American languages. In the latter n is a frequent initial, but it is merely definitive, being found in 3 and other numbers. If the root is na, and Ja, le is the poss. postf., the final gu, hu, k must be'a superadded definitive. Double and even treble definitive postfixes or prefixes are not infrequent in the agglomerative formations of Asia, Europe, Africa, Asonesia and America, and the Dravirian vocabularies siipply instances, The combination of de and gumay therefor be purely Dravirian in its origin. Gu, ku, &c. is a common definitive final in . some of the Caucasian languages, and others which do not use it have im- porte it as a substantive portion of numerals which they have borrowed om the, former. The Georgian es-gu, 1, is an instance, the proper Georgian unit being ar, with or without thi. The Caucasian affinities of the Dravirian numerals raise the suspicion that the final gu of the \atter had a similar Lesgian origin. The parent system was doubtless formed in 8. W, Asia, the great focus of all the Asian, European and African nu- meral systems; and the present Caucasian numerals are probably them- selves derivative trom some archaic formation that was not confined to Caucasus and did not originate there, for the plain of the Euphrates, and not its head basin or the vallies beyond, is likely to have been the earli- est seat of civilisation in this region, Five. Aindu, anju, anchu, Tam., anja, anchu Mal., yaij, yajjkhu Tod., ayine, Tul., ayidu, eidu Kar., ayidu T'el., eaiykhan Gond. (Root ain or an pro- bably, but possibly anj, anc). This is a peculiar term. Ifthe root is ain, or ai, it appears to be a mere flexion ot'na, 4, Ifanj §e. be the root, the only affinities I can disvover ure the Caucasian inshéu, itself an inflected combination ot defi- nitives, and the Iranian panchan dc. with the connected words signify- ing “hand” (e. g. yash, chsy, Ugrian; siv, tsu, chu, sesu, Chinese, Ultraind. Ind ; hasta, seste, hath, Iran.) If the numeral was not derived from Cauc. or Iran,, but was a direet modification ofa word for “hand,” the WTHNOLOGY OF THE INDO-PACTIFIC ISLANDS. 3 nenrest term is the Tungusian hanya, the ancient prevalence of which is attested by the European hend. Six. Aru, Tam,, Kar., Tel., ava Mal, orr Tod., aji Tal. Gond. (Root @ pro- bably, a-ru, @-ji, but perhaps ar.) . From 5 a new series of simple definitive terms appears to commence, In this respect the Dravirian system) is very remarkable, the allied Asiatic and African systems forming the higher numbers by inflecting or combining the lower, adding them to 5, or subtracting them trom 10. Lcannot but suspect that the Dravirian terms are really disguised inflexions of the iower ones. If this is the case aru must be referred to ira, era, 2, ani! be considered as ®remnant of an original binary system (2, 4, 6, 8), a conjecture countenanced by the great prevalence of such ‘a system in the Old Workd, and the frequent formation of 6 irom 4 or 2, Seve. Ezh, ezhu Tam., ezu, yezhea Mal., yele, eZ Tul., yo Tod, elu, yelu Kar, edu Tel. yenu, yetu pend. Ifaisthe root of 6, ¢istheroot of 7 du, tu and zhu being the def. or poss. postf. The Tamil zh, (Frenchj) +d. /in the other languages. At present Iam inctined to consider e as a flexion of a, or e-du of a-ru. : Eight. Efiu, vettu Tam,, etfa, Mal. yeta, ett Tod., ename, Tol, entu, yenéu Kar., enimidit Tel, © The root appears to be en, et, joined in some of the terms, if not inall, to another particle. | incline to consider the 'erm as formed from 2, 10, by agg!urination, 4 common mode of forming 8 in allied, Asiatic lan. guages (Uyrian &c.) If this is the case, en, ef, represents 2 (erd, ren, yed or en) and the definitive finals represent 10, by vocalie reflection or harmony. Kar. 8 en-tu. 10 pa-iu; Tel. S enimi-di, 10 padi; Mal. 8 eta, 10 pas ta The imi and ame of Telugu and Tuluva may signify “ short of’, ** Jess’?, orsome other term indicating that the number is 2 shor! of J0, or they may merely be the ba of 10 transformed by the phonetic action of the adjacent sounds. . Thus inthe highly barmonie Tel. ¢-om-midi 10, may have been ¢-on-bidi, ¢-om-bids, and then more euphonically ‘-om-midi, Nine. On-badu, on-bakudu Tam., om-bade@ Mal,, oram-bo Tol,, om-bed Tod, om-bhattu Kar , t-om-midt Tel, These words are 1, 10, i. © Lshortof 10. The Telugu and Tuluva prefix and infix a definitive in om 1. (on modified by 4) The Telogu term is formed like that for 8, and as om is unequivocally 1, it slrongly confirms the opinion that 8 is formed from 2, Ten. Pafia, orn-patudu (one, ten™®) Tam,, psifz Mal., poét Tod,, patie Tol., pattu, hotiw Kar., padi Tel, a | The root is eviden|'y pa, the poss. pnstf, changing in Telugu to di, the pranominal form of iis poss, pestf, With or without the postf., it isa com. mon N. Astaticand African term. Af. pu, fir, own, awo, ful, fule Sc, met, (Coptic) ; Samoiede bu, wi, di, bet (Coptic,) iat “ ~ 4 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, B. COMPARATIVE VOCABULARY OF MISCELLANEOUS WORDS OF THE DRAVIRIAN FORMATION, 1. Air. a, kal, A. Tam., cali, ghait; katta, katfu, koft, Tam. Mal. Tod Tf the k, g, of the first term is prefixual and li, la, the root, it is identical with the Lrawadi, li &e., Karen thi, Burm. le, Yuma da-li, a li, hi &e , which is also Seythic and Semitic. The other posttix tu, ta, ¢ is also root tor “Air,” “Wind,’’—Gond, bariba (H.) ita; Male, take, taphe ; Uraon tha- ka, (Comp. Asonesian 0, gato ée., Fin, at, Armen. ot &c.) There could not well be a more striking instance of the extreme caution required in comparing monosyllabic roots, Al first sight the Dravirian galt, kal aut the Yuma /@li appear to be identical, and even the rejection of the Yuma prefix ‘a jeaves ample room to contend that the root i: is at once Ultraindian, Dravirian and Scythico-Semitic, I think, it is clear, however that Ji, d, tu &e., are the ordinary Dravirian postfixes and that ka, ga common tothe two terms, both of which are found in Tamil (Kal A. kaliu Mod.), is the Dravirian root. As ¢and k are frequently inter- changed itis probabiy identical with the Vindyan dental form,—Male, taklie, daphe; Uraon, thuka: Gond hariba ida. Beyond India the dental form is common as a term for *‘ Air” or Wind” and the guttural very rare, Polynesivn ¢o, Tilanjang tiu; Mille gato, Tobi gufam, which may be the dental with a definitive prefix, or direct derivatives trom the Dra- virian kala It is the most common Scythic root, at, uta, wet, wola Wogul, but, in general, it takes a postfix e.g utlun Chuk oddun Tung, tol, tyil, tuuli&e., Ugrian, Turkish, with which Comp. the African dolah, Darjur, The same root reappears in the Armenian ot, Sanskrit vata, atima, Zend watem, Bengali datas, Lat vevtus, Persian, Hind. bad, Asonesian adi (Sabimba). In Maly badi is applied to the spirit of a person or animal that has been kijled, aud which, it is supposed, will take vengeance on the slayer it a charm is not used to lay it (See a charm nsed by elephant killers; Journ. Ind Arch, I. 316.) In Malay badai isa gust of wind; Kagayan padak ; Australian padru (Bathurst) (Sansk, badra), Binna badara, Malay &e. udara. Vhe Nicobar twfutis robably the same root with the prefix ku, and in the Malagasi rivotra, denesian ribut &e, “ storm,” “hurricane”, it appears to be com- bined with another wide spread root for “ air’, “* wind’” (ri, H, above referred to.) The only guttural forms I find are Australian’ kira, kirrar; New Guinea, giriks ; Woloff quelo; Fin gai-so; Japan &¢., ka se; Mongol kei; Mon kia; Nega tikhe. b. elaru Kar, A, This resembles the Ultraindian term (which see for the foreign affivties), but it may ve the common Dray, root a. wilh the initial consonant elided., c. ambaram Telug. amayum Mal, Korea, patam, paron; Ugrian, warma,*barsko, marwezh (Wol!s,); barshi, merga, merz &c.; Koriak, walkatt; Bods bar, Gar hal, (see Ultrain. dian, iis a common Asonesian root. From ‘he commutability of J, r audd it is probably radically the same as bad &c, (8), Pavanam, weiws, Telugn Wind” asliiri wav, Pashtu wah, Newar phaiz Ende wabo, Australian wibi, . J Berio ec? ‘OF Tim tvdh-PAciwico TezAnps. 5 wepe; Quilimani pevo, Suabili paipo, epipe, Galla bube, afoaw, Danakil haha; Koriak claugui; Semitic hawa; Kol hoyo, (see Ultr.) 2. Ant. a, uray, erumnbu, irumbu, irivi, erd (the root appears to be 7i, ru &c, the m being a common euphonic augment before b,) Uluraind. Jang, rang, miling &c. Azonesion :—Bis. langam, sulum, Lamp serem, Sund. irum, Aru areram, Pol. !o, lelo. The same root is probal)'y contained in the Celebesian biri, bere. 6 Tel. chime, Tul. pijin, ; Gangetico.Ultraind. chiji, chimechi &c, Hind, cheumta. o. Arrow. ‘a kanet Tam, A. : . Pashai, Sindhi, kan; Asam kanv, (See M. A.); Koreng, éakyen, Anam ten, ' J, ambu, amba, ammu, ‘Lhis closely resembles the Mishmi, qmpu &c,, but the JItraindian root is pu, pun &c, and the Dravirian apparently am, if itis native. It appears, however, to be E. African, Comp, Suabili mpamba, Makua izpsmba, in which /, 2m are prefixual, The root pamba is probably connected with the SMaeaas wana and its Asonesian derivatives, and also with the Sanskrit ana. ec. Tul, biru (‘* bow,”? Kar. bilu, Tam, vil, Tel. vileambu, Mal, velu,) Sanek, pilu. cd. satalu Kar. A. Although ancient it is Arian. Vindy. seri, char &e., Naga lasaz, lahan, san, Kapui then, Siam Juk.son; Sansk., Pali, Bengali, shara, saro, shar &c,; Tungus, ser-dan, Koria sar,; Georg, isari, 4. Bird. a, pul, paravci, parua, pull, . | Angami Nag, para; Andi purtie; Hind. parinda, English, bird; Galla gfala “@; Majagasi vorons; Asones, burong &c. ‘The vliimate monosyllabic root is doubilegs the same vs the Tibeto Ultraindian byu, bu, ve &c.,ibut there is appwrent!y nv direcé connection between the latter and the Dra- virian, ‘She 2, 7, r@ is probably the definitive, b. paki, pita, hak. Bengali pakhyi, Sansk pataka, Hind pakheru; Daphla pata, Aka putah, Abor petang, patang. Fin pitic, poita (Sansk. palaka), Ason.—Binua, pake, Lawp. puti, piaiu (“fowl”), Pasir io. The Indian and Indonesian form is probably pre-Arian, The root appears to be the same as inc. 5.. Blood. a, chore Mal. chore Kure, oe choma; Gaug, Ultr, chi, chai, sai, asu &c.; Cire. sha, teha, Ke . Olah, b. sennir, Tam. A.; kennirss Kar. A.; khens Ur.; kes, Male; (? Garo ieaerehat.) Su, si, chai, Ccc., is the preceding Gangetico, Ultraindian root, and kan, ke &c, may be merely the Ultr. pref., Hind, khon; Turk. kan &c.; Fin kem; Aino kim &e.; scou,—Nicob. kanak, Austral. kwuain, komaya. 7 c, notturu, nettar Telog. Tul.; nattyr Gond,; ( Lau lent, lut, let; son —Sunda let, Trusan elod), d, udiram Tam, (Sansk. rudira], G ETHNOLOGY OF THE INDO-PACIFIC ISLANDS) 6. Boat. a. pakada, padava, para, vanji, [Bengali pansi) valam. I place these ferins together because they appear to contain a Common root, pa, and it muy be considered vacertain Low wach of what folluws it is merely deli. nirive; ka-da, da-va, n-ji, l-am appear to me to be double posifixes of the ordiuary kivod. Probably two distinet terns, however originally formed, have long prevailed, lst, pak, found in A, ‘I'am. only, and, 204, the more Pravalcat uae, par, vat orvan Both have been carried to Asonesia,—avaul Pagzai; pada waka, Celeb,; vaka, poki, Polynes; wangga Viti; venau, wenau Timor; bangka Balignini; uwang Magind,; wangkang Mal &o, The Indones. pratu, peau, Pol, falaa, Viti velo, may be from the Dravirian pacu or the Ultraiadian p'luung (Gurung playa) Pa, the ultimate root of the Dravirian terms, exists in the Mille, Tarawa and New Caledonia wa, Gar Nicob. ap , Gorontalo bu, Nias bubu, - 6 doni, Kar. M, This name is-widely spread along the coasts of the Tndian Ocean and Ul rainiia. Africa ;—Saumali donie, donah; Danakil deuniki. Sindhi, Jiengali, dingi; Anam, ding; Chinese (Teo chew) ting saimpan (canoe); Kol. dunga, denza; Murmi, Newar, Magar, Sunwar, dunga, donga; Aka hu\ang; Garo, reng; Naga, Manip. Kum,, surang, rung, plaung &c., Mon #leng, galon &e. &e., Asonesia,—tina Sumba, Solor, tid ng Kis. (tis Jong Mikir), ‘The stender form ding, ting, bas spread along he coasts. ‘The broad form dung, dong, rung &¢. appears to have a distinct history In the Gangetic valley the Bengali dingi has superseded the ancient West Ulitaindian dunga, donga, still preserved in the Himalayas and Vindyas, 7. Bone. enpu, elumbu, ella, emika, elume, eluvu, elu. Tibeto-Uliraind. ruba, along, rang, aru, aro, Ason.—\oh, lolor, &e, Semitic alam, alot, ailathir; Galla Jala; Pashtu alukei.; Lesgi, root ra, ro, re, lu. 8. Buffuloe. a. karan, kara, kera. Tam. A., Kol. Kambojan kar-bu, kra-bo, Chong, ka pao, Ka, kar.pu; Indoneg, karhau, karabao, kabu &c, ‘I he same root is probably found in gour, gor, and the Beng. and Hind. name of the wild buffaloe, arna, appears tu be a contrac. tion of the Dravirian karan, b. ervma, enumu, erme &e, ‘ Ultraind. le, reh, oi, ‘The root is a common one as applied to the *‘cow,” Seythic, Tibetan, African. 9, Cat, a, pusei, Tamil A. epuchcha, Mal., puchche Tuluv,, .pusi Kol,’ pusi, Pashtu, Sindhi, pasha Kashgar; Milchanang pishi; Kapwi pishty pishik,; Fin misak, matska &e., Mong. michoi; Afiican,—ius, musune, topisa; Ason.—pusa, Born., Phil; Semitic, bis; Eogish, puss, ‘The postfixed root is probably the Egyptian chai, shai, preserved une‘\ored #in the Mongo! mi-choi, Naga mo-chi, Bodo muji, ‘The other root mi_ pi, PMs isalsocommon, Inthe Tibeian simi the p sition of the two rovs ig reversed. “ 6. beku, biku Kar., probably the root bi, be (see a) with the pos'f ku. A/fr.—paka, Kilim; Ason.—bika, Buton, ‘The Buton term is Dravirian, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 7 10. Cow. a avu, Tel Kar., a,Tam, A, | The rvol, a, is probably archaiecally connected with the Tibelto.U/train. dian ba, wa &e. (which see), bul it may be connected with the Lesgi a ka, a-la, a |, (seed ) b. petam, peta, pasu, pasuvu, hasuva payya; Singhal. vesi The ultimate rout is probably pa, pe, ve . T. U. bo, pha, wa &c., a8 in a, Comp, Himalayan, pit, bik, bit, bi; Dhimal pia; Karen phi, bing, Khumi bhi (buffaloe), But the Dravirian. Himalayan forms, from their possession of the final dental or sibilant, con. nect themselves directly not with the ‘Tibeto-Ultraindian but with the Seythico Iranian mes, misye, mis, wanch &c,, (Ugran); machala; Tungus; bos Latin; maas Kosah &c The Chepiou, Bode, Garo and Naga terms are more purely Scythe than the South Indian,—moshya Ch., inashu, mushu, tWed,, mashu Garo, masi Naga, ‘The same root isa common one for buffuloe, ec. lanma Tuda (Perhaps tan has been adopted from petam Ult, tom Kapni, atom Ma- ram.) d. akalu Kar. Caue. aka, ata (Lesgi), Probably d,is connected with qa, and c. and both with the Vindyau udu &c, Il. Crow. Kaka, kaki, kagi, kak &. Uraon khakha, Male kake, Magar kag, Singpho kokha. ‘I'his reduplicated form is perhaps connected with the Tibetan khata, but it is one of the most widely spread imitative words. Kha, ka occurs alone and with a second root er postfix ra, wa in Kol and Gond (kava, kawa, kahu), in the Himalayas (kalwa Kir, ahwa Limb., kawa Murm, ku New.) Anam konkwa. In Kumiwa occure alone Jn Asonesian he Dravirian reduplicated form is common, kaka, gaga, gagak &e, 12. Day. a, pagal, pagil &c., Gond pati; ‘* Sun,” paka-lon. Lesgi bigula, “Sun” bak; Ason.—Australian baga, baga-rin ‘sun’ ; Indo. hesian pagi &c. “morning” ; b. el A. Tam, ullah. Uraon ; F Milch. lai; Korea, lai; Koriak, alo; Arm. or, Iran, elere, uras, hari &c,, Ason.—ilah, alu, ira, &c. &c. Indon; la, ra Pol. 13. Dog. a nayi, naya; Toda noi, Gond, nai; Male allay; Ur, alla, Nayi, allay, &c. appear to be contractions of the full ancient Indian form preserved in the Himalayas and Ultraindia,—nangi, nagi, Murmi, nagyu Gurang, neko Mishmi; and its antiquity is proved by its being found in Australian, nagi, nago, as well as in Savo, nguka. Jt is Upper Asian, and the terms prevalent there, | ke the Himalayan, leave it uncer- tain whether the root is nak &c. orka &e. As na is a Scythic and Himalayo-Ultraindian prefix, and kwi &e. is very prevalent in Chi- nese, Tibetan, Himalayan, Ultraindian and Asonesian vocabalaries, if is probable that ka &c. is a distinct root, whether ma be merely pre- fixnal ora root also. All three occur in Upper Asia—inu, in, Aino., Japan; kaj Korea, nokoi Mong.; koi, choi, Lesgi; nyni, neaaki, nenakio S ETHNOLOGY OY THE INDO-PACIFIC ISLANDS. am &e, Tangus. The Asonesian iru, uli, ali (Indones, Aust. Pol.) appear te be contractions of kuli &e, 6. kukka Telug, Kokkur, katta &e., Beng. Hind, Sindh, but the term is found in Indo. hesia, koko, kito, gida; in Australia, kota, and New Zealand; and it is also Koriak and African, The root kui &c, is still more widely spread, 14. Lar. a, sevi, cheyi, kebi, kemi, kivi, kavi, kada; Gond, kavi (Todava, kavi) It is dovbtful whether se, cle is a distinct root, or merely a moflificas tion of ke, as appears most probable, If it is a separate root, it may be connected with the Tibetan sa. The guttural is found in Bodo khome Deor Ch, gaku, Ultraindian naka, cakor, anless these.are derivatives from the Sanskrit karna, (in Telinga karnam] Hind. kan, with which the Drav, is remotely connected. The root is widely difused. Baraki goi, other Afghanistan vocabularies kan, kad, khad (Hind.); Fin, Samoiede, ko, kun, kav &o. Korea kui; Tungos, kunya; Samoid, kuma; Tork, ku Ja &e,; Yenesei, kolo-gen &¢.; Africa,—gura, guru Galla. The prevalent Australian kara, kure. guri, appears to be conneeted with the African ra« ther than with the Dravirian form, but as Ja,ra 4c. is a common postf. in the Australian ae inthe Dravirian formation, the root may be of Dra-« virian origin, ‘Ihe Georg. kuri is close to it, 15. Earth. a. nilam, nela, nelen, The Ultraindian ali, le, lai, may be connected with this root, The Khamti and Tai-lung nin is near it. The New Guinea ena, Polynes. one, may be derivatives. The root is common,—libeian, Tungusian na; Chukchi nuna; Egyptian an; Mid:-Alrican enesh &e, b. pudavi, podavi; pulova Singhal. . Lepeba pat; Ason—batat Buol; hutang Kis.; budjor Austral. 16. Eyg. a ejinei Tam. A.; Yenes. shulei, Semoied. sarnu, heny; Tork. simct, semurtia; Mong! chara &c,; Chin, ch’hon; Cauc, bono, b. muttei, mutta, motto, mukshe. Simang maku, If ‘be root ix mu, ma, as is probable, itis connected with the Mon-An. pu, pung dc. (which see), ce gudda, cw, tatti, totti. 17. Elephant. a, kolirs; ane, ana, an, eniga Singhal. alia. war, Dhim. naria; Kamb., tam-rai, Ka kanai ; Burm. ane, ne, Indon. gas riya, 18. Eye. a. nattam A, Tam. [prob. from Sanskrit}. ; Kamceh. eleth ; Koriak, ilet. Jalat, &c, Sansk. netram, Pali, netra; Africon,—Tum. nget, Danak, enti, inte, Malagas inte ‘‘see,”; Indones. inte, intei, ‘‘see,”; Indon. nihat, lihat “see,” Binua “eye”. [See T, U. and Mon.-An. | : 0, kan, kon, kannu, &c.; Gond., kank, Male kane, Urao.-khan, Bras ui, kan. ‘Chin, gan ; Tork. kara’: ; Latin oculus; Asones..Pol. kano, o STHNOLOGY OF THE INDO-PACIFIC r8LANDS, ‘5 19. Father. @. endei, tandet, tande, M. A.—Kamb. ta, Anam thei; Tork ata, atai &c.; Chok. atta; Jap. titi, Urg atta, tuata &c.; Iranian Caue. atta, tata, dady, dad ke; Afri- can,—Mak. tete, atiti (Jap.) Egypt atf., (atai chief). dson —Ii don, tata. na, tatai, utha, (tua “fold”, “chief” &e. isa modificaticn of the same root), 5. appa, Kol. apung. Himal.—Ultr. appa, abu, abo, &e. Tib. pha, apa; Monapa&e; Ko- giak apa; The root pa, ba, &c. is almost universal. ¢. amme, Tuongus. ami &.; Cauc. emen, ima; Georg, mama, (Drav. “uneln”,) muma,mu. The root ma &c. is common, but generally applied to ‘Moe ther”. d. achcham. . Sam. esem, ese; Jap. tei; Ugr. isi &c.; Turk, asio, Mong. ezeges Ghara ais [see Mon. An } e, eyyan. This term is perhaps a softening of d,, but there are simi« Jar Asiatic and Asonesian roots,—aya, ayu Kc. — 20. Five. @. azhai. A. Sam. shu, siu &c.; Tork. oth, ot &c.; Caue. za, tee, mza, mze Kes ‘Oset. sing, Pers. seng; Georg. zezchii, Hind. atash, shuala; Semitic, asat, isat &e.; Asones :—Sim. us., Komr, husok. nerupu, nirpo, nippu. Arabic, nar; (Kashm. Hind., nar.) ce. tiyya, tu. Tork ut &c.; Ugr. tu, tni; Tungus. toh, tua; Ason:—Indon. tui, tauo Cee uta, “heat” New Guinea; uda, New Caledonia. (see a.) . benki. Singpho, Nag, Gar. van, ver, wal; Asones.—Pagai vange. e. kechcho; Gond. kis. Male chiche, Ur. chik. Yenes, khott; Tork. “heat,” kus, kos, kusu; Asones.—Pol. kasa, J. Kol singil, sengel, (‘‘Sun”, “Day” singi, See a, sing, seng,) Fish. @. puzhal. (? Iran. matsya, piscis, visch, fish Se. Cire. psis; but the Drav. root is probably pu ) b. min, minu, Gondi, Male min, (Sausk,) Chong mel. c. chepa. Sam, chale; Yenes. ise; Ugr. zen, &e.; Cauc, chua, psis, besuro, cha« ve &e.; Arm. tzugn, Semitico-African,—asa, esa, said ; Malagasi hazan. Ason —Indon. isa, asan, tsi, irda &c. (Africo-Sem., Malagasi.) 22, Fiower. a. @darA Tam. (?G. U) ; é. pu, puva, povu, buvuy Male pup. Ur. phup. Kol, buba, baha, bowh. G. U. Sunw. phe, Limb. phung, Kar. bungwai; Ultr. pun, pu &e. Chin. fa, we; Semitic pol, ful &c., African wab, pau &c., Malagasi yong, vuna &c. (Ason. vona, bunga &c.) 23. Foot. a. kazhal A. Tam., kata Kol. Kas, kajat, Karen kha, kho-du, Tangk. akho, ake, Kumi. akok, wkiuk ; Cauc, kok, kog, kash ; Oss, kach, Georg. kuchehi, Lasi, kassi, ig ELBNOLOGY OF THE INDO-PACIFIC IsLAND™ kiuchehe, kuskas Afghanistan,—khn Peees Ason,-Indon. pany kaja, kacha &c. . b. adi, ori, adugu, hajji, hejje Abor, ale ; Cauc. rori (? Aiico - Asonesia tina, dina, tana &c. seo Mon.-An.) : ( ¢ kal; Gondi kal¥, kev. : - Tib. kang, Kashm. kor, kwar, Hind. gor ; ;7Ason. -kokor (tatah.) 24, Goat. a. vyellei, yalladu. Ifthe root is ve, va, it is connected with the Ultraindian de, pe &ee (Egyptian and African, be) which has spread into Indonesia, ‘The im mediate affinities are Atrican,—teel. Ambharsi, ipuri. Makua, imbuhri Kosah, Aaviri, Malagasi, dubila Danakil, iwureh Yoruba &c, But the root is a Caucasian and [ranian one for ‘sheep’, Caucasian mell, mall, het; Bengali mera, bhera, Hind, bher bhera, Sanak. Tndonesian biri. 6 meka (see a?) | ‘@. edu, adu, adr; Uraon era. This ia one of the roots for “‘cow’’. d. kuri; Male kre, Ultraindian mikreh, makre Tangk,; probably derived from the Hind, at 25. Hair. a. kuzhal, A. Tam., kudalu. Caue. kodi; Tib. kra, Singp. kara b. mayir, mir, tala-mudi (tala is "head?) ; (Hind. mar) Koren muri; Austral. mori. a ventruka, d, orama ; ‘Ost. warras; Georg. weve; Arm, law ; Ason.—Anstr. uranj ¢.) tali Male (tala, “head” Dray, ) 26. Hand, a, tol. Sam. utal, udam; Kameh. tono; Yenes. ton; Turk. adem; rea! taalo; Afr. tan, tana’ &e.: .;, Ason.—tong, ting, tangan &e. . b, "kal, kayya, kayi, cheyi;-Gond, kait; Ur, Ahakhah. | 7 Tangkul akhui, kuit, Bodo khai, Khumi akhu ; (Mon. Anam, tn amb deih ) N. Asiatic, Scythic, Cauc. , Iranian, Ultraindian &o, kata, i 2 kar, kak, kuik, kua; &c. “27. Head. @. senni: Malagas. saino ; Sansk. shira, Cauc.’sh’ha, 5. talei, tala, tale, tare; Gond., talla, Magar mi talu, Kir. tan ng 5 Malagasi ins Rotuma thilu; Lasi, ti. ec. mande, mudd (Hind,) . 28. Hog. a. kezhal A.'T.; Male, Ur. kis, , ~ Cire, kashha, kcho, (root choa, cho) cha, .kha, ka, peti re Os, Pers. chug, Arm. chos); Malagasi kisoa; Mong.. kachai; Semitic pis Breton guis; Ason.—kuis (Batan.) pant, panni, pandi, handi, P ji, poti; Gond, paddi. . (t elup fune, Serakoli bule.) - he Dray, root appears to be pa (a. BTHUNOLOGY* OF THE INDO-PACIFIC ISHANDS, Ti a being frequently inserted before a, r, 7,). If so, itis Himalayan, Ultra- indian, Alrican and Asonesian, a) 5 -, sed, geT eet. kodu, kombu, komba, kommu, kurr. Sunw. guro, Him,-Ultr kung, gong &e., Chin. ka (Ultraind kau &c.;) Ugr,y keku;,Semitico.-Alric. gnng, garong; Iran. garn, cornu «ke, ; Lasi. akra, Georg. nka; Ason.-Sumba kado, 30 Horse. a. payime. This term is confined to A. Tamil. It ia probably con- nected with one of the words for “Cow,” payya Mal., Kurgi. _ — b kudire, kadira, kudure, kudare, kudre, kadar. oa ih Beng. ghota, Pashai ghoda, Deer got, Tirhai kurra ; Yenesei, kut, kus, kon; Ug. kon; Sam. kunde, yuda; Bokhara ghunt; Cancas. kotu, kooto (Andi); Asoa—kuda, The other Scythic form of the root Aon, is found, in the Manipuri diwlects,—chakon Koreng, Maram, sagol, Champh. It corresponds with the Mizjegi gaur, gour, Hind, ghora, Naga kor, Changlo korla &¢. { Welsh gorwydd.] | 31. House, a. illam, illu, ila, arra; Gond ron, Ur. erpa, Kol. ora, oa ; (Him-Ultr.}; Jap. ire; Iran. alaya (Sansk,) aula (Latin) &c.; Afric.—arre Danak. ile Yorub, Ason.—Simang hale, Pol. hari, Rotnma ri (Lepcha_li). 6. manei, mane. (?Circ., wuna,) . ce. Vida, uidu; Male avu. : 32. Tron. a. karumbon, irumdbu, irumla, tinumu, karba, (Him-Ultr. yogir, yagarah, &c.) ) Caue. ger. Afr.—Haus. karufa, karife, Shangalla sio-kar; Iran. iron, The Drav. root appears to be rua or ru, ne, with or without pre- fixes and postfixes. It occurs in Samoide ur, Milchanang run, rung (a nasal being trequently posttixetl'in this language.) Hind, &c. bite kabina, Karn. panna Urao, (the bon of ka-rum-bon may bé the samé root.) Kol merhil, merbad, medh, marhan; Lepcha panjing; Georg. beresh; Afr.—Galla sibila, Saumali bir, Danak. ‘birtr, Galla beret.’ [Georg. beresh, Kol merhd.] Amh. berut, Malag. vi, Egypt ba; Ivan, jertus ;,Ason.—New Guinea puruti (2, Alr.); binaku Buton, bunjil © Pamp,, pungal Tobi. é | < 4 < valle 33. Leaf: - a, adei, elei, ela. ele, ire, err, ela, Male atge, Ur. athha; Bodo lai, Maram alui (see Tib-Ultr.) b. aku, Gond aki. Chin, ge; Turk’ kaak 5; Ugr. kor, kuar &e. Cauc, g’a, Bie “34, Light. a. oli Av Tam’ (7b.)° a b. velicham, veluturu, belaku, pele ; Gond berachi, Ur. billi, Mal, aveli; Dhimalwayal; Manip, D. war (seeG. U.) "ee. ¢. bhaksha Tul, | a. O y auine qh Bhi 35. Man. a makana; ganasu. ; Lau kon, khun; Ugr. watan, kaiam, kom; Yukahiri, kunshi; Cauc. konachoi; Afr,—s isatil umake “husband”; Shaygalla, gunya (Yukah,);* Ason,—Kuiit, tame, kandka, kanchu; Ach, ekum “ husband, ¢ il 12 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, : b. al, alu; Ur. alla, Kol. horro, horh, ho; Mon karu, kru, Ka hloe Burm, lu; Turk ir, iri, er, arini; Ugrian alma ulmo &c,; Cane. olo- chan-chi; Arm. air; Germ, er; Afr —Mak. alo ana, Malagasi lahi, olona, Egyp, rumo &e, Ason.—Austral leah, leh, iure; Born are. 36. Monkey. a kaduvan, kurangu, koranga, koti, kadoge, kodan Bodo Mothara, Arabic kirthi, Mabra garat; Afr.—golo. Ason—Indon., kara, kra, gere, kodeh, gudeh, kate, kita, b manga, mange; Male muge, New. moko (Murm, mang &c, Hind. maimun). , Ason.—monyit, mona, amo, &e, : 37. Moon. a. pirei A. Tam; Male bilpei. , Bodo, nokha-bir; Canc. ports, bars, bers; Afr.—Galla, Tigre werhe, Danak. berra, Felup. fylein, Malagas. yolana ; dson,—vola, bulan &c. 6b. tingal, tingale Korea tal, Ug. tilaye, toles, tilos, Samoide diri. c. neln, (d. aabilli, (The root dil is a.) . 38. Mother. a. inral, Turk. ini &e.; Tungus, enei &c.; Fin ene ; Cauc. ila, illi, ennin, ninn, nana, ana; E, Afric. ina, unina, inani, Melag: nini; Ason.—ina, inde, inda, &c. &c. . , . b. tayi, ayi (Father ¢.,) Male aya, Ur, ayo, Kol. iyo, eang, engan; Lhop, Asam ai. -e, amma, (Father e.) Tib. Him, d. appa, avva (Father b,), Gond aval, Manip, aphu, avu, apwi. é. talli (Father a.) ; 39. Mountain. a. varei, malei, mola, male panw; Kol buru, Ur. parta, Kir bhar ; Samoied bore, borr; Fin ware, wuori; Turk, uba; Cauc, mehr; Sansk. parva, 6. konda, gudda, gudde, konom, hunnu. . Murmi kung, gang, Newar guo, Gur. kwon &c.; Yenesei konony, kar, kai; Turk. kirr, Pashtu gar; Fin. gora, kuruk ; Sausk. giri &c,; Georg koj, kirde; Pers. ku; Afr.—Galla, gara; Ason—Indon, | gunong, &e. 40. Mouth. a vayi, vaya, bayi, payi; Ur. bai. ‘ Caue. bak, bagga ; E. Africa, atta, aof, af, ma, Malagagi vave, vava ; Asones :—talai, baba, vivi, ba &c (Malag.) %. noru; (Magar nger; Manip mur, mor, mai.) | Sam. nal, an; Yukuh, Tungus anya; Ason.—ngari, nganga, anka, © Austrl, ngan. : 4l. Moschito: a. kosurw, kudu, Ason,—Sumatra agas; Pol, kutu, 5. domo. Ason,—Bis, tamo, &, a a ‘ - ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 1a (? Tib.) Ason—Kaili sani, Mur. I. sonney. | 42. Name. , per, pera, peru, pesaru, hesaru, pudar ; Gond batti paral, a. al, ira, iralu, iral, rav, reyi. (G. U. hor, mula, ngayul, walo &c., Malagasi elina, aline; FE. Atrican bara. Ason.-mariri, bouli &e. 44. Oil. , ‘ wee ennei, enna, nune, enue, enu; Gond ning; Limb ninge (See | 45. Plantain. a. yazhei, vazha, bale. x, Korea phatshyo, Semitic mez, mis; Asor.—Indon pisang (Mahra mis.) : 46. River. . a. varupunal, aru,[Arm. Fgypt.] eru, yeru; Kol gara? Manip. D urai, TNeeae Bin Be &o; Mong mura, muran; Turk. muren; Fin wire; Caue, kor, hor, or, Georg. oreuba; Arm, aru; Pers. arga; Semitic bahr; £& Afric-—mura (Makua), mulo, bolan, Egyptian aru; sones,—umala, brang, wara &e, b. puzha, pa, pole, hole (G Ue , age Sp ; 4/r.—Galla aba, Shangalla epucho. c, tude. Manip D. tu (“water”); Sam. te &c. &c.; Afr.—Yoruba ado, [A ‘*videly diffused root. See T, U,] 47. Road. a. neri, dari, dova, (see T. U.) . é. vazhi, pade, hadi, sadi, batte, (Iran, path, paddavi, wat, bat, &c.) ec. Ur. horah. Kol hora, horen, hor, da-hari, Gond. sarri; ? 6. ? Mahra horom &c.; Georg, shara, gsa ; Pers, roh, sarak, (Kurd re,) qee T. U. : uppu, uppa, upp, (Men bu.) 49, Skin. . a, ada), tol, tola, tole, tovalu, torra, Gond tol; Dhimel dhale. — Ugr towl, tuolye; Turk. tari, tire, deri &c.; df/r—Galla itille; [? Ason.—Pol. kili, gili, Indon. kuli, kulit &c, (Sansk, kriti.)] - 50. Sky. a. vin, vanam, Manam, Miimnu, ban, ponu (Ultr. Lau, van, fon, fa, Lungkhe wan, wyn &e. Ugr. menen, pil; ‘Turk, awa, piel’s; Ason—wono, Austrl., awan Sabimba, Sambawa, wang Madura, banua Nias, also “country”, “land’’ in this and other Vocabularies, awan ‘‘cloud” Malayu &e, b, mugilu (? megha “ cloud,” Sansk.) 51. Snake. a, kadsevi. e pambu, pamba, pama, pavu, havu, pamb, pab;*para-punu (See ) ¥ ae 14 : ETHNOLOGY OF THE INDO PACIFIC ISLANDS, 52. Star. | a, vin-min, van-min, minganna, minu,pone-min, Mal, bindeke ; Ur; inka, Kol ipil, epil ; Burm. min-ong (U,); Ason,—bin-tang, b. chukka, chukki. ; Yenes. choyen, Ug. chus &c., Canc, zuta, za &e, e, daraya (Uind,) ; 53. Stone. a. kan, kal, kalla, kallo, kal. . Kameh, koall, kuol &e ; Yukahiri kell; Fin’ kalle; Cane kera; Arm, khar; Pashtu kani, Sindh. kod; 4son ~kain N. Austral.; kala Pol, (kaa reng Indon, “Cural”, &e,, but this may be from karang,a made thing &c,) b, rayi, ° 54, Suz. a, “pakalon (See ‘Day).” Yukabiri bugonshe, Caue haak, bok, buk, “Day” bu, bigula. Afr.— Shangalla woka, Galla &e. wak, wakwak ‘‘(God);” Ason.—Australian bagarin, baga. 6. poddu, hotts, polutu; Ason—? Pol, polotu, ‘heaven ”; Kagayan bilak. e. pallilli, @ 2 55. Tiger. pol, pati, huli, pili, pirri; Gond. pulli. oria Pon. : b, nari. , Burm, nica; Semitic nahar &e. 56. Tooth. a, eyiru A, Tam, Ason,.—Austral, yira, 6. pal, paila, palla, ballu, pall; Goud. pajk, Male pall, Ur. pall. . ae pane, pankt, ponk, pu, pia &c.; Canc. kbili, Aibili; Ason.—(eoe e. kuli. Cauc. kerchi; Ason.—kuni. Kis, 57. Tree. a, sedi, chedi, cheftu, Cauc che, peha, kctiad &c. (see T. U.) b. gida, _ e, maram, mara, men; Gond. mara, Male, Ur, men, Afr.—Makua mere, Kwil, mute ; Hind per, ‘ 58. Velluge. pekkam (? Mal. Jay. pakan “markew’’.) ur, uru, Gond nar, tara. desam (Arian), palli, halli. moda, mort, S28 an oe ETHNOLOGY OF THE INDO-PACITIC ISLANDS. 15 59. Water. m, punal, vellam . ? Hind. pani &¢ , Indones, bana, banyu &o, ; b. tanni. Oss, dun, don (see river c.; T U ec, nillu, niru, nir, (Sanskrit nir. 60. Yame, valli (Ulty, bal, wiru, berha &€.) ee % a ~ - i, a Td : f stn 7 s » : ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, APPENBIX TO CHAP. VI. OF PART II. A. COMPARATIVE VOCABULARY OF THE NUMERALS OF THE MON-ANAM FORMATION, One. Tndia,—moi, midh, miad, mia, mi, men (Ko), Gond). Ultraindia,— wei Kas; muo, maai, Mon; bo Karen; po Angami Naga; aima Singpho ; moe Kamboja, Ka, Chong; mot, Anam. Malay Peninsula, —mnui Besisi. Asonesia?—amui (2) Mairasi ; abut. Bruner I. (2), Mui, moi, moe, is probably the oldest of these forms A/frica,—Kicamba uniue Siege mo-ja, 5. Af mu-sa, mo-chi &c.); Akuongo, ema; Cam, mo; Nuba wa-rum, wee-ra &c,; Rungo mo-ri, Benin bo, WN. and M. Asian,—om, uem, Samoied.; emu, omin &c, Tungus. . Two. Ind.,—bar-ia, Kol, bar-ea, Gond. Ultra —ar. Kasia; ba Mon; bar Ka, Chong; pia Kamb. (3 of Kol); hei, Anam, Mal. Pen., —be, Simang ; mar, ja-mar, ma, Binua. N #.sia,—mal-gok, Chukchi. E. and §, African,—biri &c.; W. Afr —fire &c, Akuanga, epa; Cam, ba; Karab. ebah; Rungo mba-ni; Calb. ma; Mok, tba; Bong, baba ; Bin, be; Ib, aboar, abo. Three. Ind.,— op-ia, p-ia Kol. Goud. Ult,—pui, pai, Mon; peh, Ka, Chong; bai Kamb., ba Anam, Mal. Pen.,—wiup, Sim.; am-pi, am-pe, am- pet, Binna I can find no decided foreign affinities. The term (pui, Mon, wwi-p Simang) appears to be a flexion of mui 1. The Binua form appears to have been modified to accord with the Malay am-pat. 4, the Binua terms above 3 being Malay, In the extreme N E otf Asia and the ad- jacent Polar American Janguages of the same formation, pi is an ele- ment in 2, pi-gayut, Chukchi, Eskimo. It does not occur in N, Ameri-« can formations, There isa distinct term for 3, having a very limited range,—lai Kassia; jui, Car Nicobar; luha Nankowry. Unless the Namsang van-ram, Mu- jung and Tablung lem, are connected with this word, it has no other di- reot affinities that are very obvious. The following terms may be rela- ted to it ;—Kassia han-dai, kon-dai 2.( 26,3; kuis G in Kuki, Karen &c, but here it may be merely the Kassia prefix ka,-d and r being fre- ap tly preceded by min Kassia); rai, 8, Bongju, roe Kuki [See Erenr]: liere is another, and, as it appears to me, more probable explanation of this form, Ifthe basis of the Mon-Anam system was strictly binary, and pui, wai, is simply mui, 1, a little disguised, the Nicobar lui may be Ay original form of the Kasia lai, and, like pui, a mere variation of mui. Four. Ind.—pon-ia, ti-pun-ia, u-pn-ia Kol ; o-pun-ia, u-phun’Gond. Ul¢r., —pon, Mon; pon, Chong; puan, Ka; buan, boan, Kamb.; bun Anam ; fuan, feun, fen, Nicobar, (? /emang, Simang). This numeral is very remarkable. Itis a modification, found in Aso- nosia, of the Atrico-Malagasi term which, in another form, has spread . ETHNOLOGY OF THE INDO-PACIFIC ISLANDS: so widely over Asonesin. Mid.-dfrica (Hausa, Galla, Saumali, &c.) fudu, fulu, ofur; Malagasielar, efad &¢. Asonesia,—an-fa, Nias; an- far, Kel); hai-phar, Tanne; fan, fang, Caroline ; far-fat Marian ; owang Pelew; havi, New Guinea The more common Asonesian foym is the dental pat, am-pat &c, The rogtis Egyptian and Iranian ( fu, four, chat-var, fuso &c.) Itis simply a variation of the similar root for 2 (i. e. the dual of 2, asin other binary systems), The Ultraindian and Indian forms cannot be deriyed from the Iranian ehat-var. They are evidently connected with the ancient Asonesian form prevalent in Mi- cronesin, and derived trom Malagasi. Taken with the fact that the terms tor T, 2, (and 3, ifa mere flexion of 1) are also African, they afford some proot (hat the same long enduring western civilization which car- rie! Maldgasi and I.-African words to Asonesia, at one time embraced Ultraindia ia its influence. : Five, Sang Kasia; pa-sun, Mon; thanin, tan’, Nicobar; chang, Ka. The Tau jim ajpears to be a modified contraction of san, tha, The Kuki sin ka, Bongju tswur-kar, Car. Nicobar sum, 10, is probably the same term. It is Alrican, beimag found in the same formation to which the | Malagasi owes so wuch, and from which the previous Vindyan terms mity also have been derived,—Galla, Saumali, shan, zan. “That it is an ancient Mid-African root, belonging to a diffusiveyivilization, is evinced by the progress it hos made to the Westward and Southward, Binin, tang, !apah a/-tong, Cam. ma-tan (this langueze has also the Vindyan and Ulirainiian 1,2), Calbra son-ni; Rungo otani (Comp. the Nico- bar forms); 8. Atrican sanu, tanu, &. The same root is also Samoiede, Tungusiau aod Aleutian (sam, tong, chang, san, sun) an Asiatic distri- bution which shews that its diffusion in Asia and Africa was anterior not only todhat oi the Lranian, Semitic and Caneasian, but to that of the prevalent Seythic, numeralsystems. Radically the word signities “hand”. It is fond with this sense in Dravirian as in many other lan- unges. ~ In the Menam basin a second term is preserved, pram, Chong, Kamb; nam, lam, Anam. The root appears to be ram, lam, nam, and to be a prefix, asin Mon, The Nankowry lam, 10, is the same term, The Vindyan term is mor-ia, mona-ya, mone, mo-ya, Kol; mun-ia, muna-e Gond, which may be an inversion of the Kambojan and Anam, or vice versa If tle Mon term had been Tibeto-Burman, there would have been grounds tor identifying the Viadyan and Kambojan vocables and considering themi as representing the original Men. But as the Mon term is native or African, this explanation appears to be in- . niimissable. If the Vindyan term has displaced an older one of Mon origin, it may have been derived from the Tibeto-Burman pungu, phungu, Naga; bonga, Garo; phong Mikir &c. ~The Kambojan and Anam term is not only found in the Nanco Jam, 5, but in Daphla rang, 10, Mon, klom, 100 (Ka dam, Anam ight and in shorter forms, ra &c., in the Naga dialects, Garo, Mikir, Bongju, Kuki, Kambojan, and Lau, with the power of 10, 100, 20 &c. Allthese forms appear to be referable to the binary nomenclature, which some of the terms jor “ eight’ prove to have co-existed with the uinary. Ram, 4, is probably an abbreviation of ra-ma, that is ra 4 for 2 dual) and ma, 1. Inthe same way the Vindyan 5, muna &c., may be a flexion of pun, 4, Prom the evidence atfordedby some ot the higher ; ; PA 4 HTHNOLOGY OF THE INDO-PACIFIC ISLANDS, 8 numbers that the Mon system used both the binary and qninary methods of expressing t nuinbers, a usage by no means singular, [ have little hesitation in referring both the Vindyan and Kambojan terms to the single Mon-Anam system, Although Lean tind nething to warr@pt the opinion that the Vindyan and Kam jgjan isngusges might have obtained separate terms from East Africa, for L have no doubt that all their African terms were received through the Mon, it should be remarked that analogous*words are cur rent in some African vocabularies. The true explanation, .I conceive, “is, that the Alriean terms in question are formed from the same binary definitive roots, mia, ba &e; ra, la, na &e. A Suabili dialect has manut, and to the westward forms similar to the Indian aud Ultraindian occur, —nun, Bullom; mu, Kru; num, Akin ; aum Amina, : 2x. The Vindyan terms, like some of the Ultraindian, appear to be Tibetan. Ind.—tur-ia, turu-ya, tarwi Kol.; tarm, tarume Gond. The Gond has an exceptional term sa-rong (sa is a prefix in 5 also, s-aij-an; yaly Toda) which appears to be simply ru of the Kol dialects nasalised But it may be direetly deriyed from the Naga form so-ru, In the Gawil form the ng becomes m. Ulir,—ka-rao Mon; the Bongju, Kasia, Burma, Sing- hu, Chong and, Ka terms are all similar antique modifications of the Tibetan, The Kafrao is a derivativeot the Mon. The Chong ka-don is a nasalised form similar to the Grond sa-rong, In Bodo, Dhimal, ongju and Naga, forms in ¢, d and y also occur. The anomalous terms are tha-ful, ta-tul, ta-fad Nicobar; shauk Kyeng; sau’ (abrapt accent for +) Anam. The Nicobar term may be composed. of tha 5 (trom thanin) and ful, fad, which should represent 1. A similar term for 1 does not exist in the Indian, Ultraindian or Asonesian pro= yince, save in the Egypta-Atrican wotu, uotu, motu &e. but in the latter it is used for 10 (i.e. one tale). It is probable therefore that ful is a Mon~ Anam binary term formed flexionally from pun, fun, 4. Tha may either be from the previous term on the repetitive principle, or it may be the Mon-Anam prefix. Shauk, sau’ has a deceptive appearance of affinity with a wide spread Alrican, Iranian, Causasian and N. Asian term, the final of which is generally #4. African, shita, sita, seda &c.; Semitic shat, hat; Ugr, chut, hat; Iranian shash, sechs, six. But it is merely one of the numerous variations which the Tibetan root undergoes. The origi- nal may have been the sibilant thauk or thuk, The Hakhoing khrauk preserves tie broad vowel. . Seven. . ; s Mon, ka-bok; Ka, pah; Anam, bei. This term isa flexion of 2 (the word for 5 being omitted, as itisin most of the other formations), I have alreaily mentioned that most of the Ultraindian and Himalayan languages adhiore to the Mon-Anam quinary principle in forming the teri ior 7, and that a large number of them indicate the commencement of the higher series of numbers, or those above 5, by the prefi. (general- ly ta, ka). Lepeha preserves ka in all the terms from 6 to 10; and Kiranti, which, in its word for2 (ha-sat), retains an ancient root which reappears in other languages in terms tor 4 and 8 (i-sat Namsaig Naga), has another archaic term in hag 9 7, which is evidently the Mon bok, The Nicobar sat might appear to be Hin li, but as the Nancowry dialect has ja-kiat, which resembles the Lau form of the Chinese term (chiat), 4 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. ant is probably Chinese also. The Chinese root is very widely spread (Ugrian, Iranian, Alrican &e.) fd. i-ya, €-ia, i-air Kol; a-ya, a-ieah, Gond. (Some Kol dialects haye taken Hinditerms). This is the Dravirian e (e-zha, e-l, ye-du &c,) ight. Ind. ival, irl-ia Kol; ilhar, elar-ic, Gond. This term appears to be an archaic binafy on”, a flexion or ,eduplication of the Dravirian 2, ir, and to be relate! to that for 9 aud 10, as in the Dravirian system. — Tn some of the Ultraindian and Himalayan languages the term for Sis asimilar flexion of the ancient Mon-Anam rootinrfor2or4, It is found in the Yuma group ani the Nicobars,—rai-fart Bongeu; rae ° Kuki; prab Kasia ; awera, Cor-Nicob?; Kiranti, re-ya; Murmi, Gus ee pre (comp 4, re, pli, &c.) . The other prevalent terms appear to have been adopted from the Chinese. The,Mon ka-cham, Ka and Anam tam, appears to be the Chi- no-Tihetan sum, tum, tham 3 (5, 3) on the seme prinviple that 7 is 2 (5, 2) in many of the Ultraindian anil Himalayan langusges. The Bur- mese shit, si, Chong ka-ti, Kyeng shat, Singpho ma-tsat, Naga cheth, chet, thuth, chut, sat, sep, te, tha, Garo chet, probably involve a misapplication of the Chinese term for 7, ch’hit, ch’her, sit, thet, tshih, The Abor-Miri pu-nit-to, Miri pinye, Daplila play-1ag are 4,2. Binary terms for 8 appear to have formed the limit or highest number of the scale at one time, for they have been applied to 10 and even 100 [See Ten. ] : Nine. Mon, ka-chit; Ka, chin; Anam chin; Karen chi. This is the Chi- nese 1, i. e. 1 short of 10, as in Dravirian and Mikir. The Chong ka- saris peculiar. It is perhaps from the Chino-Tib, san3, — - ; dnd.—ar-ea, ar-e, ar-he, ar-aiak. Although ar is apparently a flexion ofthe ir of 8, which is 2, itis probable that it represents 1, as in the Dravi-ian terms. In the Male or-?, 1, the Drav. on takes a vibratory form, end in Tuluva the common term for 9, om-bodo (i, e. 1, 10), takes or as a pref. (oram-bo). Ten. Ind.—gelea, gel Kol, gulea, gil, Gond. The Angami and Mozome- Angami kerr, kurr resembles gel. Kerris evidently a derivative from the Naga thelu, taru &c, The only analozous toreizn form appears to he the Chukchi kulle, and both are connected with African terms for 1 (kulle, Sokko &c.) Hissi, 20, is evidently the Hindi bis, the conmuta- tion of the labials and the aspirate being easy and common. ' In several of the Ultraindian languages the African root for 2in r, which enters both into the Dravirian and Mon-Anam systems, re-ap- pears in higher numbers, az in African Janrnages, a Consequence of the ultimate binary basis, Rae Bongzu rae, Kuki is 8, in Chong it is 10, in Lau and Kambojan it is 100 (roa, roe K., noi, hoe L.) In other Jan- guages also it is used for 10, Tt appears in the Anam mare, Naza faru, tarah, ‘helu, dérr, kurr, and Kumi he-re 10. With these compare the Burmese tar', taya, Karen teraya, Mikir phar 100. [The Nancowry lam 10, Kadam, Anam tam, Mon klom, 100, appear to be formed from 5, nam, lam, ram, or from 8, tam, Anam, Ki.] Asa connection be- tween Sand 10 or 100 exists in the case of rai, and is also remark« ed in the Tibetan aud Chinese systems, the latter is probably the true ETHNOLOGY OF THE INDO PACIFIC ISLANDS, 5 derivation, In the Namsang Naga, which uses the Chinese chi for 10, the ancient term is retained in 20 and the higher terms. 20 rugkngi (1), 2). 30 rogkram (10,3), The ak is probably a connective like Aa in the Khari tarvhonet, 12, (10,2), tarahasam, 13, (10,3). 'm Khari and Angami ra occars, following the lower aumber, Khari, 30, samruh, (3, 16), 40, lirah, (4, 10); Angami,’30, seri, (3, 10, in this dialect sam be- comes she), 40 lhida, 60 rhipengu (10, 5). Mozome-Angami, 30 surr, 40 lhide, 50 ripangu. Ra also enters into the terms for 100 in some of the Naga dialects,—rakru Nagaung, rudrah Khari, contracted tu kra, kre in Angamiand M, Angami. Some of the Yuma dialects also preserve it. bongju, 20 rabu-kar (10, 2, the term ter 2 being Mon-Anam wiso.) The Abor-Miri has no trace of this term, but in Dophla it maintains its place a RA 10 rang, 11 rang-/a-akin (10 and 1) &c., 20 rang- chang. In Bodo, Dhimal and the Nipal languages I do wt remark any trace of it. In Garo it occurs in the Mikir form for 20, rung, The Ari- arised Gangetic languages possibly retain it in the nwnbers between 10 aud 20, e.g. 11 ega-ra, 12 ba-ra (in which ba may be Mon-Anam*) 13 te-ra &. With these comp. the Sanskrit eka-daslan, dwa-dashan, trayo-dashan. n several of the Kumi dialects the Mon-and Kasia sun, san, 5, re~ appears as 10,—Kuki, sun-ta, sum-ka; Car Nicobar, sum; Bongju tswur-ker ; Kyau, tchuow ; so Kumi, 100, tcbun wai-re. . In most of the Naga dialects 20 appears to have pertially retained a Mon-Anam character. The terms are ma-chi, ma-tsu, tha, tsa, cba, niakhi, me-ku, ma-ku, in which, cha, tha &c. ere the Tibetan “ten.” In Namsang cha ts also used for 100, cha-the. In the terms{or20,ma must stand for “two,” that is, itis the Mon-Anam ua, ba &e, 2. This is confirmed by the Mon ba-chi, Chong bar-se, Ka bar-chit, all signifying “two-ten.” The Kambojan ma-pai is a similar term, but ihe use of pai for 10 is anomalous it it be the Kamb. bai, 3. It may be connected with rai, hai, Mon-Anam flexieus of 2, but it is more likely to be a form of “one,” Kasia wei, It is very remarkuble that the same term appears to be preserved in the Murmi 10, chi- wai (one-ten), and in the thence derived Sunwar 100, swai-ka (hun- dred-one.) In Kumi it occurs in tchun wai-re, 100, in which tchitn and re both represent 10; from 40 to 90, wiis used for 10, and it is probably a contraction of wai, In some of the Murmi numbers | o-kol is used for 2) or “score.” Itisevidently connected with the Nova and Mon-Anam ma-kn, ba-chi&ec. It may be inlerred from ‘his that the Jepeha and Lhopna kha, khe, “score” are fragments of similar terms, the postfix for “ten” having come to represent ‘-twenty,’’ like the corres- ponding cha &c, in some of the Naga dialects. ‘The anomalous sau, 20, of the northern Lau dialects (Lau, Ahom &c.) is probably a variation of the sme Naga form, ‘The purer Siamese retains the Chinese term yo sip (2, 10). Io four of the Naga dialects the term for 10 is ban, pan, which is pros bably from banga, pangu &c. (Kumi pang, Mikir phong)t The Kasia —+s * Bopp derives ba from the Arian dwa,and ra, re from the Arian dash deka (Comp. Gram $319), The contraction of dasha into da is not improbable, the conversion of da into rn would be easy. + It may be connected with the Chinese wan or dan, “ten thousand”, origi- nally the highest simple term of their system, aud which ike Mancius have appro- ' priated to 1,000 twuun, ° 6 ETHNOLOGY OF THE INDO-PACIFIC. ISLANDS. - Bhi-pon appends |he same word to the Chines: chi. The Limbo thi.how appears te be direcily derived from the Kasiaterm, In both systema bom &r, represents 10 in the higher numbers, 20 Kasia ar phon (z, 1"), Limb ni bong; 30 K lai pon, L, sum bong. Kasia continues to nae pon in the terms above 3), but Limba discards it and wloptsgip which is the Kiranii kip, 10, the Mikie variety of the Chinewe chip, Kumi bas.aleo apoog in 2U and the bi, her numbers, B. COMPARATIVE VOCABULARY OF MISCELLANEOUS WORDS OF THE MON-ANAM FORMATION. In the following list of Mon-Anam words I include all thnt are foun in the UltrainJian languages that remain prepositional. Buta considerable namber are Chinese more than Uliraindian, and many do not belung toany glossarial formation that has predominated in, Ultra- india, Several have also been derived by the: Mon-Anam tro the Tibeto-Ultrainiian vocabularies. On the other band, mny words, widely diffuse! in the N. Oltraindian and Gangetic languages, that do hot wppear in this list, probably belonged originally to thy prepositional formation, and have been lost in the progress of those changes to which ali glossaries are subject. : }. Air. a. An hoi (“wind” jo); Kol hoio, hove, hoyoh. , Ar. hawa; Mong abur; Tib. abur (‘wind’), Aborasar, “wind” ; Be isali swar; Burm. hong si w. . 4f.—Danak. haha, hahaito &c. wi, Dalla soueta w., Malagasi isoute, isonte w.; Malg.-Asones, angine, auze, angin, hangin&c., w. Asov,—Kyan ba-hoie, Pol. suu(emitic, Malas. ), b. Kamb. akas ; Manipuri masu, masi, ma-hia, marti, numg-sit 5 Sunw. pas Milch. bash, (Root probably as, ash, hash, sa, su, st, shiy thi, se &c.) / Jap, kase; Fin gaiso, aiseb; Persian, Turk, yosi, awasy ; Ug. wes- ges; Mong. Tib. asur, (Abor asar); Samoii. masi, bursi, barsbt &c. w. The Ultraindo-Himalayan root is evidently as, sa, &c. and iden=' tical with a., but in some of the above Mid-Asian terms si &c. may be merely the def. postfixed to other wide spread roots. It oc~ curs in numerous other Asiatic and African terms, com'ined with formatives or with other roots. It appears to be ra:icully identical with the preceeding term, (a.). In the Kambojan form, the vowel ol the pret. as been euplionically transposed, ak-ns for ka-as or ka-sa. ds0n.— Wagi usa’, Ende, Soloras, Pol. sau, (Bajo srua, Pagal rusa.) . | c. Mon kya, kia (also “wind”); Nag, tikbe ; Goud kueyo w. [? Kol Royo}. Comp, Dray. * Mong. kei w (See Drav.). Ason.—(See Drav.) | d, Mon, bloeis (An. “Shy”. Sansk vata, Beng. batas, Eorop. ventos &e., Hind, bad, Bod, bar, Gar. l; Ug. wire, wot, wat, Sam bar-ehi &c,; Korea paran,» pharong ameh, epulud &¢. (Manipori phanra &,.), Af—sevh.miaro, Aso —- (“wind”) Binoa barua, Meri beroai, Cel-bes puire, pori, New Gain woe rei, Aust, porowu, maeilo, boran, Wirienguma, padro; Sabimb, badi, Keg, padak (Drav.¢, a) a Lan lom, Gare lam par (See FT, U, lung &c.; Dray, 2.) RTHNOLOGY OF THE INDO-PACIFIC IsLANDS. 9 2. Ant. a, An. kin, Kas. ka dakin, Singp. gagin, (? Singhal kanhi); dvon —Taraw, kino, Ach, kimo, Sumba kama.ola, Kis: ogaia, Timor kuva, 6 Laat, not, pouk; Kamb, sar-moit; Mon sa.mot, kia mwot; Dhim aha moi, Kol wai, moe; New. imo; Male poks Ur, pov (laa) (? Siauh makvro); Ason.—Mal, Indon samat (Mon J, Bina pos, isos. 3. Arrow. a. An ten, Ka tong, Manip. than, the; Nag. lawhan, /asanc. /usate, Jahan; Siam lus-so., Mik. thul, Koyan thar, Kaw, tai; Gimou toup, Lepch. chong. Crasant tem, tom &o ; Tongos, sir-dan, 8am. changa, Mong. so1nn, chos mo, sumun; Chin, ten chi, Sam. tise; Beng, Hind. tir (Manip. tef)s [lhe roots in kare probably identical with thosein ¢, ch, 8. Chin -ehi an, dian, &e.; Lau kong; Sindh, kan (arrow), Asam kanr, (are); Dav kanei, (arre); Mon Knya, kon, tanga), Thesame root is used for s’bow" ia many languages,— Lau tanu, thanu, Ka tongah, Kamb. ting, Mou tangas tangah krang, Singp. ndan, Pabi tanu, Singh. donni, Beng Hind dheank.} Ason.— Bis odong, odiong, Bin. Mal, &¢ datnak, dama, Bin lamak,, (but this form may be connected with the Semitie rama), Jav, kandews (bow). The Malagasi pawa, luna &c, has kept its ground io most of the N.bae Polynesian vocebularies, b. Mon lay, levu, Kamb. pircen (See T. U.) » Ason.—Pagai rorou. ; e. Kas. ka knam (prob, 4—kam from kan, with the initial nasalised), 4. Bird. a, An, chim, Mon sin ngat, kachim, Kas ka sim; Gond (Gawil) sim 3 Silong sisom; Nic. thom (? Kul. chene, Kor, chongwa, Sunw, chiva, Bodo ilon-then) [ree T. U.] 7 Ch, chio, clisu ; Mong, sibechu, shobon, shobo &c, Sam. teshundas ehiacha, Korisk atschel, Aino tschipkar, Arm. trshon, Sansk, porchi, Af—Saumali shim-bir, Galla sim-bira &e, Ason.—Biaua chin; (? Mad acham), Kamb. sat; (prob. T. U. sa &c , with a Kamb. eonsonantal fins! ) c. Lav nok, nuk, oaut, Burm. ogbak, oget, Kapwi oghet, Mormi nasu, Urao orak. In Abor Miri itis preserved in rok pi *‘vird’s-egg”, bo h word being Vindyan, “ f. Wauc, angko, anko, anako, woenucho, aka-onoch, anakwi; Sansk, beyanggo, bihanga¢ | Malaya &c anggas, angkus|, Af,—iuko, Galia (‘ fowl)". | Ason,—Niha-Pol. (common), manok, manu, manuko ‘auc. woennko} &., N. Austral, aluk, lukaluk, @o. (? Galla), The great prevalense of this term in the Indo-Pavific vocabularies and the prefix ma, render it probable thait existed in the Naga Manipuri and allied dialects before it was displaced by the Tibetan vo, sa &o. [lt is dot probable that a root for “biru’ is peculiar on the Aaoe European Continent to the ©, Caucasian gronp on the ofe side, and to Lau aud a*few other adjacent langasges on the other. Is it evidenily connected with the almost universal root for “dick” ‘Tangus, oiki, Mong. negusvo, nogd so, Turk. urak (Urao, orak “bird "’) ordek, ortale- &u.. Pashta ordek, Ug. batta, wise, wasig, &c. (t, s for d, r, 0 )¢ E. Afr, bvitak ; Ind —batak, boguk &o,;. Arab batog Span. pato Indones, iteb, ite; itik, ii Gv. (Gund, ite, ** bird”) ; Bogl duck &e, du j & ETHNOLOGY OF THE INDO-PACIFIC ISLANDS: 5. Blood a. An man;? Kol myun, [? Ugr. male, woor-ak, wyar, urr, ver &o. K ciak in levo-mui, Korea ko mor; Afr Shang, mohbu, } Ason —Simang, Binva, maham, mohom {? Zend, wohone], Kia, meany (Ko!) Torres St. mam; Austral komara ( Korea.) b, An. tiet (Chinese) ce K. chiam, Mon chim, Kar. thwi, Yuma ti, thwe, si, thi, hi, i, sai, chei, Koveng tezai, Jili tasai, Garo chi, Dhim, hiki, Deor, Ch. chui, Bo!. thoi, Tiberk. shui, New, hi, Limb. mekhi, Lepch vi, Che. wi, Churgl vi, Sunw, usi; Mag, hyu, Kir, hau,? Male kesu, (lin, chiue, hint, hue, he’, Japan tschi, zi, Aino. kim; Tungus. anyi, shoma, Turk chan, kan, yon &c, Uireass, cha, Mi-je}, zi jioh ; Arn. anvun, Lat. sanguis. Afy.—Malegasi razana, zanada &o, Ason —“imang cheong, Bin. za, zuis, . . d Lau leut, lut, let. Kaslm., Sin ‘hi, rat, Tamil ratam &c., Sansk. rudira, Dray, udiram neturn. Ason —Sunda let, N, Austral elod. oP iy ¢ Kas, suum (?a Kol; but probably sam (c.) with the initial nasa~ lise.) : 6. Beat. a An. ding, chi liuang, Kas. liing (See T. U. and Drav.) b, K. tuk, tup, tutuit, Chong ‘dok, Kaduak (T. U. and Dray. tha, oa, Jonga &c, with a Kamb. consonantal final), Ason —haruk Kayan [Naga surung, Abor hulung].; bidok Bajo, Pa« sir. e. M. Keng, galon, Naga lung, ihe &c. (See T. U.). Leng iscone nevted with the slender Burman torm the. d. L, reua, ru, heu, hu (See ‘I. U., Mishmi, rua &c.) 7. Bone. eo, An. shung, chiang, K. cha’ang, Mon. tsu, Deor Ch, pichon, Kas. shing. Ko) jung, Ur. Male fochal, Kirant. sai-ba, Limb, say -ef, Turk. 8 oma, syomjok, suuk, suik &e.; Jap. hone; Lran. os, asthi, &e. Asn —iti. jis, Bin, jahang (Kol), Tobi chil ( Male, Ur, ), Pol. sui, hui, siv!. : b. Lau duk, nuk, keduk, (Manip. arakhai, Tib,rako, Lhop. rutok Be.) ' . . Cuuc. rekka, rata, rotli &c. 8: Buffaloe. 4 o An. klong-nvk, Mon. priong, pyen, Mikir chelong, jelong (’ hoibu mnamuk, “cow” )p Chan H brung; Kas. s inreh, Nega le, teli; Manip lu, iroi, woi-:hoi; saloi, alui, aghol, malul, suci, ra.oi (See Dias.) ; Kyo cha-lawe Ason—bin, Mangk. Vug. tidong, Mandh, Dor. tercng, Mok. nang, : 6 Kamb. krabo, harbu, kwr-hai, Ka kar-pu, Chong, L., khrai, khwai, Bum. kyue, Kol kura, kera (Drav ‘he final bu, bo is pro- bably the wide spread word for “ox” &e,, which, in the Manipuri and Yuna dia ects is uiso used in compounds e g. (wei-rhoi * bufiaice”, woi-tom “fent’’), Ag/n.—\ndon, krabui, kripue, karabao, horobau, karbau, kabu, ku- bu, kebao, kibo &e. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, ~ 9 9. Cat. @ An. m'‘au, L. miau, meau, Kas. miaun, Garo myou [See T, U.] 4. K_ chinia [Manip. See T. U.] ce. M. pa-khwai (See Dray.) Korea, koi; A:.—Kwil paka. 10. Cow. a. An, bo, bou; Lau woa (T. U. ba, nwa &c.; the amplified vowel is found in the Songpu woi prefixed to other words for “‘cow”’, “buffidoe”’ and “elephant” [See Buffaloe], 6, An. sung-krau, Mon kleau, Kar, klo, ‘ Dray. akalu, Cauc. aka, Hind, goru; Tungus, dokor, ukur &c, Fin. sagar, iskar &¢. Arab. bagar. _ c . ku, L. ngua, ngo, ngoa, hu, Mon. nua, Burm. ngi, Blut, ngo, Jili ‘anga. ‘The last term may be connectetl with some names for the “buffaloe” in adjacent languages,—Naga chang, tyang, Dhim, dia. The Mon nua is evidently the same as the Burman and Karen nwa, Which may be the T, U. wa. The Kambojan ku appears to be nearer the Chinese (gu) or Indian (go, guo &c.) forms of the wide spread guttural root. Ch, ngni, Turk, ona, ina, inak, .Magv. uno. d, M.kwon hban, Kar. wa bing, Rakh. min, Naga-man (See T. U.)L e. Kas. masoi, Naga masei, Garo. ynashu, Bodo muashojo, Tami, asi. ; « Ugr. mus &c, Latin bos (See Dray. 3). Il. Crow. @ An. kon-kwa; L ka, kawa; Dhim. Gond, Kir, Murmi kawa; Limb. ahwa; Mund. Bhumij kova Ka iga com. Drav., Tib,, Asiatic and Asonesian root. Siodhi kaw, Kash.kav, Ason —Bat, Bug. gawo’, Tojo gawap, Parig, kau, Vel. kaon, (Indon. kuwau, kuau, * bird’, ‘pheasant’ &c.) 12. Day. : 13. Dog. _ @& Hair. ‘ a, A.tau’, Kar. thu (see c ) Chin, thau mo &e, (thau is “head’’), ‘ &. A long, Kyen lu. war Aino ruh, Arm law. rego pp oli, Dt. uiri, Tar. ira, Pol. Iau; ulu (Ludon. &¢, ulu, “‘head'’); Wirad. uran ; Kayang inang, e K, sok, M. sok, thwot; Binua sok. Probably from the Ultrain- do-Himalayan song, som &c. and ‘lib, sha, which again are connected with the Mongol usun, Turkish asim, Ason.—Tobichim, — L phom, phram, Vib pu (a wide spread root) e) “Kas shnin, 11 sh is part of the root, it is probably a modification of the I’, U, etliwon, thung, sam &c. If the sibilant is a prefix, the root may be connected with the Naga min, Garo aman, Bodo khomon, 26 Hand. a. .* In the terms for 100 Chinese and Tibetan difler. The former has pe’, be’, p»’, equivalent to pak. The latter hasgya, The ‘libetan term appears to be uuknown in Ultraindia, The Chinese is foun! in two Naga dialeets,-puga. The ancient Tamil paka has an aceidental conci- dence with the Chinese term, But the root pa may be ultimately « Gap Lhe two paragraphs marked * should have followed “Ten” p, 27, 30 ETHNOLOGY OF THA INDO-PACIFIC ISLANDS: referable to a similar source with the Chinese. Both the Chinese and Tibetan terms have some appearance of being flexions of the terms for 8. Chinese 8 pat, 100 pak ; ‘Tibet 8 brgyud, gye, 100 brgya, qya. IE ’ the scale is based on u binary one, asis probable, the ivoeiiiance is real. In some ot the Mon-Anan languages the same root is found ex~- pressing 2, 8, and 100, Addendum (p. 18.) 47 Road. a. A. dang, L tang, M dan, ga-lan, K, kalanti (T, U.) b. K. chira da. o, A,ngaba (? Bodo, Dhima] lama, dama Tib.) ba ETHNOLOGY OF THD INDO-PACITIC. IsLANDS, 3) D. COMPARATIVE YOOABULARY OF MISCELLANEOUS WORDS COMMON To TIBATAY, INDIAN AND ULTRAINDIAN LANGUAGES, l Air. a. T. w. lungma, s. Ihakpa, Serp. langbo, Milch. lan, Kir. hak, Murm. ihaéa, Gar. nangmro; Mishm. arengga; Nag. rang-bin, rayg- che, Manip, nung-sit, thirang, phanra, khiraug, nong-lit,,Garo lam-par, Lau lom (1). Barman alliance, li, le, kali, fali, Changlo ridi &e. (Gyarung). The Tib. lhak, Kir. hok, probably occurs iu the Lepeha sag-mot ‘fait”’, sak-ni “day”, and in the Limbu tam-sak-pe “sky”, sa-chalk ‘sun’, Drav. elary Fin lil, ilma, lemin, Yukahirvi ili, Turk. il, eil, Aino rera ; Cauc, churi, Georg. kari; Iran. aura, aer &c.; Semitic (“*wind’’) re, ire, ira; *Alrican ahru, &¢; dson,—Meri longlangi, Sambawa langi. Pol (‘wind’) malangi, Rotuma leang, Sumba riru, Mandh. iri, “wind”, irt Pol., savili, Parig., pu-ire, Kaili powiri (lhe same root is found in “Sky”, “Sun” &c). 2 Ant. a, T.w. grogma, Serp. rhunma, Sunw. rog-machi; Aka farak, Ab. taruk, . b. T.«, thoma, Lhop. kyoma (2). Ason.— Binua tami, Silong kedam, Hok. hutom. 3 Arrow. T. w. mah, s. da, Serp. Lh da; New. Bodo bala, Sunw. dla, Ka- ren pla, Singph. pala, Jili mala, Nag. thelu, Manip. la, lu, nla, malu, mala; Burm. mra, mya, Murmi, Gur. Mag. mya, Kir. me (3). Iran.—pilu, Sansk. ; [vil, veliu, billu, “bow” Drav.,] pilum Lat, pijl Dutch, &e. E. Afr. mpamba Suah., impamba Makua (Dray, ambu &e.); Malagas. avana, Yoruba offa, Sech. bura “bow”, Ma- kua mura, “bow” Asonesian, — Pol. taa, Pagai rorow, Lobo lara-kai. The most prevalent Malayu—Polynesian terms ere of direct Malagasy and African derivation, Indon. Pol., fana, pana, &c, 4 Bird. a. T.w. byn, Db bhya; Tiberk, Milch, pea, pia, Lep. pho, Limb. bu, Chep, moa; Male puj; Nag. vo, 0, thevu, Manip. va; Singh. wu, Kyo wa, Kumi ka-wa (4), Asones.—Samb. pio, Kis ban. ae ia b, I. s chya, serp. jua, ? New jhango, Sunw. chi-va, Mish. tsa, Dhims _———— SS eae. — ee a) fala tali dane tali, Karen, Khy. kai, Burm. &o,'; Manyale erdah; (Guar mro)]. ; me) (Tak seehok . Manyak darah, Gyarung korok (Tib. grog,) Thochu tu-khra, Sok pa arokb we, fora skhro.] . : (3) [Takpa mla (Burm.), Manyak ma (Burm. &c.), Horpa ida (Tib.) : ‘ (4) [Gyarang pye pye, Takpa pye (Tiberk, Milch.), Gyami sphuie chher (Chin. chio &c.) ; ; 32 ETHNOLOGY OF THE INDO PACIFIC ISLANDS, jiha ; Manip. masa, macha, matsa, Nag uso, uzu, ozah, auha &e, (é b). Chin. chiau, chio, tio; Korea sai, Aino ziaf, chirpu &ec. Tungus. gasha, Turk chush, kash, kugha, kus, &c. Magyar katsa (iow!). Caucasian.—Lesg heso, weza, uetzu, Cire. zis, chshi, Georg kinchi Asonesia,- ima janga, Mank, jangang jangang (Newar) ;N Aust, bijinj; Komr. sisu, Lump. sia (‘fowl’); Sain, tuhu, chundo, chiacha; Mong. shobo, shobon, sebechu Hind. Beng. chiriya (see also Mon-Anam), 5 Blood. a. T w.khrag, s., Serp thak, Lh. thyak, Milch. pulach (5), Bengali rakta, Sindhi rat, Sansk. rudira, Nic. kanak; Ugr wuorak; Afr.—Saumali, Galla dik, diga [ihe Asones Pagai logow, Buol luku, ag: lugu, rogo &c. Roti daak, are probably from dara, lara] 6- Boat. a T.w.gra, Lh dru, du, Chang dru, Serp thu(6). Abor efku, Mish, rua, Garo rung; Gang-Ultra. dunga, Nag lung surung, arong, ru; ae rung, dunga ; Khyeng, Lungkha laung, Kumi plaung, milauug, Kyo plaung, Mon klenz, ga-lon, Lau ru, rewa, An, ding, Kas, liing, Burm. the, Singpho li, kar kili. * Ason—Mair. era, Tilanj alina, Tag. longa, Jav. palang, Indon, bare, bulu, parau, prau, Pol. falau. | hos b, T.s. koa; Nag. koa, khuon, kho, khung. Asones.—Savu aoe. 4 | e. T.s, syen, Nag. yesang ; thseng. Chin, chiu, ch’biang lang. : 7 Bone. a, T.w. ruspa, Mag. misya ros, Sunw. rushe, Chep. rhus; T. s. ruko, Serp. raba, Gur. nug-ri, Murm. nakhu, Lh rutok, Mish. ruboh; Nag: tah: rha, rah, eru, uru; Lepch arhat; Manip saru, karau, maru, para, soru, aru, arukau, uru, thuru, khAru; Yuma ru, aru, ar, Singph nrang; Burm aro, ayo; Lau duk, nuk (7) Drav. elume, eluva, elu &e. Pashtu alukei, &c: Semitic, alam, alat; Cauc ratla &e Malag taolana, tolan. Asones.—Juav. balang, Lamp. belu; Austr pura, Baw, loh, Komr lolor, Solor. riuk (? Lau), Erub lid, Taraw. rij ludon. tulang &c (Malegasy), 8 Buffaloe. T. w..mahi, s, mahe. Luh. Lepch. Murm. mahi, Gur. mai, Serp, meshi; N. & C. Tangkhul shi. - Hind. bhaina, Beug. mobish. Semitic gam-bus, jamus. The original term was probably the wide spread mos, bos &c. “ cow” [See Naga,— Himalayan “cow '’; Naga masi &c.) (4 6.) [M anyak ha (Naga auha)]. (5) [Tukpa khra} (6) |Gyarnne bru, Takpa gru, Manyak gu (7) [The-chu ripat, Gyar. syeerhu, Lakpa rospa, Many. rukhu ; Horpa rera) ETHNOLOGY OF THE INDO-PACIFIG ISLANDS. 33 9 Cat. | 53 a. T. w byila, Lh. pilli. “ Hind. billi&e., Lat. felis. . &. T. s. simi; Nag. ami, miang, mish, mochi, mesa; Ny & 0. Tangkul fami, dame; Yuma mi, @mi, mim, nim-boi, mi-ynung; Kar. mu-uiyn; ? Kamb.chima; Bod. monji (9). - Mongo! mii, Japan mio &c, Chinese miau, bin, niau, nzio; Asones.— Thden. miong, miau, miu, meo &e. bd oor) Isthe lib simi connected with singa, “lion.?” In Indonesia sing, Auching, wching is aterm for “eat”, | 10. Cow. | a "T w ha, phachok (Tungus chyukwr], Sunw bi, Limb yopi, Dhim pia; Bir awa, nu, Kar, wa bing, ga phi (10); Lau woa, An. bon; Dray awuy avu, pei &e. Indonw sapi. - b Bh. Milch. lang. } Fin lebma, lehima: Caue ol, als Semit.-Ta (root): Galla, Ambaria, Jam, Danak. fah, Gallu lawom, lawuim, Suahili lombe; Ason.—Lndones. limba &c- . 1] Crow. T w khata, Sunw,-khad; Magar kag, Ts. ablak (11), Serp,. halik, Lh. ola, Lepeh elok, Gur mioneya. j {Drav, Vindy, Gang-Ulor kha, ka, kiawa, kag &e J, 12 Day. — T. w. nyinmo, s nyimo, Serps nie, Lh nyim, New nhi, Meg. nem-sin, ~unw na-thi, (hep nyi, nei. ( hang. ngewe, Dhimn- ngi tims Naga auyi, ui, dint, whi; Singyh -sini, Jili tdna, barn wey Kars ni, Yu- nia fend, haweni; Awaw nywt (12). Tangus ininyy manyvi; Yenes na; Caue kini, dini, Asonea.—Born, ngo, nga, uugu, Sambaw ano, Buol nu &e; (foun, Indun, neno, inil), a. T w khyi, Serp Lh. khi, (13) Mileh. kwi, Tiber: kool, Limt. khiay Murm. navgi, New. Khi-cho, Gur nagzyo, Mage Chep kui, Chang. klu, Aka, Abor cki, Dhim khia, Gar. kai; Naga, kui, hi, tia; Mani; wikte, Singpl kwi, durm. khwe, Kur, hewi, Yau. wi, ui, dui; Kamb. chake Chinese Khian, ken. keo &e., Korea kai; Mong. xekoi, Fin koira karre && Caucas koy, choi, woi, gwai, kari &¢ Asoxes —Bin. koih, koyo, Phil kua, avai : A & Pos, uyo.(@ Scythic, from a form similar to the Dinna koyo) (10) {Vhoohu gwa’, Manynk wo-mi (mi is used generically as in dinge. mi ‘*buffuloe”, See also *Cat’*)] a ; (11) [Phoehu nyugewo, Gyarung fabrok, Tukna skpek { (42) [Gyaraung nye, pish-nve, (Burm.), Horpa ye-le, Takpa nyenti' (Dhint Nag). Many, nasheha ] e (18) [Gyarung, Takpa khi}. ~. e wos (9) {her fochi, Sok pa simiy Manyvak machen, Pukjpa syinthy | 34° ETHNOLOGY OF THE INDO-PACIFIC ISLANDS.” ; 14 Ear. a, TT. w. rna, rnawa, ne, Lh. navy Kir, naba, Murm. nape, New, nhai pong, Gur. nabe, Mag. na kyep, Sunw. nepha, Chep. no, Chang, na; Naga na, SPAGO RS Iece tenhaun, anve; Maui). nu, Aaa, khe- na &c. [but this may be from kan); Singph. Burm. Kar, Yam. nu, Kar, nho, ka-na. (14) Cauc. en, in, banka, b. T. s. am-clo, Serpa am-chuk, 15 Earth, T. s. Serp. Murm. Gur, Chep. Chang. sa, Lh. sah, New, cha, Mag, jha; Bod. Gar. Nag., Karen ha; Naga ha-waw (10.) Samoid ya, Jap. tsi, ai; Turk, yazhan &c.; Cauc, misa, musa, mit- za, sach &c.; Zend sa. 16 Egg. T. w. sgonga,s. Serp. gongna, Lh. gongdo, New. khyen. Cauc.—Lesg. gunuk, kor-kon, gaga wc, ( 17 Elephant. a. b.T. w. glang-chen, s, Serp. langdo, Lh. lang-chen, (17) Chang. Jang-pehi; Champh. lamun; Burm. w. chong. s shen, Kar. kchong, khsa, Yutu. sang-bung, tshi, kusai, dusiiai; Mou shen, tsin, Lau tsang, chang, tyang, chiang. . Chinese chhiang, sio, siong; Suahili simba Asones.—Jav. leman, liman, (Champhung), Binua, Mal, deram, Bin, braugte, bringkil, 18 Eye. | a. T. w. mig, s, mik. (18) Him. mik, amik, michi, mide, mak, mi- kha; Abor amig, Kol met, med, Dhim«) mi, Bod. magon, Gar. ma- Akar; Naga mit, mik, ¢enik, fenvik, tenok, awhi; Manip. mit, mbik, mik, amak, @micha, omit, amit; Singp. mi, Barn w. myak-chi, Dar}? Kas, Aamot, Mon mot, pamor, An. mat [See Mon-An. and ray. Chinese mok, ma’, ba chiu &c.; Jap mamige, mey. Africa,— Makua meto, mezo, Suab mato, Kihiau mess, (these EB Afr, terms are plural), Kongo mesa; Malag massu &c. sun —mat, mata &, 19 Father. «+ T, w. pha, s. pala (19) Serp, aba, Lh, Murm. Chang. apa, Lepeh, Gur, abo, Limb. amba, Kir, ba, New. abu, Mag. lai, Sunw. -bave, Mileh, baha, Chep. pa, Aka aba, Abor babu, Dhim. Gar. aba, Bod, dbipha ; Vindy, aba, baba &c; Nag. apa, apu, opa, apo, va, taba; (14) [Gyarong tirne, Manyak opi, (Murmi) Takpa ne-blap, (dlep is leaf”) Horpa nyo | (15) [ho-vhu zip, Gyar. se’, Tanka sa’) (17) [Gyar., Tapka lang-chheny Sokpa Ihabo-che, Horpa lamo-chhen. | (18) [Gyar taimyek, tammyek (Burm), Takps me-long, Many. moi, Horyg mo (Chin )] } (19) [Gyar, tape, Many,, Takpa, Horpa ape}, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. » 35 Manip. apa, pa, ava, iba, papa, avu, Singph. wa, Burm. phae, Kar. pa, Yum fa-pha, phai, ba, bo, abha ; Mon bah, bha, Coutmon in all parts of the World, Mongol, Samoide, Turk., Semitie, Atrie, aba; ‘Turk, Tomali baba &c. &c.; Asones.-bafa, papa, bub, ibpa, pua, pupa &c. &e., [ibu (Manip) '* Mother’’] 20 Fire. T. Him. me (20), Him. mi; Aka ummeh, Abort eme“Dhim. one; Naga mi; Manip mai, chami; Burm, mi, Kar. me, Yuma mi, me, mai; ? Mon miot; (See Mon-An ) Chin, we; Aino abe, apeh, ambe, Jap, fi; Fin bi &e; Afr —Tumali ibi, ibe, Kuam. mo, Malagas ale, apo, alu ; Ason.—Niha-Pol. ale api &c. 21 Fish. T w New. nya (21) Morm far_nya, Lh. ngva, T.s., Serp Limb. Kir. nga, Sunw ngau, Gur. fangna, Lep ngo; Aka nyay, Abor engo, Bod. Gar na; Naga nga, nia, nya, angu, angi, kho; Manip kha, chakha, khui, khi, nga, sanga, thauga; Auam kha, Mon ka, Kas, dokha, Nicobar ka. Fin kal, kof &c, Samoid ual, Korea koki. Asones,-ka, ika, ikan . (Men-Anan) | | 22 Flower. T w metog, s. mentok, Serp, mendok, Lh, mentog, Murm. men- du, Tiberk, ments, (22) 5 | Galla doko. . 23 Foot. T, w rkangpa, s. Serp. kango, Lh. kanglep; Mikir keng; Manip. khong, ki, akho, ake; Singp dagong, Kar, khong, kha, Yunis akauk, akok, ya-kong } Mon chang, dzong, Kamb, chong, An. kang-shun (23) Drav. kal, Chinese kha, Asones,—Australiau kana (Drav ); >imaug chang, Tobi chem (Mon-Kam!); Indon, kaki &c, (Yume); Fin, Chukchi, Eskim, Cauc. See Dray. (int 24 Goat. Tf. Lh, Wurm, Gur. Mag. ra, (24) Changl. raba ; Kol meram, Urao, era; Naga roan, ron, Garo purun, Bodo barma. Seniitic aron, Saumali arre, Danak. illa, Galla ri. 25 Hazr. a. T. w. skra, s, Murm kra, (25) Lh kya; Singph ‘kara, Fin karw, Aino karnu, Koriak kirtshiwi, kirwyt Cauc —Lesg, chara; Afr-—Mukua karare, Saumali dukore, Galla. &e. chegur. | rot (Pho-chu, ‘Takpa me’, Many. same’, Gyar. timi, Horpa oma’ (Aka) ‘ . (21) [Gyar. chu-nygo, Takpa nya, nga Horpa hya], (22) [Many., Takpa mento, Horpa meto}. i (23), | Vhoerhn jako (Yama) Horpa ko, Sokpa khoil (Dray,)). (24) [Pakpa ra) . (25) [Pakpa kre]. : 36 EPANOLOGY OF THE INDO-PACIFIC ISLANDS. Ason,—Cer wkar, Australia tulkure - & T,s Serp ia, Limb, thagi; Kar tho (See Mon-An.) Fin ata, at &e. Asones, 7 Panel tawn e Vow syu, Dhim. mui tu; (26¢)? Kol ub, up.- | F Fin ‘up, ip; Chin bo Axon.—bok, but, ba &c (probably Dravir,) 26 Hand. / T. ww. lag pas (21) s. Serp. lango, Lh lappa, Dep kaliok, New pa laha, Gur, lopfa, Aka luk, Abor, elug . Noga dak, chak, yuk; Singh Jleta Burm, vw, lnk, s let. i Turkish ilik, Ost Jagol, Ason —Indon. langan (generally ‘arm”’), Sunda lingun, Pol, riunga, linga: — 27 Head. T, w. mgo, s. Serp, go, Lh, gufoh: Naga kho, tako, Manip. kok, kni, kau, akav; Buriv. ghaung, lu-gu, Kar kho; Nicobar koi. (27) — Ason. —siiang, Bin. koi, kue, kai, (Manip. Nicob.), Balignini ko, Batan ogho. Jap. kavbe, Kam kabbel; Aust. kabera ; Cauuc,—Cire. kah, aka Ke 5 Lranian kapala, caput, &e | b Gur kra, Mish, mkura, Bod “khoro, Manip, ¢akolok (Tib, kra “heire’’); Yum Ju, blu; ? Kas fli, : Cauc —korte &e Iranian kula, cranium &e, Asones —Sim kala, Aru guli, Born ¢akolah, Mal, **seull’’ ¢fankora,. | 28 Hog. T. w phog, s. phakpa, Serp phak, (28) Uh. phagpo, Chep pink, New pha, Sunw po Mag wil (? Aka akon.) Abor eek, Garo vak; Naga vak. ak, auk, thevo, thavo; Munip bok, Aabuk, avak, wok, hok, ok, Singp wa, Burm w. wak, s. wét, Yum wok, wet, wut. Malayal porki, lraw pig, hog, porces &e.; Cauc huke, khuka, Ka bkake &c. (Aka kuk-pu) Asones-tiatan bugu, Sirawi kapot, (Mahip kabuk), Pol, puaka, 29 Horn. T. w. ra, (20) s rajo, Lh, rou, Murm. rhu, Gur, ru, Sunw, guro ; Lepeh, arong, Chep rong; Chang, warong ; Abor arerig, Mish riu, Dhim>) dang, Garo korong, Bodo gong; Kol daring, dring, Ur marag, Male marg; Naga rong, wong; Singp. rung; Mon kreang, greang, Kas. ka reug, ; 30 Horse. T. w. rta, s, Serp. Lh. Mur. ta (30), Yumatsa, sha, Kar kthai. kthe ; Tark. ut, at, Yenis kut, kus &c. [Hence kuda, ghora, &e ) Arm, tsi, Caue shu, tschu, &¢ Sam, djuka, tschunde; Sunsk ashwa ke , (25¢) [Pakpa pu, Horpa spo, Many. mui (Dhin.)] (26) Haar tayak (Naga), Many. lap-che’, Takpa la, Horpa tha], (27) [Gyar tako, (Nag) ‘Takps gok-ri (Manip )Horpa gho]. | (28) [Voh-chu pi, Manyak wah, Takpa pha, (Newar) Horpa val]. 7 iso) Cho-chu ‘rak, Gyar, tard, Many, rude, Takpa reba, Borpa- rumbo) (80) [Takpa te’). ETHNOLOGY OF THE (NDO-PACIFIC ISLANDS, 37 3L House. a TT. w. Uh. khyim, Kir. khim, Murm. dhim, Sunw. khi, Gur tin, Limb him, Serp kangba; Abor ekum; Naga hum, ham ; Manip, yim, yin, shim, shin, tsun, chim; New. chhen, furm im, eing, Kar, i, gueng, Yuma ing, eing, um; Magar yum; Mon he, Kas? root ini, Kuki teng. (31) (Samoiede ma, ne, men &c. 2?) Ason.—Tobi yim, Mille im, Sunda ima, Sav.emu; Indon, ruma, huma. 6. T s. nang; Mrung nao, Bod, nou, noo, no, Anam dang, na, ya, nya, ngua. Cauc, unneh, Ason.—Lamp nou (Bodo), 32 Iron. T, w. Ichags,s,Serp, Lh chhya; Naga hache, hatse ; Mon pasoe, pathway, (32) Korea soi, say, Samoid yese, bese, basa, &e.; Canc. achik, icha, ask &e.—; Iranian ws, eisen, ayas; Ason—Indon base (Mon pa- zoe), bosi, basi, besi, bisi &c. 33 Leaf. a, T w loma, Lh dama, Lep lop (83) Murm. New lapte; Gur lau, Mag tha, New: haa; Aboranne, Mish nah, Dhim Thava, Bod lai; Manip na, thi-na, thing-na, sing-nu (“tree-leaf”’), Singp Jililap, Burm rwak, yuet, Kar In, Yuma la-kang, An la; (Drav. elei, ela &c ): Fin lopa, lopat, lopta &e (Tib, Murmi &c); Malagas. ravi; Ason.—Erub lum, Bima rupa, Savo rau, Pol lau, ran, Indon, daur, raun, Sumatra, Phillip. betong, Mal. &e. lai, asegregative used in enumerating flat objects, as cloths, sheets of paper &c. b. Tib. s. hyoma, iyowe, Serp. hyomap, Lh. syoma, New. hau. Chin, hio, 4 34 Light. a. T. w. hod, Limb. ot; Naga oitike, (34) Turk, syod. : b. T. #, hwe, ev, Serp. Lh. ew, Singp. thoi, Jili thwe, 35 Man. T. Him mi, Abor ami, (85) Mish, name, Garo miva; Naga mi, ami, theme, thema ; Manip. mi, thami, mu, samu, mai, chamai; Yuma ke- mir Fin mios, mis, mes, pioa; Turk, bai; Zend memio; Afr.—Galla mi, - ma, Manding. mu, Ason.—Indon, mama. 36 Monkey. a. T. w. sprebu, (36 a) Lh. pya; Aka lebe, Abor sibie; Naga veh; (31) [Tho-chu kih (Sunw.) Gyar. chbem, Takpa khem). (32 Iho-chu sormo, Gyar, shom, Horna chu, Many. shi]. (33 Horpa bala’, Takpa blap (New., Dhin)}. . (34) [Tho-chu uif, Horpa spho (Turk.), Many. wu’, Takpa wot, Mru watai}. ‘ 35) |Gyar tami (Nag Manip.), Takp, mi’}. (33 a) [Gyar, shepri, Takp. pra]. * 33 ETHNOLOGY OF THE INDO-PACIFIC IsSLAND?. Singp. we, Jili fawe, Kar. ta-aoe, Ason.—Inilon. ? brok, belo, ubal, b T. s. tyu. (36 5) 37 Moon, T. w. zlava, s, dawa, (37) Serp. oula, Lh. dau, Lepch, lave, Limd lava, Kir. la dim, Murm I[hani, Changlo lani, New. mila, Gur Jaa ngi, Sunw. la to si, Chep. lame, Aka pala, Ab. polo; Mish, nalua; Naga da, Teta, luta, yita, letu, le; Manip, lha, fangla; Singp. Jili sata, Ruma. ar. la, Yum lo, slu, hla-pa,ta; An. klang, Law lun. Fin mano, Korea oru, Samoid ivi, ireda &c., Chukeh. iraluk; Iran, luna, moon, mon! &e. Afr.—Danak. bera, Felup fylin, Malyges. vula, vulan. Ason;—Austr. palu ; Indou.. Pol. vola, bula, bulum &e. (through Malagas.) . 38 Mother. T. Him. ama, amo, am, ma,-mang, amai, (98) Mish. mama; Dhim. Gur. ama, Bol. bima; Burm ami, Kar, mo, Mrung amo; Kas, kami, Kamb mi, An. me. ° _ Nearly universal, e. g. Yukahiri, Yenisei, Samoid , Pin, ama; Africa, Malagasi, ama; Asonesia, ami, ma &e. 39 Wountain. a, T Serp. ri, Aka nodi, Dwphla miodi, Abor adi; Dhim ra; Yue ma lai, mu-ra; Lau loi, noi, dui; Lhop ronc; Manip malong, kaiong, Khiuug; Kar koe long, Yum &lang, Alang, slang; ? May. Sunw, dunda. Chio tia, Tangas alin, uro, uray-Mong ala, wa, Fin ur, Ason— Indon lulob, ieleh, lada, lede, alanga, olono, gunong (Manip fa-long). & Tos (W Tib) dak, Lepely rok, Mileh dokaug, ‘liberk dang- kang, Chamang donk, Limba tok-song; Male toke, Goud dongar ; Jili satong, Burm tong, taung, — Turk tak, tag, dag, tau &c , Japan dake, Aino tapkub Ason.—In- don letek,? Erub tulik [Viti toka-tau, but here toka probably meaus *“fixed’’]. thuang solo Rotuma. : 40 Mouth. T., Serp Lh kha, Milch, kagang, Chamang kehk; Aka gam, Mish, taku; Bodo kugha, kbouga; Angami Naga ata; Kar, kho, Lung-ke aka, Kum. uk-kha Chin, khau, hau; Yenisei ko, gou, khan; Sam angan, ak, agma, ake; lurk. ukse, agus, &e; Japon kusi, Kameh kasha; Caue, haku; ‘Semitic kho (Gara, Malra); Mongol kurgo, gurga. Ason — Austr, ka, kuraka (Mongol); Jay; ebikam (Aka) : 4\ Mosquito a T. w sunbu, mcburings; Abor sunggu;? Kol. bhu-sundi, bhu sendi; Manip sangsan, facheans, kachang; Kumi chang-rang. Ason, Kaili sani, Murray I souney; Bima samulan (Kumi ) ; b. T s. sye-dougma, Lh, zen-dong, Him lam (with other roots join- a '/006€0€@©«O}O0 ' (86-4) [Gyar ti}. = (37) [Gvyar, ésilée, chile’ (Nag le), Many. The’, Takp, le’). (88) [Horp. Many. Takp. ama, Gyar, fomo). * ETHNOLOGY. OF THE INDO PACIFIC ISLANDS,, 3Y. ad); Aka farang; Naga mangdong; Kumi chang-rang. Ason —Sas. tutang, Bat, tirangkas. o 42 Name. Tt. Him ming, min; Abor amin, Vish, atanng; Dhim ming, Bod, Gar, mung; Naga min. man, feaung; Manip ming, armin, oniin, mi, amang ; Singp ming, Jili faming, Burm. amin, ami, Kar, mi, meng, Yum aming, amun, : Common, , - AB. Night. : T. w. mtshanmo, s. chenmo, Serp. chemo, Limb. kusen, sendik ; Dhim, eae Naga asang-di(Limbu); Manip rasa, rosa; Singh. sana, Vil SOME. P wee! Su, y Turk. achsham ; Mong. so, chei, suni; Yenis sai; Aino asi, asiru ; Jap. joru, Fin ose; Semitic asar, azar, (Mar. Ghar.); Afr.—Malagast asine; dson.—Balig. sanguna, Tid. singi bungi, Mang. chan. 44 Oil. a, TT, w. hbrumar, marku, Bhut. makhu, Japan abra; ? Kilimani,-makura, & IV. -s. Serp. num, Lep. nao, Limb. ninge, mingay ; Kol senum ¢ Naga manga; Singh. Jili, Lau nam-man. Ason.—Bis nana, Pol. faugo, pani; Indon. miniak, minako, miua (Limbu, Nega), 45 Plantain. : T. s. Lh. ngala, Kir. ngak-si; Naga ngo, mongo, mango; Manip, ngo-shi, nga-chang; Singp. dango, Jili hhungo ‘ Ason.—The Indon, pisang may be an Ultraindian term, pi-sang (Comp. nga-chang &e.) - 46 River. ; T. w. gtsangpo, s. chang Serp. hyung, Lhop chhu kyong, Lepch, ong yong, Limb. wohongy Mar syong, Kir hong-ku, Gur. khwong; Mishmi tsalo; Naga joan, shoa, swokia; Manip, shinggu; Anam song, sung, som. . Pashtu sean, sin, sint (heace Sindu, Hindu, Indus, Scinde &o.); Ver. shor, shur; Mongol chun, usun; Turk -wsun, su, sug, yei-su, dsulga &c.; Ugr. iyaga, yugan &c.; Sam. yacha &c. ; Ug. yo, yozi &e.; ‘Korea ha syu. Ason,—Indon. sungei, sungai, sunge &c,; Cel, salo. 47 Road T. Hin. lam. New lon, Sunw. la; Aka lam-tau, Ab lam-be Mish. ailam; Dhim dama, Bod. lama, Gar. lam; Naga lam, unglan,, lemang ; >; ingp lam. Jili tanglong, Burm lam. lan, Yum lam, lang; Kas. lanti, on dan, ga-lan, An. dang, Lau tang. Chin, lu, lan; Gara orom, Mahra horom; A/fr.—Malagasi Jalambe, Sushili jira, Sech sela, Ason.—Indon rorong, lorong, balan, laravke. Janga: turang, taluna, dalang, dala, jolo, jalan, Pul. sale, hala, harunni, ara, 4@ ETHNOLOGY OF THE INDO PACIFIC 1SLANDs. . 48 Salt. T. w. tsha,s. Serp. Lh. chha, Mur, Gur. chacha, Mag. cha, New. chhi; Dhim. dese, Bod. shyung kare, Gar. syang; Naga matsu, metsa, matse, machi, sum, hum; Manip. ntau, machi, miti, ti. nchi, matai, kasam, thom; Singp. tsum, Jilichum, Burm chha, tsha, Kar itha, Yum ma-tai, sheve. Japan shiro; Ugr. sow, sol, sula. sek, so, &¢ ; Samoide si, sir, sak &c ; Cauc, shug; Indo-Gur sal, salz, sout, salt &c.; Afr,—Galla usu, Danak assebo, Malagas sira (Jay Sam.) Ason—Indon sia, asin, si- yok, sien, asin, asiad, sira (Mulagasi), mase (Manip, machi ), masikh, masiti, penasim, mengahi Pol ulane, masima, masi. 49 Skin. T. w pagspa s pagpa; [? Bolo bigur, Garo bigil]; Naga ¢akap, ane ); Jili maphik ; Singp. phi, Kar. phi, Yum moe-pik ; ? Kamb. si- kk * Chinese,-phi, phue- Ason.—Austr, bokai, bakai. 50 Sky. TF w nam kha, s. Serp Lh nam, Kir namecho; Naga aning, anung. Samoid,—nom, num, nob, nyoa; Ugr in, ninak, inniyn, numma, no- men, nair; Kashmir nab; Asor.—Vimor neno, Kissa onga [See Sun, Day.) : 51 Snake. Tw sbrul, s dew, Serp dral, Lep. beu, Mag bul, Sunw. bu-sa, Gur bhu-guri; Aka fabuk, Abor tobi Mish tebu, Bod jibou, Gar dupu ; Kol bing; Naga pu, phalu, purr, thofa, ahu; Manip. maran, pbaru, se phrui, mari, plharun, phrol, lil, nrui; ‘ingp @u Jili tapa; urm mrwe, myue, Yum. ral, rai, poi, pwa, marui; An ran. Malagasi bibi; Kwamam, oria; Bengali uraga, Hind Pash mer; Asonm —Tilang. bio (Lepch. Yuma) N Austral, ambit; Indon. ular, ula, orei, alhin &e, 52 Star. -'T, w. skarma, (52) s. Serp. karma, Lh. kam, Mur. kar-chin, Gur. tar- a ..k takar, Abor. tekar; Manip. ¢ikrou; Singp eagan, Jil sakan, Burm kre, kve. Uvrian,—chur (Ost ); Koriak gor; Yenes kaken; Korea kurome ; Mong odon; Iranian, tara, dara, staranm, astrum, stella, star &c.; Afr —Suahili tara ; Ason.—Kayan kraning, Viti, kalo, Indon entara, ndara, dala, etab, tawar, war &e (war is probably a different root, be- ing found in New Guinea, Torres St., Australia &c,, tara is probably of reeent Arian origin.) 53 Stone. T w rdo, « Serp. Lh. do; (? Gond tongi; ) Anam da. : Korea tu; Ost to; Other Ugr. lang, ko, ku, kiwi &c ? Malagasr vatw, —_—_—_—_—_——_———EEEEE es Sooo (52) [Horp. sgre, Many, kro’, Lakp, karma}, EYHNOLOGY oF THE INDO-PACIFIC ISLANDS. 4i 54 Sun. T. mnyima (54); ? Kol singi (See ‘‘Day”). 65. Tiger. Se s T. w. stag, s. tak,(55) Serp jik, Lh. tah; Yuma tchak-ke, tuke oe, tagain. Tran. tigris &c. . | 56 Tooth, T. Serp. Lh. so, (56) Tiberk soa; Murm, swa, New. wa, Gur sak, i ae Changlo shia; Nag. pa, Manip ava, ha, hu; Burm swa, thwa, japan cha, ha, fa; Ugr. pu, hut &c.; Semitic sin; Turk. tis, tish &c., Tungus. it; Afr —Malag: mifi, nife. dson,—Indon, yus, titi, ngisi, isi, lisik, niso, niho, nifo &c, (Malag ) : 57 Tree. T. w. jjonsshing, s shing-dong, tam, (57) Serp, dongo, Mur, dhoing. Lh. ade Tiberk Milch botung, Limb. Mag. sing, Kir, san -tang, Gor sin-du, New sima; Aka sangaa, Ab. sine, Mish masang; Dhim shing; Naga, sang-tung, sau-tung, sun-dong, si; Manip. thing-bang, sing- bang, thing-kung, thung-rong, asing, hing-tong (Tib.), hing-bang; Kar theng, thi, Yum ting, teing; Kas. ke diing, Lau ton, tun. Chinese shi, shu, ch’hiu, chang &c; Yenes hochon; Sam. cha; Ug. guy &c; Canc che,she; Kemch. uthun, utun, unda; dson —Lamp. Land batang; Phil. dutung, Sunda tang-kel, (Mal &c. tang-kai ‘‘stem”’), 58 Village. a. T yul tsho, a. thong, (58 a) Kir. teng, Chang.dung; Abor do« lung, Mish mating; Gar. song; Naga ting, ting-kbua, ching. - 6 T. w yultsho, Serp, yul; (594) Naga yum, ayim, yam; Maring yul, you, 59 Water. | . T. Serp. Lh. chhu, (59) Limb. chua, Kir. chawa, Gur. kyu, Sunw. yvankhu; Naga tsu, dzu, zu, atsu, tu; Manip, aichu, tu, tundu (Gangs ltr ji, si, ti, di, ri, tai &e. Mon dai}. Chinese, chui, shui, sai &e.; Jap mizu &c:; Samoid, tui, ita, Ugr. uit, ute, wesi; Turk. shiu, su ke: Mong usu, wsun; Ason,—bisany mazi, meze, mazi&c. son.—Indon, chie, Yam: T.. w, dona, s thomia, (60) Serp. dhoa, Murm, teme, Gur, taya; Bodo Mrung. tha; ? Kol da sang. Chinese dua tu, tua chu; Ason.—Indon uda, New Cal. uti. 55) [Horp. stak, Takpa tee]. . 56) [Thoch" swe’, Sokpa syu-chi, Horpa syo, Gyar, tiswo, Takp, wa’, Many, adic racer) ip RS shi’, Many sapo’, Tapk. sheng-dong].. 3 tn, ak Many. nyimay Horpa nga}. (58a) [Gyam twang-cha. | (585) [Takpa yu, Many. hu, ? Gyar. wo-khyn, éu-khyu); (69) [Thochu chab, Gyar, ¢ichi, Takpa chki; Many. dya’, Gyan pa (ay (Uyat, séten), 42 ETHNOLOGY OF THE INBO-PACIFIC ISLANDs, APPENBIX TO CHAP. VI. OF PART II. =. VOCABLES NON-BHOTIAN IN ROOT OR FORM COMMON TO THE NOnTH ULTRALNDIAN, HIMALAYAN, AND MIDDLE GANGETIC LANGUAGEs, * ] Air. a, Changlo ridi, Aka dori, Yuma aii, kali, &c., Burm. Je, Kar, bli, khii. 1a). : ¢ oe ilma, iim, [ila, élaran, elem &e. “Life’?}. Wog. Wi (Ost. Hl, wula ta, Mag. elet, Life’ |, Turk muil, chil, (Comp “Wind”, Yokabiri id, Aino rera, Turk il, eil, chil, sil &c.3 Mabrah era, Gara ire, Arab. re &c ).—Ason.—Sumba rivu; “Wind, Mandh. iri, Ut. lauri; Celeb. puire, pori, &c.; Aust. mailo wiri-nguma &e.; Pol savili. [>ee D 1) 6, Limba samit, shami, Lepeh. saginat Mag. names, Sunw. phase, Milch, hash, Ab asar; Manip, nung-sil, Khoib. nong-lit, Maram nhilut, Luh. masi, N.'T. mas, C, 1. mashia, Maring marthi; Nag, rang-che, (L}). a * Mong. achur, ahur, uhr, Tib. hur (wind) [See B, Mon-An. The Bima sinet appears to be connected with the Linbu skamé]. 2 Ant. a, Serp. rhunma, Abor-M. mirang, Mish. aruang, Jili_ teang-lang, Luh, chaling, N. T. lang-za,; Khoib miling, Mar, phayang ; Yuma pa- leng, maling, pa-lein-tsa | Drav, Ason—>ee Drav. Some of the Asone- sinn words are immediate derivatives from Ultraindian. ‘Thus the Marine miling-is found in the Bagis and Dore‘biri, Kand, bere and with a final sin Madur, Baw. dédie, but this may involve a separate root tor Simang has les and Panilas-ga. The Abor mirang and Mishmi ruang, render it probable that the Tibetan reg, Sunw. ragemachi and Burman Spa-rwak-chhit or pa-yuot-sik present the same: root] b. Gur. chiji, Sunw. rag-maghi (rag, Til.) ; Murmi syon-ri, Bod. ha- sa brai, Ahom mynehu, Deoria Ch. chimechi; Nag, mecha, muthang, tik-sa, tik-ha, hache, hung-zah, tsip theh &o; Manip DD ling-za, chameha, kak-eheng, nteang, \iteng, tangin, mateang-\wi, ching-khia ; Yume wadsi, pa-lein-tsa, Burm par-wak-ehid; (Dray. china, pijin ) ec. Lepeh. tak-phyul, Limb sik-chem-ba, Kir, sa-chata-ya, Nag. tik- sn, tsipeehak; Manip. hak-cheng. . 6 Boat Gurung pla-va; Kumi plaung [See Tib-Ult. and Dray. (a)). * In App. Ei the roots, not the de anivan. are italicised, I have substituted ‘non-Bhotian” for “non Tibetan’, as Mr. Hodgson’s East and North Tibetan vocabularies now show that geyeral of the vovables are Tibetan, (la) [Gyar., Towng the tadij | (1 6) [Thocha mogyw (Manipari D, magn &¢.)], ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 43 ? 7 Bone. en ie nugri; Garo gring, kereng, Maring Fhru &c, [See Tibeto-~ lur, 8 Buffaloe. a Limb. sawet, Kir. Sanwa. Sawand san are probably the root used for **Cow”’ (b,),—et and wa being def. postfixes, &. New. me; Deor. Ch me, Asam mok (‘ Cow” Murm, mhe, mih, Sunw. i. Burm. me.) (8b). : e- Aka tien-dek, Abor men-zek, Inen-jek, men-jeg; Nagn teh. “Cow ” Ch, ta-gu, Turk. inak;? Fin traks, Ason.—* Cow” Bis. daha. 9 Cat. ; Yuma mim boi, Newar Blow (9). 10 Cow. | : e ‘ a, Sunw. bi, Limb. bif, vapi, Kir. pit, Lepeh, 5:4; Dhimal pia, Ka- ren hpi, Jing, Bengali ga-bhi (Dray. pei, peta &c., and see B,, Mon,— An, and D, ‘Tib,-Ult.) (10 @). ao* ; b. New sa, Aka shye. Abor sow; Mishmi ma-tso-kru ; Singph kan au; Gar, mashu, Pod, mosiu-go; Naga masu, masi, ust, moh &o.3 Kasia ka-ma-si; Yuma siya, tsi, teho, ma-chou (“ Buffuloe” Tib., Ultr.-Semitic). (104) . . . Chin. sua qu, sha gau &c. Ost. mes, Perm mos, mus, mys, Wog, misye. Lat. bos.—Af Sumali, Gella, Bishari &c. soah, sua, osha, *The Ostiak chosy, Kamchat kusha, Lap. husa, kos, Shangulla kasa, appear to, compe thé sibilant root with a prefix or another root. Comp: Singpho cansu, “ a e. Gur. myau, Mur. mie. These forms are” probably related to.a, See also ** Buttaloe’’ b. . : d, Mag. nhet. Niet is possibly the Turk, inek. (10 d), ll Crow. : Aka pak, Ab. pivag, piak, puag, Nega vakha, Yuma uak, wut, S. Tangk awak, Wasonar Maram chag-/ak, (11). ‘ ‘i Ason —Tug. avak, Yoko weak. , . 12 Day. : si Limb, len-dik, Kir, len, (12) Abov longe, Songp. kalhan, Mar. lan~ fa &. ' (8.6), [Manyak ding mi | (9) ['Fakpa syim- bu, } doa Mang, wo me ] (10d). [Sek pa salto, Thocho, “bull”, zyah (Yuma “cow” shya)}, (l0d) [Gyar nye nye, Gyami nyen, new) 2 Cl) £Yakpa akpo, (Maring af), Shook, nyag-wo). : (12a) [Vakpse ayen ti (Limb Jen dik), . The Horpa nye-le appears,” to show that nyen, leno, is the Tibetan nye with an » postfix], . As HTHNOLOGY OF THE 1NDO-FACIFIC IsLANDs, 6. Magar nam-sin, Suwn, na-thi, Tibetk. zhang-ma, Singphu ning thoi, Nay. asonga, tsing, Deor. Ch. sanja. Lungkhe sun Manip ngasun, fuaswag, asun Rakh ‘Sun’ tshan, Burm. tscheng, chung, Nancow. han, Bod. shyan, Garo san, rasan, Kol sing, The word is radically ‘syn? in most of these languages : ) Ug shun-du, Tangusian, “Sun”, shun, Ost, siunk, Semitic sham, hkhams &c. wh 15 Earth. a, Limb. kham, Kir ba-kha, Sunw. khia-pi (? Tib. New. cha, Tib, sa) Bing! nga Jili taka (whence probably the Naga kafok), b, Lepch. phat, Manip. lai-pak. (15 a). 16 Eqq. | Dhim fi, Naga ati, Limb. thin, Kir, uding, Ohisngl. go-tham, Ti- berk. tum, Dhith. tui, Bodo don-doi, Gar. ton-chi, Singphu udi, Burm. v, Kar. di, Milchanang li, lich, Asam koni, Yuma wati, adM, atui, Nag. ali, oli, nit, uisu, aust, hadsu, Manip, nroi-dui, maka-tui, hacku, atu, artu, wayui. (16), oe ; | 17 Elephant. . a. Newar, Chepang kisi, Sunw. so da, Abor-Miri siti, site, Manip. _ ka-sai, sai, Kar, ka-isho, Nag. sali, shiti, suti, tsw (see Tib.-Ult.) 20 Fure. Garo wol, ter, Manip.-2wan, tavean, Singph, wan, Nag, van, (Mon.« An,—“*Sun”’, “Sky”,) : ; 22 Flower. a. Limb phung, Kir. burig-wai, Sunw. pin, New, , Aka pung, Abor. apun, Rag. tober; chuben, nhopu, popu &c. Manip. D. abun, won, Pie; Burm. pen, Kar bpa, Guma pa, par, papa, Deor Ch, iba, Garo par, od. bibar; Dray, pu, puva &e. Ch. fa, hua &¢ Japan fana, Semitic ful. Af—Bagnon guefon Felup ba fan, Malag. vong, vono, vuna &e. Ason.—com. fonga, bunga &o, ‘Flower’, is not included in Klapfoth’s Scythic yocabularies.) Mag. sar, Lep. rip, Chep, ro, Manip, par, rai, lat, cha-va@ pen (pon &c. in other dialects) Nag. naru, nolong (? Dray.) * 23 Foot. a. Milch bung, Tiberk, dung-khat Manip. wang. Limbu lang dophe y raon dape) Changlo bi, Nag. upki, Bodo yepha, Maram, Songpo phat, or. chap, pa is a common Asiatic root, but in most formations it takes final 7, n, d, t, &, &e:,e- g. Korea par, Kash. dhoer, Sindh. per, Hind, pair, Sansk pada, pad, Europ, pes, vado, foot ke; Beng. pa,.paya, Semis tic pa’im, pa’am, Af —Malag. pe, fe, Gallafana, Ason.—Pol. wat, wae, fae (Manip.), Mak. bangkang, Sol opat, (15a) [Herp keha). , bis) [Gyar. kisat, Gyami chi-tun). o ETHNOLOGY OF THE INDO-PACIFIC ISLANDs. Eb 5b. Mourm. bale. New. pali, Gur, bhate, Abor M, ale, Mag. mihi, co. Suow. khweli, Kir, okharo, Burm, khre, khye, Kor: khodu, 'Tiberk. bong khut (Drav ) d. Dhimal khokoi, Kumi akok, akauk, (Tib.Ult,) 24 Goat. ; a, Aka shabam, Abor shoden; [Dhim eecha.] Nag. nabung, ne. bong, Manip. hameng [See Mon An J (not included in Kiaproth’s Atlas). Af Malag. beng. umby, Woloff biente, Ason —Iodon, bembe, bimi, embe. ambo, imbe &e,, kabimbi, kambing, &e. b. Sunw charsye, New. chole, Lep. saar, sarchru, Chepang miche, Dhimal eecha, Songpu zyu (sindhi chelo Hind). 25 Hair. a. Lep.achom, Mag. chham, Sunw. chang, New. song, Changlo cham, Manip sam, tham, kosen, Bongju som, Kuk, sam, Burm. ehhan- bang, Yuma tsam, tsang, chang &c., Toung-tha athwon, The Mon- Anain thwat, sok, tau’ is probably the same root, . Mong. usun, usu, chor-chun, Turk asim. Ason—Tobi chim. (The numercus other affinities of Tobi, Pelew &c. with the Ultraindian languages muke it evident that c/im is of Ultraindian origin and pro. bable that the 'l'urkish form once existed in the Ultraiudian province). b. Kir. moa, Gur. moi Chepang min, Dhimal mui tu (? Hind. mu), Bod khanai, khomon Garo kaman, Naga min, Burm chiding. ec. Nag, Aho, ko, |? Bod, Ahanai, khoman, Gar ka-man (‘ head’’y,] 26 Hand. Mileh. got, god, (Chang! ye Kir, chuku-phe-ma, (‘foot ’” ubjia- ro), Limb huk-taphe (footlang-daphe), Mag. hut pink, Chepang /et- pa, Nag chak, yar, Manip. D. kok, kut, hut &e Binua kokut, kokot, A Ugr. kat, kata, kasi, ket, Kit &c. Sam, hatte &c, Indon, Europ. hath, = 27 Head. a. Lepeh. athiak, Limb thag-ek, Kar, fang, Mur, thobo, Abor mi- tuk, tuku,-Nag. tek. Silong atak Chin thaukha. Sansk mastaka, Zend wedege. 4f,—Tum. adg. Ason.—lIndom atak &e. b. New. chhon, Burm khong. Nag. khang, kho &c., Gar, dokam, shi+ kam, Jili nggum ; : c. Gar. kra Bod. khoro, Mish mkura (“ Hair,” Tih. kra, Singp, kara). : ou Mag. mitalu, Manip. lu, als. é e. Tiberk pisha, Sunw: piya, Manip. pi; api, chapi &c: Turk a. : | f. Abor dum-pong, Aka dum-pa, Singph. bong, Deor, Ch. gudong, 28 Hog. a. Murm dhwa, thua, Kar tho, Korea tw, tot, Ch. tu, du, Ug. tua, fue, b, Gur, tili, Mish. badi. - 46 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 30 Horse. Milch. rang, New. sala, Chepang serang, Singp. kamrang, Burm, mrang, myen, Lungkhe rang, Kel sadam. Ug lo, lu, log, ; Ason.—ludon, jaran, jara, dala, ndala, nyarang. e. Mong. Tangus. Korea mourin, moron &c. 31 House. a. wepch i; ? Manip in, Yuma ing. (Tib.) b. Mur. yum, Manip, yim, Abor eum (Vib.) c. New .chhen, Manip, sang, Dhim. cha, Singphonta, Manip kat, shin &c. (Tib.) 31 Iron. @, Milch. pron, rung, run, Lepch. panjing, Limb. phenji, Kir Mag. Chepang phalam, Murm phai, Gur pai, Sunw wa akli; Uraon pauna, Jili taphi, Singph mpri, Maram kapha, b, Nev RE bg Kas TLP . ce. Lepeh pan jieg, Nag jian, jan, yin, yen, &e Deor Ch. sung, Burm san, than, Mishmi si, Manip nian, thin, tin, thir, thiar, Aka kakdhar, Kumi hedang Dhim chil, Bod. chur, chor, Garo shur, shil Tangus shelle, zhilla, solo &c. Af,—Suah. chunga Tigre achin. 32 Leaf. a, Kir ubava, Snnw sapha , nabar, Neg tuwa &c, Burm rwak, Lau bai(seeM A) - & Lepeha /op, Singph lap, Nag nyap, Tib tama, Dhim lava. ¢, Ahom anne, Naga am, Mishini nah, Manip na, thina,, panu, &c. 33 Light. a. Lep, aom, Changlo ngam, Sunw. hango, Aka hang tepa Chep. angha (,? Bhut dam,) i Lep. achur, Litab, thoru, (2 Bod, churang); (Tib, hur “wind’’, achur Mongol ‘air’’). ce. New jala, Murm., ajalo. 34 Man. a. Lepch. maro, Kir. mana, New mano, Sunw. muru, Aka bangne, Mag. bharmi, Chep. pur-si, Isthe Avur of Murmi not the same word ? miis Tibetan. (Burm lu, Drav, Vind. horo, ala male, oraon &c. Pashtu Sindh, maru &c. && a wide spread root, ; b. New mijdkg (mase ) Dhim. diang, Changlo? songo, Jili nsang, Nags nye sung, mesung, sauniak, Deor Ch. mosi, 8 Tangkh pasa, Kyo mash. Ug. chum, hum, Korea, sana; shanan Malay jan-tan, 36 Monkey. a. Sunw. moro (“Man"’,‘muru) Mish. tamrm. 6. Gur timyn, Chep. yuth, Burm., myauk, Yuma yaung, Manip: yong, kozyong, nayong, hayong, ying khayo, Dhim. nhoga, + ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, ay 38 Mother. Aka ane, Abor. nane, Singth. nu, Nag. anu, onu, Manip. anu, onu, 9 al 40 Mouth. a. Lep. abong, Abor-M. napang, napung, Nag: tepang, tabany, Kum. lndaung, Semitic pam &ec. Af Gal. afan, Dan. afa, Malag vava Ason —pang, ban, fafam, baba, Safa, fefa. b. Limb, mura, Mag. nger, Manip. mamun, chamun, khomar, kha- mor. e. Kir. doh, Mar.* muthu, Gond. udi, Murm. Gur. sung, Sunw, go, Chep. mothong, Newar mbutu, Kir, doh Gar. hoteng, Nag. dun, d, Qhangl. noang, Dhim. nui, Manip. ania, - * 41 Moscheto. a. Lep. mang kong, Nag.mang-dong, Manip kang; chakhang, timg- kheng, tangkhang, Burm. khyeng kkyen, Yuma kang, b Abor swnggu, Mish todze, Kar.patso, Manip, kachang, sangsan, thangtan thangkran, karehi, Bodo tham-phor, ce. Chepya Aka Mish ia Nag ayah, d, Changlo binang. 44 Oid. Muorm. chigu. Gur ehugu, New chikang, Mag. sidi, Chang] si, men- si Chep sate Mish sua, Burm achhi, shi, tsi, Yuma tsi; Nag tanthi, totsu, kakizw, Manip to-chai, Dhim ehuiti, Bod thou, Deor Ch, tu, Manip. to-chai, thau, thao, Karen tho, thu, Bongju, kerst. 45 Plantain. a. Lepch kar-dung Singph lungei. b. Limb la seh’ Ker guak sit, Murm mache, Mag. mocha, Sunw mu-Ai, Chep. maise, Mish, phaji, Manip ngachang, ngashi. 46 River. a. New. khusi, Dhim. Garo. chi, Ahor-M_ asie ( Water’’.) b Magkhola, Chep. ghoro, Kol gara, Uraon khar, Naga khar, Sunw. kha, Kum ta-gha , c. Bodo doi, Manip. duidai, tui-koak, tuthau, tu (Water). AD Skin. . a, Lepathun, - ‘ b. Limb horik, Garo holop, Kol harta, ur, Manip. ohwl, arhun, ¢. Murm di hi (Gur dhi) Singph phi, 50 Sky. Lepch: ta Zang, Sunw. sarangi, New. Mag. sardg Bod no khorany, Abor taling, Mish bra, Nag rang-tung; Male sarange, (Day &c.,) Rang. lang, lun, lungo kc, is a wide spread root applied also to at, cal 48% ETHNOLOGY OF THE INDO-PACIFIC ISLANDS: day, sun, God, Tib, Ugrian, African, Asonesian, In the more radical form /a, ra, it i8 still more common. ‘ e <= 52 Star. J ae Changlo murgeng, Singp sagan, sakan, Manip chagan. 53 Stone Milch. rak, rug, Tiberk rah, galhing, Nim. long, lung, lohong &c, com. (Tib do); Garo long, Aka elung Abor iling, ilung Mish mpla Singph nluny, talong Nag long Manip nung, lung, talo, tlulung, ngalung, Kum lung, dum, Car long, tu. ? Mong. cholon, chola (ordinary form of Seythic ual, kel &c.) gin« dhi rehans Af.—Saum. dugha, lugha Galla duya, daga, daka Amh, dengya ‘akna maludw, Ason.— Mille rakah, Tasm, lowai, loine, Aust walang,, marama &e, ’ ° 54 Sun. a Abor arung, Nag rang-han (See Sky.) 6, Bodo shan, Gar san, ra-san, Deor. Ch sanh, Nag. san, rang han, Singp tsan, Jili katsan, Kol. singi . Tung, shun, Ug. shundy &c. Semitic sham &c. Indo-Eur. sol, sun &e, é 55 Tiger. a. Limb keh va, Kir kiwa, Dhim kkuna, Nag Au, hayi, takku, ys Manip kai, takhu, chakwi, khu-bui, akhu-li, sang-khu, sukwi, sa- hu, é Magranghu, Singph sireng. ¢ Sunw. gupsa, Chep ja, Male sad, Nag. sa, sahnu, chianu, Bod mocha, Garo matsa. . d. Ab. simioh, Aka samnya Mish tamya, 57 Tree. a. Lepch kung, Manip thing-kung, Kumi ahung, tagom, Simang kuing. b. Tiberk pang, Bod bong-phang, Garo pan, ‘Deor. Ch. popon, Burm, apang, apen, ~ingphu phun Nag bang, pan, pe, Manip thing-bang, sung- bang, bing-bel &e. Ug pun, pu, fa &e. Sam: pu, pe, poi, Tungus mo, mo, Pushtu wana, Af. vahad, Ason.—pon, puang, pohon, puna ko. . 58. Village. a. Murm namso, Gur nasa, Nag ha. 6 New. geng, Sunw gaun. Lhop. Lepch kyong. This vocable is wide- 1 spread in Ultraindia and Indonesia, but frequently applied to “fam ily’’, “ tribe” &c. | ‘59 Water. a. Kumi, Kyon tui, Kyan fuwe, Mrong fei, Lungk. fi, Rakh. rif re, Kar biti, hfe, Chep, Milch. ¢i, Magar di, Cheng ri Murm kui tui (see T, U.} 7 fife wt jet i 4 ETHNOLOGY OY THE INDO-PACIFIC ISLANDs. * = APPENBIX TO SEC. 6 OF CHAP, V., PART It. AA8 wo, sr \. COMPARATIVE VOCABULARY OF SEMITIO AND AFRICAN NUMERALS; . I. SEMITIC Acs the Semitic aystem occupies a peculiar place with relation to the Asiatic on the oue side and the Alri¢an on the other J shall give it separately, referring tothe Semitico-Alricun list for the detailed state- ment of the affinities whea they are*both “Atrican and Asiatic. , One. a. (a.) wa-h-id, or wa-h-ed, wa-h-ad, a-kh-ad, mase.* wah-id-a.t.a, a-kl-ad-i fem. Arabic, e-kh-ad” muse., a-kh-at fem. Hebrew, kheod masc., kh-ada fem. Chaldee, ta-ut Mahrah, Gara, vo a9, va Egyptian, The Semitic forms are most closely connected with the Ugrian, in this agreeing with the Indo-European, [Semilico-Atrican J, A.] Asa delinitive, separate or concreted with other roots, a 3d pronoun, a demonstrative &e, the most archaic Semitico-Libyan form o! the root A appears to have beep the aspirate and sibilant, passing into the dental aud guttural, ha, sa, tu, khu, ku &e, i Asa unit the purely aspirate form is found in African languages, but itis rare. ‘The sibiJant is the most common, not only in African sys- teins but in the higher numbers o! the Semitic, from which it follows that Semitic dielects had originally s forms in 1 also. The variations were simply the definitive in its different forms, ‘Thus forws similar to the Babylonian su-va mase., su-a/, fem, “this, he, she’’, su-2z, su-na, sil-n, ‘pu-nu-t, su-na-t plural (consequently not | but 2 as in sa-ra 20 and the common Semitic 2 she-na &c.), the Egyptian en-tu-/ mase., en-tu-s fem., en-te-s-en, s-en pl (3d pronoun), the Arabic ho-wa masc., bi-yatem., ho-m pl. masc., ho-nne pl, fem. and other Semitico-African forms of detini- tives are found as units in Lor in higher numbers. Hottentot, which has the guttural form of the unit, shews the archaic range of definitives and consequently of numeral elements very clearly, quei-mb masc., quei-s fem., quei-hw pl. masc., quei-tee pl. fem., quei-ma@ or qu-an pl. com. Many of the numeral terms bave a redundancy of detivitive elements, , The primafy proviowinal postfixes appear at a remote period to have become concreted with the root, when new or secondary postlxes were puperadided, Thus the nucleus of the Semitic 2 was the detinitive with its dual or plural postfix th-m, sen, t-r. At luter stage a secondary plural posthx was assumed asin ath-in-ay musc., ath-inet-an fem. Arabic, ta-r-t-en Chaldee, Sh a dad Hebrew, (in which the original na is elided although preserved in the masc. she-na-yim where the two plural elements are in juxtaposition as in the Arabic term.) Ln the current forms of 1 the dental postfix only is used. This isthe © definitive itsell, with or without its secondary feminine power. In other terms the liquid plural na, Ia, ra &e. and the Jabivl masculine aré preserved ag posttixes, and it is pret that in the original Semi tic system they might all be used in | also, the plural power of the liquid being sécondary. mine ‘ < (6.)° The sibilant unit takes the liquid postfix in 10 ash-ar, as-ra, Ar. — oo _ * Some Provincial and extra~Arabian forms are included, to shew the phonge tic variations to which the terms are liable. ; ” ?. ETHNOLOGY OF TNE INDO-PAGIFIC IsLANDS,: 7 which appears to be preserved as a pronoun in the Hebrew relagive ash-er. Terms tor 5,10, 100, are nnits in mort lnogunges, and This - term is still used for 1 as well ag 10in Atrica, [Semiieo-Atriean Nomerals }. A g.). The same form overs in the semitie 8. The ow form of the posifix, varmble jn some dialects tor and in Atricen ones tod, occurs in-2, with itsdaal or plural farce, but hero it is probably to « be conaislered as the secon elewientin a compound of two definitives (¢.)) The labial (masculine) posttix oceurs jin terma which must have. * been originally unite, In tho Semitic system this form first appeaia in » the highest term.ol the primary quirfiry system, 5, and im Alrican sys- tems it is tsed in 10 as well ax 5, In Egyptian ib is yreserved in. 3 and. it enters with the same power into.the Semitic. and Reyptian 8» Tf the- Jabial hal occurred in Band Bamly, it might hive heen consilered asia second radicul defivitive in a compounml, but in & and Pir nnust represent au archaiv form of the-unit. In African systems it is fonud inother numbers as a postlix or prefix, in the sume mode as it oecars attache) to substan tives and. qualitives. [See the remarks on 3..AL) > ‘B a wal mase., ¢ wali fem “first”, Ardbic, [par Persian.” | ' | This archaic unit is preserved in A‘rivan languages os a eardinal terin. Ih Semitic it vectrs ns suvhin the contracted form avin 4d. It. isa N anil E. Asinv and Druviro-Australian definitive and unit, In the Semifive- Libyan formation the Iubiul wes an important. archate definitive. tis largely preserved as a postfix and prefix As a S71 pronoun wid demonstrative, the sibilant, deatelor guttural cetini: tive appears to have early gained axelgsive currency. The dalial, how<9 ever, keeps its plore even in some separate terms, as In the interregative | pronoun, mi, mal eb, rosa, man, wa Ar, andin the Gonga 34 pron. i. !tigalscrehtivein Arabic, man, Gara, mon, Mobrabl, moll. In finusa itis demonstrative Ovong-ga ‘this’, won-an that’ )and rebtiye (won-th won ne) fy thy Zinhian fonily the libial is one of the ohiet dehnitivesy: Bd provotns iil denmoustratives, Froyea very early period the labial defi itive wequired w wesenlive jiowerand it i possible tat (le horberal wa-} an the ailied Atiitn basri, ma-ri, viet, pa-ka, ha-si, mo-si &G, - were lorie alter it hid nequived mat power lar while secunidary dofi- pilives might ‘be jseposed as well ‘as postposed. ‘The sibilabe and deutel was comodn before it becaine temipine, and the Wasculfve fuice tion of the hebinl may have been 9 consequence of the other principal definitive becoming fewinine, Butitis more probable ahat the Jnbial wusdhe first to regniv: wsexaal (masenline) power, and 1bet as a nn meral eb ment in the archaic Semitiio-Libyan sistem it is to be consi-« dered as mascujiar, ‘the other definitive certainly occurs in that system a8 i: the pronominal hoth asan archaic common aydas a later or secondary ieminine particle. [See Semitico-Alrican Numersls 2. B. e.] Two. ith-na-ni. atli-iman, is in-in, s-n-in masr., oth.in tan fem; Ar, shee De, shie-ne.m, sient yin, mase,, sih .to+yim, fem, Heb, sroh. Mahroh, te ren mise, ta-r-t-éen fem Chald (s-an-va or saen-na 20 Biliy. loninn), seu, s-en-fey sn fi, snov-s Ee. Copt,, the-ne¢ Berb., (si-n ith 12, 2) Ae )s shen Shillah, up Bullum, & ileal Kulobi, ‘Kt-le-tes kive-lj-ce, quirle-/, bu-l-et, Abysyy he-lita, Gatat, ki-lli Aukiko KPHNOLCGY OF THE INDO PACIFIC ISLANDS! - & The initial sililant (ith, she, sh, s, the, variable to klie, ki, he) is fonnd in Baul othemtertis, woud tie t and k of 1 re only varivtion of it, In the Pade Buropean 2, 8 tod 4 i OCeUTs in the dental orm t, d, ands inesey thie and wtherN. aot Ep» Asian Janguages ass, t, k &e,) In the ludlos Bury peso and several Sey thie forms (2, 4) a lnbiulis inters posel between the initial and the faa elements. In the ToigsEds rope 2 the lubial only is preserved, ind bor ile jobial and the Beal, tevaier &eo fo the absence ot the labivk Semitic resembles the Cauca. sian, (Mingreliat sli-ri, whieh has the Agibic vowels, Samoiede si-riy ai-de), Ik thes) waeesiau gear, Moncolian k-e-yar, (d-u-r jin 4, z-i=r in 6) the infiuence oa lost hebial is perhaps sill felt in the broad vowvl, As awessemialbelement-ot2 n, 1, ris counnion in N, and cB, Asian onmuwerals, © divese ditsoit in the apparently controeted il, ni, urh 2. Waring, whielr hus dost hin 2, preserves iti 4 ni-la, nisl, ni-leit- nid oin 8 ies ke Aino hes itin 4 i-ve, y-ne, which wppears te be a contraction o@ the ullderm preserved iv the Yeniseian 2 kisi, also his ne, iene, hd ie the Aino d zieut, Pye inst, ike the Kamechatkan ki-ni? is dowhtless y remoantef the original system in which the term was used asaunit. [tis a -rythic dd pronoun, The posthix na, ue &¢e, of Vis the Semitico-Libyan plural definitive, . The numeral s-0u, s-en, she-up, uth-in wor thereture ja all probability the plural orm of the detintiive, and Weuticn) with su-pu, s-en Se. In” otter ormiutivns also the liqhfd vetinitive is not only” a frineljal ele- mentiu 2 (and olteh dy higher daal wumbers, 4,8), but ia piutab oP® dun} particle. Iw Arabic it'is dwii as weil as plural. 1 the sibilant” were considered as an archaic prefix asin Ziwbien, as in the Malagasy” iz-tho “1, aud asin many sabetentive words of ‘sewitico-Libyan® glussuries, na would! heeonte the radical element of 2, | The wide prevalence in Avrica of a labial terin, full and contracted, (ba-ri, -ba-r, wer, tru-l-ud, biert Ke. Se.) and the porsistence ola siitilar team nthe semitie 4(2 dual) reniers it probuble Mat it was archaically a tera) for Zin the Semitic family, or nm thal western brench which firs’ gave numerals to Africa. See Afriun mumerala 3,4, 7 (5, 2), 8. ahd Semitic 4,7. be ibe occurrence o: the lwbial Lath im 4,. an | 2 the aréhsi: Semitico-Libyan system rveoembled the cy tui und the Draviro-Australieu. : Ay Three. Aw Vhe Egyplian sho-men, shame-t, sho.m.t, ‘sho-m-ti, sho. m-te OppRars (a preserve & ler Gnve possessed hy Semic. Ji recurg” fe the upit in the Beyptian 10, withaul the sibiivnt prefix. meu-/, mot me-t, me-fi, me-te, and ia tie Bgyptiioa and Semitw 8 (5 3). The Epyptian 8 is shun, shomen, shemate, shome-ne, sh meni, 80. h-me-ne, kivme-ve, kheem.ne. ‘The Semiic & is sheemon, tha mun, tha-man-id- ¢. Bishwri like Evyption lias the labal in 3, mib, as well aa in &, + u-mhoi. ‘ = A siwilar term was used inan archaic N. and E, Asian system it, whieh the labial was thy qualitive postfix aud the sibilant the numeral root, primarily definitive aniupit tu the archnic Semitico- Libyan stom the labial does not appear to have been qualitive, but masculine, liein, sh-mn é&e. is the masé, form of the silihet unit, 3 Veing very commonly’ a nit (properly 2, 1, but asin other terius one of the -wo % ETHNOLOGY OF THE (NDO-PACIFIC ISLANDS, was early dropped for brevity’s sake). ‘The superadded t of Egyptian mikes the term fem, a . Asa onit the sibilant reenrs by itselfin the Egyptian she 100, and sha, sho 1,000, and it. is the inital element in the Semitico Eyypuan 2, Semitic 3, Semitico Evyptian 6, 7 and 8, corresponding as We bage seen with the Indu European t, 0 of 2, 3. 4. with the Lodo Kuropean silent of 6,7 and 8, and with the N. and E. Asian sibilant, dental and guttoral unit of | and higher nombers. If the sibilant be considered as a prefi the root becomes the labial. ® ; | In 3 the same archaic form is preserved by Caucasian, 8e-mé, 8a mi, 80 mi, ju mi Lesgian cha b go, sha-l)-go, sha-mmba, ch’-ba, Chinese sae mi. The sibilant alone recurs in Ciréassian (shi) Korean, and Kamechat- kan, The broad form of the dental reeurs in Yeoiseian with the postfix dong-em ‘The comman double form of the Seythico-Chinese sibi- lant, dental and guttural unit oceors in the 3 of Ostink, chud-em, In Scythic and allied N. and E. Asian languages the Jabial definitive and postfix orcurs in the forme men, man, mon, me, ew, m &c. a6 in the Semitico Egyptian 3 and 8. : The first element of the Egyptian term (she, sha) is the same as that of the Semitic (tha). The vowel isa variation from that which the purti-, ele has in Egyptian as a definitive tu, su, and in 4 {-tu, in 5 tu, in 6 sou &c., and in 1,000 sho, ; . In the Nilotic languages generally the form in ooru is equally common with the Semitic in 9, i, e. ‘Thus the Berber 5 is su-m-us, the Dalla du-su-me and the Darfur us. Bishari has also su as the unit tor 5 in G (su-ggoor), 8 su-mhai, and as the unit for 10 tu-mmun, lis se- veral terms in the Galla group the same form is followed, to-ko J, su-dide 3, koen 5, tu-r-ba, t’du-bah 7, su-ggal 9, ku-dun, tu-ma &c. 10, Gonga and Malagasy preserve the form in Li-so, Gonga in 5 huch, in 6 ho-su, in 8 hos, andin 9 ho-da. Even the more purely Semitic Jan- uages of Abysinnia have the’ Peyptian form in some terms. Ambaric u-l-et 2, so-s-t'3, au-mi-st 5, su-blia-t 7, Tigre shu-ha-tt? 7, sho-mun-é 8, Harragi su-t, su-d8. It fs needless toadd examples from the more west- ern African language. The connection between the African and the Semitic numerals is mainly through Himyaritie, and it is probable therefore that in the early form of the Hinryaritic the definitive and unit was prevalent in the u aud o forms, as well asin a, i, e, for the Jatier are found iw Egyptian, Gonga and other Alrican systems current along with the former, The modern representatives of Him) aritic preserve severn) examples of theo, u form of the definitive as a numeral elemen’, both principal and accessory, ta-ué J, s-roh 2, ar-bu-ud 4, (arr-ué Amharic), kho-m-sa 5 Mabrah, shu-a 7 Gara, thu-ni 8 Gara, Babylo- nian has it in su-su or su-si, 60, which is similar to the Amharic so-s 3, As a definitive the broad form is the Babylonian 3d pron su-va mase , su-at fem., suenu, su-na pl., -su poss, post, which agree with the vu- meralform. In Hebrew and Arabic it is preserved with the aspirate cénsonant hn, ho-wa. In the Hebrew zo-th tem. this”, the sibilant is preserved. With these forms the Egyptian su, tu, Davakil us (3d pron musc.), Hausa su (pl 8d pron), Galla tu, ku (demons.) agree. It appears therefore that the African forms of the unit in U, 0, correspond with the Semitico-Libyan definitive and with an archaic form of the Ssuiltic unit, . - ; ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, a] The variations in the vowel were probably to some extent flexional, U was an agentive or nominative postfix in the archaic Semitic system, It may also in some cases have been a soltened form of the masculine postfix, fis feminine. Itisalso possessive A plural power cannot of course be ascribed to the u or i of the definitive when used as 1, B. tha-la-th, sa-la-tha, sa-la-sa Jfem,, thi-la-tha.ta mase, Afi, she-lo--kah mase., sha-lo-sh fem. Heb., (si-la-sa, 30, Babylonian), te-la-ta mase., te-la.t fem. Chaldee, se-le-ste Tigre, tha.th-t¢ Mahiah, tha-k-it Gara., ke-ra-d Berb., k-ra-¢ Shillab, (so-s-¢ Ambaric, to-s-% Nubian, ta-tu, sa-tu &c. «ec. Zimbian &e.) In sa-la, she-lo &c, the sibilant unit is followed by the liquid la, Io &o. Radically the compound may be the same as in 2,—1, n, r, being variations of the same definitive in the Semitico-Alrican as in the Scy- thic systems. From the Himyaritie and Zimbion terms, and trom the occurrence of such forms as su-su in higher Semitic numerals, it is clear that the radical term was a double or reduplicated unit, which varied from sibilant and dental to liquid forms, s-s, t-t, r-r, I-l, s-d, a-r, 3 |, ter, t-s, k-r &c. The two forms, the sibilant or dental, and the li- qultiey ine their combinations, must have ve-existed from a very remote period. , The primary form su-s, tha-th, tha-k &c., connects itself with a very common double form of the unit in the Seythico-Uhineze systems. The variation of the 3, t &c. tol, r occurs in these N. Asiatic systems. In the Koriak 3, which is not reduplicated but is simply the sibilant unit, jt varies from sho, so, to ro and 1S: The ludo-Zuropean t-ra is a similar form to the Seythic ko-r, ko~l, ha-r, ku-j, chu-d, and to the Semitico-Atrican forms in t-r, t-l, k-r &c. If the initial, in all these terms, be considered as a prefix, the simple unit remains tlre second and radical element, as in the remoter E, Asian systems, Chinese, Koriak &c. Lf both elements be considered radical, and this appears to be the correct view, the term is still merely one of the archaic forms of the unit and defivitive. Comp. the demonstratives zo-th Heb, fem.,i-za-to Mala= gasy, dza-ka, dza-li-ka ta-ka, ta-li-ka, fm., Ar., thi-na Berb. f., za-na Gaila f. The / form of the definitive, although common asa single particle in the Semitico-Libyan languages, occurs rarely with the preposed sibilant, the common form being n. The Hebrew ha-lla-zeh m., ha-lle~ zu f. “that”, isan example of a demonstrative compound similar to the Semitic 3. - Four. A. ar-ba fem., ar-ba-da mase. Arabic ; ar-ba-ah masc., ar-ba fem. He. brew ; ar-be.a’ maec., ar-ba fem. Chaldee (ir-ba-ya 40 Babylonian); ar- ba ud, Mahrah, Gara; ar-ba i Tigre; ar-ba.ta Gatat; a-f-t, Flu Bg, 5 sr-af, ar-ut Ambh.; u-bah Arkiko; fou-so, ‘Vibbo, fu-du, hu-du &c. Hausa ; fu-lu Kalahi; hau-da, av-da &e. Gonga ; a-tu-r Saumali; fere Dauakil; e-fa-r, e-fa-tra, e-fa-d, e-fa-tu, e-fu-tsi &&c, Malagasy ; Zimbiua wa.na, wa-n, ba-na &c. These are terms for 2, i.e. 2 dual. The Semitie collocation, it will be remarked, appears to follow that ofthe cogoate Galla, Malagasy, Inde. G ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. European and Scythic term for 2, The Egyptian, Sodanian, Galle and Malagasy collocation foll: wa that of the more prevalent African 2 (Zimbian, Nabian, Nigerian) which is also Scythico Ausiralian. But as the labial is a postfix in the archaic Semitico-Libyan as in the archaic N, E, Asian, and the initial ar of ar-ba has the same elli)tie ap, earance whioh ar, an, al, ir, il &c. havein the Scythic system, it is pro. ble that ar-ba, like them, haslostits original initial consonant or prefix. The general Scythic affinities not only of semitic but ot all the other S, W. numeral systems of the Old World,—African, Euskarian, Cauca- sian and [ido-European—refer us to the Scythico-Chinese province for illustrations of the Semic numerals, and an example of an allied term is probably extant in the’ Mongolian dor-bo, dtir-ban, tir-ba &u. (in Turkish, with a dental postf., dor-¢, dur-t, dwa-ta, the r elided in the last asin the Indo~European dwa, 2), in Indo-European cha-t-va-r-as, with the sibilant post, The Mongolian ar-ba, 10, (also ar-ban), is o precisely similar term to theSemitic ar-ha, but although the form of the final consonant and the postfix doultless identical glossarial y as well as phonetically, the initial consonant of the root may hove difter= ed. The Mongolianterm is probably a contracted uniblike the Cauca- sian ar—, er—. ‘The Semitic must be reterred to a term for 4 or 2, The Georgian r-wa, r-uo, ar-a, ovr, 8, [4 dual], appears to be a similar elliptic term, and the Mingrelian bar, bar-| probably preserve its lost initial. In the other Caucasian languages itis also the labial in other Scythic forms, m-itl-go, be-itl-gu, m-ik-go, me-i-ba. The Malagasy e-far, 4, and the corresponding African terms for 4 and 2 preserve the full form ofthe Sewitic ar. The Danakil mal-ud, 2, bas the labial postfix as in ar-4a, and the Malagasy r-na, 2, is probably a similar contraction of far-wa or faru-wa, resembling the Geogian r-wa, The labial definitive postfix ocewrs concreted in the Semitic glossaries as well as in those of the allied Libyan languages, and Hebrew has it in the modern term for 2. (For the evidence of the wide prevalence of the Li-« byan labialin 2, 4, 7, (i. e 5, 2) and 8, see African Numerals.) The Egy p= tian f-tu, the cognate Alricau terms in d,s and r, and the existence of nearly all the varieties in Malagasy (r, d, tu, tsi), corroborate the in- ference drawn from the Semitico-Airican terms for 3, that, in the archaic Semitico-Libyan, asin the Seythic, definitive and numeral system, the detinitive and unit in}, r, n was merely a variation of that ins, t, k. Although the liquid r, 1, n was early combined in the Asiatic systems with other definitives (labial, dental &c ) in 2, 4 ke. it appears to be the essential element in the Sythic, Semitic aud Alrican systems. Frve. A. kha-ne-sa, kha-m-s fem., kha-m-sa-fa mase. Arabic, kha-m-ish- shah masc. kha-m-esh fers Hebrew, kha-im-sha mase. kha-m-esh fem Chaldec, kha-m-is-t@ Baby lenian, kh-ish Gara, kho-m-as Mahrah, aem-is-f, au-tm-is-¢ Amharic, au-m-ish-te Figre, su-m-us Berber, su- m-os-t Shillah, tu-m-at Timani, bu-su-me Dolla, @-m-us Arkiko. This term is probably a unitasin the African and Scythic systems. Radically kha-m is identical with the sha-me, sho-m, tha-man &c. of the Egyptian 3 and the Egypto-Semitic 3 of §, But it is remarkuble that in the proper Semitic languages the unit roottakes the gutture! _ ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 7 form as in the Semitic 1 (akh—), whilein 6, 7 and 8 it returns to the sibilant form it preserves in 2and 3. . The older Atrican terins—the Berber, Shiflah, Timani, Dalls—retain the sibilant and dental torm ofthe initial ‘unit, wad the Egyptian (B) has it without the lvbiul. [See African Numerals, 3, 5.) The term is similar to the consmon Seythic unitin k,t,.s &c which appears in 1,3, 5 avd hizher numbers. Examples of its occurrence in 6 are ko-m-leh Kamschotkan, which reappears in the sibilant form in the Samoiede so-mba-lach, so-bo-riggo, sa-m-lik, and in the dental form in the Mongolian ta-bun, ta-bu.* The Kamschatkan and Samoiede terms afford examples like the Semitic of a secondary postfix, and show that in these languages also the labial had lost its primary qualitive force an’ merged in the root when the native pustfix was superadded. B. The Egyptian tn, tiu, tie, in 50 taiu, teui, was probably a native unig derived trom the dental definitive and demonstrative (comp. en-tu-f “he,” en-tu-s she’, su “he &e.”’ ta, ti, te “this” fem , tai “this,” tui - relative tem.) In the analogous form su it was probably the oldest form of the sibilant, dental aud guttural unit. It is still preserved in the Babylonian 60, su-su or su-si, and in the initial elewent of the Semitic 2, 3, 6,7 and 8 the sibilantalso keeps ts place, In the older Atrican forms of the Semitic 5 (A) it is alsu retained, as we haye seen, in the forms su and tu, | Six. 6 is simply the unit—for 5, | - in the prevalent sibilant form. Egyp. tian preserves the lubial postfix of the archaic mother system 8-ow or s-oo, 6-00, 8-0, but in 60 has the pure unit or definitive se (as in 100 she, and 2000 shu). The Euskarian sei isthe sameterm. Semitic has si, si-ta, se-te fern., si-ta-ta masc. Arabic, shi-shah masc, she-sh yen Hebrew, shi-tta masc., she=t fem. Chaldee, sha-t-id Gara, ha-t-id Mabrali, se-de ist Amh., se-d-ishte Tigre, se-d-is Berb,, su-th Shillah ‘These terms are the same as the Indo-uropean sbu-t, sha-sh &c und the Scythic double forms of the sibilant, dental and guttural unitin 6 (chu-t, ku-t &u.), 7 (sis &c.), KO Xe The Himyaritic and derivative African forms shew that the s-cond sibilant or dental is not a secondary postfix, and that the Semitic term was immediately derived from the double unit. This form was an archaic Semitico-Libyan term for 1 probably te- winine, but ns it is best preserved in a ternary series (3, 6, 9), the Semi- tic 6 may be 3 dual, Seven ga-ba, sa-be fem. sa-ba-fa masc. Ar,, shi-balb masc. she-ba fem Webr., shi-beg mase. she-ba fem Cheld, ha-ba-id Mubrab, sh- va’ Gara, siu-ba-te Tigre, se-ba-t, su-bha-t Ambaric, se-i-ech, Eg.; se-¢, se-d Berb shash-t Eg. zos-pi Eusk This is the Indo- European sa-j-¢ and’ Ugrian sa-b-et, &c. in which the basis sa-p, su-b, si-m &c. is the sibilant unit with the archaic labial qualitive posttix, The Egyptian and Euskarian unit preserves the double form found in’ the Ugriam sis-im &c, in the semitic and Indo-Enropean ®6jand in ‘seyeral Ajrican terms, The term, in its original form, was 6, 2. See Iudo-Europeau and Seythic Numerals) A Similar form of the unit is preserved in the Egypuan 3 and Semitico-Bgyptian 8. 7 Pal : ETHNOLOGY OF THE INDO-PACIFIC ISLANDS: Eight. tha-ma-n, sha-ma-n fem., tha-ma-ni-fa masec Arabic, sha-mo-n-ah masc., she-mon-eh. fem Hebrew, te-ma-n-ja masc., ta-m-n-e fem Chal- dee, tha-man-id Mahrah, thu-m Gara, shu-mun-te Jiygre, se-min-t Amh., sh-men Eg., t-em Herb , t-emp-¢Shillah, This term is evidently not formed from 2or8, but from 3 in the Egyptian form (i. e. 5, 3, as in all the African and many other systems), : Ws oa Nine. ti-s’, ti-sa’ fem. ti-sa-ta mase. Arabic, te-sha fem, ti-she-ah mase, Hebrew, Chaldee, sa-id Mahrah, Gara, ze-tti Amh, ze-te-in Harragi, za-te-na Gafat, tish-ale Tigre, y-sit, p-sis Eg (p-isin 90), dza Berb., trdn Shillah. This is a Scythic form of the unit, occurring in the Kamschatkan dys of 1, Hungarian tiz 10 &c, and in the Mongolian dsi-sun 9, (1, 10). It is also the Semitic 6 an'l 3, so that 9 is probably 3 trina!’ But asthe African terms are generally 5, 4, it is possible that the Semitic 9 is the term for 4 found in several African languages, and recurring in 9 in forms similar to the Semitic, Agau si-za, sa-dja, se-dza 4, tsai-cha, se-ssa, se-s-fa 9, Gonge ach-ech 4, dje-ta, yi-dea 9; Shangalla zaa-cha, an-za-cha 4; sa-sa9 The full terms are pre- served in some Zimbian systems Makua dialects.ma-che-che, mu-tye- tye, i-tye-tye 4, ma-tanu na ai ma-che-che, mzana-m-tye-tye, uhyanu na i-tye-tye, 6 and 4(9). From these terms it might be inferred that the Semitic 9 was also a term for 4, but it has noresemblance tothe cur- rent 4 either in its contracted or full form (ar-ba, war-ba &c) It ap. pears to be related however te the current term tor 2, and was probably one of the forms in use when the numerals varied regularly with the ender of the noun. The Hebrew fem. she-ta-yim is a similar term, In e Semitic terms for 3 a similar variation occurs, Arabic, Hebrew &c. having tha-la-th, she-lo-sh, while Mabrah and Gara substitute the den. tal and guttural for the liquid tha thét, tha kit, and « Jike form appears to have existed in Babylonian su-su 60, with which the Am. haric 3, 80.8, is cognate. Ii the Semitic 9 be considered ag 3 trinal its resemblance to terms for 6 and 8 is explained But even in the cure rent terms for 2 and 3 we have found a radical resemblance, so that a resemblance between 9 and 3 or 6 does not oppose but rather confirms an identification of 9 as ultimately 2 dual. Whether 9 be 3 trinal or 4 it agrees radically both with 3 and 4 because these agree radically with each other. Ten. : A. ash-ar, ash-ir fem. ash ara-fa fem, Ar., 98-ir-@ mase., es er fem, Heb, as-ra mase, as.arfem Chaldee, ai-ish.r-id Mahrah, ish-reid Gas . Fa, aseur-fe Tigre, as-ra, as-ir Amh. B. men.t, me-t, mn-t Eg. This is the Scythic labialunit end post. fix, occurring os 10 in the same form in Tungusian menz (in 1 min), But the Egyptian term is evidently the second of the definitives found in-8 and 3 and here divested of the initial sibilant unit, which it retains in some other Airican forms [See Airican Numerals, 10 B, 4.) ETHNOLOGY OF THE INDO-PACIFIC ISLaNpa; g APPENDIX TO SEC, 6 OF CHAP. V., PART UL . . B SEMITICO-APRICAN NUMERALS. One. r _ A (a) wa-hi-d, wa-hi-du, a-ha-d, a-kha-d masc. wa-hi-da-fa, a- kha-di fem. Arabic, e-khnd mase., a-kha-t fem. Hebrew, khid mase, kha-da fem Chaldes, hha-di, a-dde Tigre, bhe-d Harragi, énd Amharic. * The rootin these formas is hi, ha, kha, a, variations of the “emitico-Libyan definitive, and the other variutions, si, ti &e., were doubtless archaically current us the unit (6). ta-nt Mahrah, Gara, ta-¢ + uria Muria, From the analogy of ar-ba-ud, 4, this analysis is clearly the correct one I) the root hus no vestige of the labial, it is the Semirico-Lihyan dental definitive, also occurring in the sibilant form in the higher semitico-Egyptian numerals, and in the dental formin the Egyptiun 6. . (c). uo-t (or u-ot) Egyptian, (also u-ai, u-a, u-ei, u-i, comp. the demonstrative ai, ei, ¢, i, in pai, tai, nai &c.) If the guttural and aspirate tm (@) be the root {li, kha, ha) and the vowel or labial (wa, a,@) a mére augment or prefix, it must be ranked with the Himyacitic ta and Egyptiun t, because in Semitico-Libyan the aspirate and guttural definitive and unit passes into the sibilant and dental. In the Semitic 5 the unit recurs in the aspirate guttural form, In the allied Scythic systems the guttural, dental, sibilant &e, appears rei had an independent definitive power even wheu postfixed to the labial. . In Africa the aspirate, sibilant, dental and guttural definitive also oc» curs as the anit inaccordance with what we have indicated as its arvhaio Semitic range. Even in the same groun the consonant assumes dif ferenttorms. This inthe Gonga group we have.— ' (a). i-ta, i-so, i-sta, i-ka, e-koe. In this double form the final element is not to be considered as a postfix, because the definitive itself occurs in the same double form, and the initial element rather than the final appears to be aservile. Inthe Egy) tian 3d pronoun, en-tu-fmase, an-tu-s, en-te-s, ber the nasal ig prefixual as in the 2d ani 1+ pros nouns, and the dental is the radical definitive as in ta, ti, te, “the”, fem. But in definitives and units the prefix being itself definitive both ele- ments may be considered ravlical. The unit is geverally not a single defi- nitive but a double or intensive one, being a numeral application of a de- moustratire or 8d pronono in which two definitives are usually come bined. The recurrence o! the dental definitive, variable to the guttural, as a postfix io higher numbers in most of the Nilotic languages does not appear to reduce it to a mere postfix in 1 and raise the initial ele« ment to the chararter-of the sole anit root. In some of the other Nilotice languages the initial element assumes its full nasal form as inthe pronouns. Danakil in-ike, Shiho in-ek, Bisharye engrat or en-gat, Tumali in-ta These are clear vestiges, found from the Wed Sea to the western portion of the Nilotic province, * In App. A the vowel of the root is improperly separated trom it and carried to the postfix. 2 cf : 10 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, of the ancient nse of the double definitive as the anit, The Darfur d-ik is a variation of the same form or of the cognate (d), The Malagasy i-sa, I-so, i-si corresponds with the Gonga sibilant form and the Se- mitic hi, ta &c, In Asonesia several varieties are preserved, e-sa, i-ae, a-si, a-sa, i-cla, ji, i-ta, ta, ta-si, ta-hi, ta-ka, sa-da, sa-ra, se-ra, Bi-ti, me-isa, ma-isa, m-esi, sa-mo-si &c. Some of the Nigerian lan- guages retain the nasal prefix in higher numbers, and prove that it was prefixed to the numerals thronghout in some systems. Wolof has it in 2, dand4.+ Some other Nigerian systems bad the labial prefix. It is preserved in some of the Ashanti and Gabun dialects (bi, mi, mie, ba &c.) In the Zimbian systems substantival prefixes are used in the substantive form of the numerals, the possessive rendering them ordinal, | When used as cardinals or qualitives they take the de- finitive. prefix of the connected substantive. Thus in. Kosah 1 has the substantive form isi-nye (so zsi-bini 2, isi-tatu 3 &c.) and the qualitive forms um-nye, li-nye, in-nye, si-nye, lu-nye, bu- nye, dAu-nye, and higher numbers take the plurals aba—, ama—, e- zin—, ezi—, emi—. In some of the published lists the prefixes are confounded with the roots, particularly in the common contracted forme of the prefix as in mnye 1, mbini2, ntatu 3 &e. In other cases the consonant of the pretix is elided and the vowel only preserved, W hen it is recollected that in the archaic Semitico-Libyan formation definitives which in their variations embraced the whole range of con= sonants, might be used either as prefixes or postfixes, and that both are found in many words, the difficulty of analysing and comparing the Semitico-African systems will be understood Itis only where a con- siderable number of concurring lacts are obtainable that satislavtory conclusions can be arrived at, 3 The common Zimbian term appears to preserve the same form of the unit, with the labial definitive asthe initial, although the latter has become a stbstantive part of the root. The most common form of the ultimate root is si, variable to ji, yi &c. and cor- responding with the Malagasy si ana Arabic hi, {[comp. the Eg., Galla and Malagasy def. si, zi) but forms in a and o also occur as in Semitic, Gongaand Malagasy. In some Janguoges s become t and in others r ort. Thelabial. prefixual element has generally the form mo. The same combination is tound in other formations, aud whether both the definitives are to be considered as being primarily a def, compound used as the numeral, or a merely servile function is to be ascribed to the postfix or prefix, must be doubtful in most cases, the relutive position of the principal and accessory definitives haviug va- riwd even in the same formation. If the original form of A were wa-lii, wa-kha &c., it would follow that all the Semitico-Libyan forms o! the sibilant, aspirate, dental and guttural definitive and unit might at one time prefix the lubial. Butin the archaic non-concreted con- dition of the glossary it is clear that each definitive had a seporate currency and was capable of bving used as the unit, The cotsbinations t Mr. Koelle’s Polyglotta Africane, received vince the text was writ- ten, enables me to make some addilions, He gives yaas the Woloff 1 oy Iu Bollom né-(evph min-, nim) iv prefixed lo the Simple no. merale. , ETHNOLOGY OF THE INDO-PACIFIC (SLANDs. Tl indioated special distinctions? sexual &c. The replacement of s, t, r or js found in the Arabic ordinal 1, takes place’ in the Zimbian cardi- nal 1, ‘This tends to the conclusion that wal, war, wat, wab, wak or pak, bas, mos &c. are all variations of one archaic term, whe- ther simple or compound. A similar range of variation occurs in those forms ot the Scythic unit which have the Jabial prefix. From the ge- neral structural analogies of Semitic in its most archaic stage and of Zimhian in its existing one, it is probable that in mo-si, mo-ri, as in the Semitic wa-hi, wa-li, the labial was primarily accessary. In tormit corresponds with the Egyptian ue of uo-t=wo-t [See App. 1, One 8}. The tollowing are examples of the Zimbian term,—mo-ja Suaheli, u- mii-e (contracted) Ki-Kamba, mo-dya Makua, yi-mo, du-mo Mud- jana, (probably contracted tornis with substantive prefixes, similar to those taken iu other Zimbian languages when the numerals are used ag qualitives), mo-yi, mo-ji Makonde, mo-si Takwani, po-si Masena, dolala, mo-esi Sechuana, mo-si Benguera, Kongo, Kambinda, mo- shi Angola, i-mo Mundjola, Yoruba, which in one dialect uses the Zimbian labial prefix with its numerals, has the same form of the unit in 9 (1 from 10), ma-i-sd. . Considerable variations are found{. The prefix alone is used in some lauguages, and in others the root changes to t, k, r, 1; o-bo Kuefi, bo Binin, mo Camancons &c.—; a-fo, a-fo-k ‘Timbuktu, wo-to, uo-to Bongo &e, va-ta Panwe, pa-ka Batanga ; [ba-t Euskarian.}§ In the Mpongwe, by the common change of stor, ritakes the place of si, ma-ri. ‘The Panwe va-ta 1, becomes va-la in 6 (5, + Although I Mave placed the Kuafi with the Binin and Camancon terms, Ido not infer that the latter were derived trom the former. They are probably contractions of one of the varieties found in Ni- geria, wo-r, wo-to &e. ‘The Kuafi itself must be considered as a reme nant of a once prevalent Nilotic term from which the Nigerian were derived. The full E Nilotic term is probably Vpermnd = in the Agau wal.ta, wol-ta6, and in the allied Nubian and Nigerian terms, cor- responding with the Semitic wal, In higher numbers the labialis common. It does not occur in the second term of the unit series, 3 (save as a posttix.) In Africa it was au archaic term for 6, (the 1 tale in the quinary system), under the forms pona, puna, fung, mon[5.G] Insome Nilotic and Nigerian t In Koelle’s Zimbian vocabularies the comtnon form of the second element isai, shi, #, sh, zi, z, hiion the western side. In the 8. E. both the i and @ forms occur, dshi, ts, tsa, za, ta, a, Inthe West, Musentans do has ko si, which may be an archaic variety preserving the root with the gattural in place of the labial prefix. Butit is probably a contrac. tion of the furm kiemosi (comp. the Mimboma bozi, ki-mozi), In the Jenwu group (Camerouna &c.) mo re occurs, but the labial generally ape peare alone, mo, imo, i wo. In-Baseke the sibilant ie replaced by the dental, i wote. A similar change of ihe postfix is found further south in Kabenda dso.e (comp. mosi, mos, koi), The Calabar dialects have similar forms dsi.dsi, dee t, edsi, In Basa and Kamoku the sibilant changes to ihe gspirate hi, hia, the form thus returning to the Arabic. . Koelle gives mo.ko. Undaza, eo.ko Murundo, mbo.g Ndob, fo.g Mfut, powg Ngoten, ¢ ko Ashanti, . - 2 ETHNOLOGY OF THB INDO-PACIFIC ISLANDS. terms for 6 it occurs as 1 (6; 1), under the form wal, wol, far, wor, fa, va, wake. In 9 it occurs for 10 in Tumali, Masena, Sofala and the Kongo group. As |0itisfound in Nubian, Berber, Shillah, Tibbo, and in several Nigerian languages In the dual series it is stil] mure common, The Semitic and Nubian wal, war ia the most jrevalent term lor 2 under thy forms-bar, mal, vali, vili &c. As 2 dual it occurs in4 in Seinioaconireyes 'o ar), in most of the Nilotic languages and Mala- asy, bahr, far, fur, ind, fut, tus &e. In 7 it represents 2(6, 2) in tishari in 8 it occurs (as4 dual) in Danakil and Shiho, bahr, bahara, and in Malagasy, valu. (e) Other varieties are found in the Galla ta-ko, ta-k, to-ko, kow ; in the Hausea group dai-ak, dai-a, de-ah, nai-a, da. Probably the Darfur d-ik is to be referred to this variety rather than to (d) In some of the Sudanian aud Nigerian systems the dental and guttural oceur separately or combined, and in some cases reverse the Galla or~ der. Ga-li Kallabi, ki-de Begharmi (comp Bish. gur, gir). Mendi, Pesa and Kosa e-ta, toh, i-ta, [Gonga forms] Kra ku, Fulah go, Fan- ti e-ku, mi-e-ku, Akin di-a-kun, Amina a-kun, Tambo ku-ki, Moko kia, Karaba ke-t, Karapay e-di, Yoruba o-k’ka, o-ko [Galla to-ko}, The nasal occurs alone in the Yoruba ine [Shiho in-ek] and Ibo na [adjacent Hausa group nai-a, Reetnal: the Hottentot kui, koi-se, ui, resembles the Gonga e-koe. Another Hottentot term itswi [=i-tsui, i-tsoi] also resembles the Gonga sibi« Jant forms, The double vowels of several of these forms, ai, ui &c. are ancient Libyan (comp. the Eg. def. and units). ({). Inthe Agau group lo, la, appears as the goot, lo-2a Waag Agau, la-ghu Agaumider, la-gha Falasha ‘the Bornni Ia-ska, la-ka, and the Malagasy i-rai-kia, re-k, are similer terms to the Agau la-gha, In 6 (5, 1) the r becomes 1 (ene, en &e.) In the Dankali-Kuafi 6 the nnit has « similar form (leh &c.) In Malagasy it is also preserved as a definitive, corresponding with the Semitic and Atrican le, re, la, na, al &e Inthe Mpongwe ma-ri it \akes the place of the common Zimbian si, The Zuluand Kosah nye, 1, is probably the same root, It occurs in the Mpongwe iva of ina-gomi9, In fe? 10) the form la-ka &c, occurs in Bornui, Sangsadi and Mazambiki (g). The Bornui tirlu has the same form [Agau lo) with the dental refix, a reversal of the ordinary collocation, similar to ki-de, ga-di, e-t The same collocation, with the guttural in place of the dental, is found in the Bishari gu-r of suggoor 6 (5, 1) and the Mandingo ki- S| Several liqoid forms are given by Koelle, e. g, wo.no Okam, ke bos ne Nki. inga Kambali, un¥i Yasgua, kon Akorakura, wi an Wololl, fa. nod Fulup, a nod Filbam, bane, vane Gadsaga, pu.lolo Bola, pu-laian Bereres, ov lon, » lon Prpel, pa ini Padeade, pe le Kisi, ke le, ke-len, kes ren, ke den, ida, ira Mandingo Group, pin, Baga, Timne, bu! Bulom, Mamna, do Grebo Group, de Dahomey Gr., e ni, @ ne, e |i, € ne, me-ne, i nye Yoruba Gr, ka lo Kaem, ka ni, ka-n (deo, we ni, ne ni, e.nyi gha-ni, ma.ni, gma nvi onvi, &c Nufi Gr, pa-le Mandara, bala Ebe (Nu. 6 Gr.) The liquid appears in these examples with wariable prefixes wo—, bo—, bu—, ba—, pa—, we—, me—, &e; ka—, ke—; 1—, e—Ac, like the sibilant and denta! root, They afford s'rong evidence tbat the ri, li, ni, i, no, lo, &c. are bn! variations of the same root. . . ‘The form of the prefix serves to some extent Lo trace lines of special con. ETHNOLOGY OF THE {NDO-PACIFIC ISLANDS, . 13 Milchanang to the Singpho and Lau, could not have been exerted by « Tibetan tribe which waa confined to themountains, like the present. western Bhotians and eastern Lhopas. Tibetans or Tibetanised Himala- yans most have descended into Ultraindia or India and acquired a poli- tical and social predominance over a considerable area, The second inference is that the diffusion of Tibetan elements on this side of the Himalayas has not. been caused by a single movemect of a Tibetan tribe confined to one era. - These Bhotian irruptions into the sub Himalayas and India which may be regarded as historical bave pro- diced a comparatively smallinfluence, Although it has been continued. until the present time, or for about twelve centuries at least, owing to the permanent advance of the Bhotian ethnic frontier iota the sub. . Himalayas, it has failed to assimilate the conterminous Janguages of that tract. From the Tibarkhad to the Abor a nearly oninterropied band of - Jangnages is preserved, which retain oon-Bhotian forme of pronouns and particles, and two thirds of the vacables of which appear to be non Bho» tian Even the most eastern of these languages, as the Daphla and Abor, which are spoken by bighly Bhotoid tribes, have a very considerable ba. sis of noo-Bhotian traits in phenology, glossary aod ideology. In the Gangetic plain the influence of Bhotian hae been still less. It ie obvious, from these facts, that the pure Bhotian tribes and jangoages of the sub. Himalayas have always been separated from those of the plaia by a barrier of only partially Bhotised langaages. In Bhatan the influence of the historical Bhotian advance tothe southward hae been more powe erful and extensive than in Sikim and Nipal, bat the physical aod Jin- guistie character of the Bodo and Dbimal shew that beyond the moun« tains it was comparatively feeble and superficial. In the basin of the Irawadi and the connected western territory as far as Bengal and the Bay,—the tribes of which are entirely separated from the Bhotians by intervening ones and are less Bhotian in person and customs than some of the Himalayan tribes,—we find that the Naga and Yuma vocabolaries gre twice as Bhotian as most of the Nipalese. The Mishmi, Jili, Sing pho, Naga, Yuma and Garo appear to have a direct glossarical connee- tion with Bhotian—whatever moy be the chronological and ethnic relation of the Tibetan movement which induced it—distinct from that which Tibetised the more western languages, Bot to ascertain this relation satisfactorily it is necessary to advert to the non-Bhotian traits of the Naga, Yama and of the Gangetic languages, and these will be discussed in & separate Section. » In the following detaila my principal object will be to shew the extent to which Tibetan enters glossarially into the languages of Ultraindia and India. In the comparative lists in the Appendix, compiled befora the publication of Mr. Hodgson’a Sifan vocabularies, I had indiseri- minately entered all words that have Bhotian affinities, A few are thue included of which the derivation from Tibet way be doubted, Mid. Asian and ather remote languages having forme that are nearer the cis Himalayan, while others are given which now appear to be Sifan and not Bhotian. But making avery allowance for these, this Section with the Appendix will afford a general view of the inflaence which the Sifang and Bhotians have exerted on the vocabularies of Ultraindia and India from the era when they firat found their way across the soowy barrier. od 14 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 8. Pronouns. The Bhotian pronoun ofthe lst person, nga, na, prevails in mést of the Himalayan lavguages and in Ultraindia, bnt as it is not common in the Naga ialects, it is improbable that the Ultraindian nga is of im- mediate Khotian origin. { should rather have supposed that in some of the eastern Himalayan dialects it was of Ultraindian and not of Bho- tian derivation. The influence of the Gangetic Fhotian of the Pal -ra on the Burman fainily was far too slight to lave extirpated the ancient Burman pronoun of the Ist person throughout all the Burman dialects, The Burman pronoun agrees with the Bhotian and Chinese (ngo) be- cause all have ultimately derived it trom the same mother formation. The origin of the Ultraindo-Gangetic nga, na &e has now been set at rest by the Sitan vocabularies, et The Bhotian 2nd pronoun, khyod, kiiyo, Akhe, has been carried by the Bhotians into most of the Himalayan languages, but not beyond them. ‘This pronoun is of itself almost decisive ag to the relation of the Ultraindian to the Whotian, and of each ,o the Gangetic languages. The Burman nang, na is found in several of the Naga languages, in Garo, Bode, Dhimal, Abor, Miri (ne, nan), Daphla (ne), and even in Magar (nany), one of muny prool’s of the connection between the pre- Bhotien Himalayans aud the Burman family, The Mon and Kambo- j#n pronouns, the former of which have spread into the Malay penin- sula, are distinct. : The Bhotian rif pronoun, kho, kiu, is found in several of the Hima- Jayan vocabularies, but not in Dhimal, Bodo, Garo, Naga or Burman, the prevalent forms being Sifan and Draviroid. The Singphu Adi is pertem Khotian, but as a similar particle is a common definitive (¢. g- asia ka singular, & plural ) this is uncertain. : The Bhotian plural particles nam, chag, dag do not appear to have made much progress in India, unless the Bengali dig isa derivative from the last,* Tha postfixed definitives po, mo &c. are found in several of the Himalayan and Gangetic languages but not in the Ultraimdian. Most of the Himalayan pussessives are apparently modifications of the Bhotian, but some may be Dravirian, the latter having a wide range of possessive particles. The extent to which Bhotian forms and parti- cles have been engrafted on the Gangetic languages appears sufficiently from chap, IV., so thut it is needless to pursue tlie subject bere, _ Prom the evidence ofthe pronouns it may be inlerred that the Rho- tian dialect intruded on achain of Gangeto-Ultraigdian dialects which possessed the Sifun forms of the Chinese. Saye an the vicinity of the southern Bhotiow dialects, the Bhotinn pronouns have made little pro- gress. Kventho highly hotian Takpa retains the >ifun-Ujtraindian 2nd prorioun and has not borrowed the Fhotian one trom Lhopa, Changlo, like Hodo, Dhimal, Abor and all the proper Ultraindian dia- Jects, has the Sifan pronoun, and in the sub-Himalayan band the Bho- tian spears not tobe fonnd to the eastward of Nipal, ‘hat the Silan branch preceded the Bhotian even there and further to the westward, appears from Tibarkhad preserving the Sifan pronoun, A full comparative list of the Silan and of the allied U)traindo-Gan- getic pronouns hat already bein given in chap," V. sec. 11. Among « Magar ku-rik, (See chap. V, sec. 11). . 7. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 15 the correspondences there noted will be found the contracted form a of the Ist pron.common to Manyak, Angami Naga and Mikir; ang, the Gyaruug postfixual form, found as u. postfix in Naga and as a separate form in Bodo, Garo and Kiranti; nge Takpa and Singto; ka Thocha, Dhimal, Lepcha, Lau (kha, kau, ku), Toung lhoo, &c. Most of the Ultraindo-Gangetie forms of the 2d pron. are Sitan, nan, na The Manyak variation of the yowel too is found in Daphia, Abor, Deoria Chutia, Angumi, Mozome Angamiand Namsan- gya. The other Sifan particles are also Gangeto-Ultraindian, Ladd a tew examples, ‘ : The Gyarung particles occur in Ultraindian languages. Ma, m &c. is common as 4 negative and caritive postfix or prefix (Abor, Dophla, Dhi- mal, Mikir, Garo, Burman &c.). The Khamti ma-,mo-, and Chinese pre- posed m &c, is the same particle. Da denotes the present in Abor as in Gyarung, and a Gyaruny-like combivation of it with 4a, another form . of the same Bhotian definitive, rendersit emphatic, lada. In another form, ta, it is completive, corresponding with the Gyarung /a, past. In the Dophla pertect pana a combination occurs similar to the Gyarung, pa by itseli being tuture in Abor (in Dophla 60). Changlo has -le pre- gent, -ba past (Wurm, byi, Bodo bai), -dong future (Burm tho, Kham- ti fa—), Bodo has -dang present (useil as a verb subs.), dai, nat, imperlect, dang-man pertect, nise, gan, tut. Dhimal has / past, hi, mhi, nhi present, Garo has -na, -enga present, enya-chim imperf:, =a, -aa pert., -chim perf. def, esa, kieng, fut (Bodo). Naga has -t perfect, da-prefixed, with -¢ postfixed, as a second pertect (Abor, Gyarung, so in, Tibetan b-&c. pref, with-s postt.), ¢is future, Mikir has -loh past (Naga, Kas.), -ye future (é Se -bo, -bang empha- tic futures (Abor, Daphla, so Burm mi), -si participial (Gyarung). Garo — has -na present (Gyarung na-). Kasia hus la-past (Mikir, Naga, &c.), n-luture. In Singpho -Aa is past (Dhim.- Ai, Bhot.-s also ha-, h- -ha-dai perfect, -a future (Blot -a). Takpa beiog at present the only known language that appearsto be conterminoos with the proper Tibetan dialects on the one side and with the Ultraindo-Gangetic on the other, it will be useful to advert to the affinities of its pronouns and particles. I add the oumerals to give great. er breadth to the comparison, As Takph is the langoage of the Towang raj, it must be conterminous with some of the dialects of the Bor and Abor tribes.* At present we are only partially acquainted with those of the southern Aka, Daphla and Abor- Miri. ‘Tbe two first appear to be the same and to be closely akin to the Jast. How far they are spoken to the north, and whether any other dias Jects exist between them and the 'Takpa, is not known, i ee *. Arethe Tog ad or T'ag-ad, one of the tribes of Bhutan who in. habit the district of Tug-na, or l'ag na, Tagana or Doka, not a section o£ the Tak-pa who happen to be uniler the dominion of the Deb Raja? (As. Res. XV, 146, 140 Pemberton 111). The Pilo of Tag-na’s territory lies between Bakshaaod Cherang. He has two Dware or passes, aod the Refu Jado and two Tomas are under his ordera. His territory is — eight days journey long and four days from east to west. He pays al. together annually in two instalmenss about 3000 rupeea and rules about 8-16the of the country” (Aa, R, XV, 139): 16 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. The prononns of all these dialects are Sifen-Ultratadian. The Ist is nge, nye, in Takpa, the vowel being exceptional. The East Mishmi ke appears to be a variationofit. Daphlaand Abor have ngo, the Chinese form,—the Sifan and Bhotian being nga. Mishmi has ha, the Manyak, Naga and Mikir a. The 2d has the Chinese form also fonnd in Horpa (ni), but coniracted toi. Daphta and Abor have no—the Manyak form—and it is also posseased slightly modified by the Western Mishmi, nyo. The 3d in Takpa ts pe, be, whivh is not Sifan but Chinese and eastern Mishmi, we. The lebial ia also Daphla ma, and Abor bo. The Weatesrn Mishmi com- bines it with the dental mta, The Takpa‘pl. pos'fix i¢-ra, the Daphla lu (a var of the same particle), the Abor-/u or-du-ke (comp. Horpa 7't- gi); a-rong ‘8 wnother Abor form, and the Mishmi long corresponds with it, F. Mishmi has thal (comp Singpho theng, Angami toleli, Tengea hala) &u, The Takpa pose. is -kw, the Daphla and Abor-g. The Takpa da. live is sga, ia, the Manyak we Duaphla has -bo, Abor -nd-pe or-ke-pe following the poss. (-g-ke-ne), bo and pe being the same asthe Manyak we, and na the Takpa /a, Horpa da, Bhotian na, la, da,ra. The Tukpa ablative “from”, is & which may bea contraction of the Manyak ni, Chi. nese fi, Bhotian ne, di-ne Daphla has -g-ga-m, Abor -g-ke-m, or -g-lo- ke-m (-g poss. -lo locative), in which ga. ke are Horpa, gha, Thocho, ge, k. Changlo gai, Borman ga, Garo -mi-kho = The instrumental is in Takpa and Gyarung gi, Bhotian gi-s, Abor -ko-ki, in Daphla -mo-na (following the poss -g-', comp. Sunwar mi. Limbo nu, Lepche nan, Burman nkena, ~ The nomerala present some coinridenves. The ] of Abor a-ko may be the Manyak ta-biand Thorho e-ré, (not Takpa thi, or Gyarung ka-thi), bot as a- is a prefix and -ko 4 postfix in the other nomerals, the root appears to have been Jost in 1, althongh it ia preserved in 6 a keng ko, and in the adjacent Changlo dialect of Lhopa, khung. Daphla retains it in laken, 2, Daphla a-ni, Abor a ni ko. is Bhotian, Naga, Himalayan; Tekpa has nai «33, VD. a am, A. @um ko, Takpa som, Gyar sam, 4, D. ap li, A.@apike,(Changlo phi), T. pli. 56, D ango, A, a ngo-ko, ue ngo, pi li—ngo-ko (4 repeated), T li-a nge (4 repeated). 6 D. a-k-p- le (a-k= a-ken |, p-le properly 4, bot here contracted from the original full form of 5 p-li-a-nge), A.a-keng-ko (the 1 of D.), Changlo khung 7 #+%D. kaena-g A ki-nit-ko, ku-nid-e (Borm.” khu-nhit in 2 nhaik, nhach—nag of D) T nia (2). 8, D plag-nag (4. 2), A. pi-ni-ke (4, 2), 9, D. knyo, Lepeha kyot (nearer tothe Chinese kin, kieu, kau than the common Tibeto-U\tr. gn, ko &., A, ko-nang-ko (? Gyar, kung-gu), 10, D. rang (Karen lang 1), A uying-ko. {t may be inferred that Takpa hae a close special connection with the Daphla and Abor, but that the period of their separation was very ancient. The aouthern dialects retain some archaic full forme not now found in Takpa or the Sifan languages, and they have non-Takpa traits in common with Bhotian, and with Sifan and Uitraindian dialect, 4. Numerals. 1. The Bhotian gchig ie the original of the Mormi ghrik, of which the Guorang kri isa contraction. The Bhotian ep. chik is found in Ser. pe. Lhopa has chi, and Newarebhi. The Limbo thit preserves the Chinese final, and the Takpa thi is a contraction of a oe form. The Naga ka-tang, (ka-t in higher numbers), ka-tu, a-vA-et (? a-khet) BTHNOLOGY OF THE INDO-PACIFIC IsLANDA, lf t anil itis preserved @the Kissi mia. Other Nigerian forms are e-pa Akongo, ba Camancons, é-fi Nerapay, e-ha, u-ba Kuraba, ma Cal- bra, i-ba Moko, Kaylee, ba-ba Bougo, be-ba Batanga, (be-lala 3, be= nai 4), fe-be Panwe, be Biniu, a-bo Ibu, a-ue Whidab, a-uwi Papah, e-mo Akripon. The prefivuul de- of the Pongo torms is repeated in some higher numbers Ina few o/ the more northern languages it ia used it ulso, Ip 2itis comlincd with a nasal root, Ashanti e-di-en, a-bi-un, mi-enu, (1 mi-ch-a, biak-ung &c.) Amongst the purely labial forma common to Hausa and the lower Niserian longanges, the full or compound form baris found in Nufi oju-inir 2 (in 7 it becomes bi, ogu-dwu-bi). Sbabbe hous-war-ba 2, hoo-ah-wera-har &e, 7, and Binion a-boar 2. These forms suggest that _ the Wololfand Nabian ar &c, of ni-ar, vr-o Ke, is siinply a contraction ot bar &c. This is supported by the Tumali ar-ko 1, mar-4a 7, Kolda- gi fell-ad 7, Dalla bar-de 7, (belle 2), shangalla am-bund 2. The Koliagi fell of 7 and the Dalla belle 3 are Zimbian and Mandingian Gees (bili, lela &c,), corresponding with the Shiho mel-hen, Danukil me|-nene. (4.) With the contracted Tumali ar-4o, Koldagi ora, Kensy ow-um, Naba oug-gha, Darfur ou, we must place,.on the one side, the Kuafi ari and, on the other, the Woloff xi-ar and Tembano-ali, In these Nige- rian terms the nasal ia a prefix ds in ibe VW oloff ni-at 3, ni-ane 4; Teutl- bu nodose 3. . \ (c.) In the Galla family a labial follows the liquid instead of prece- ding it asin Zinbian. Lit Bishariadobial both precedes and tolloway the liquid, Bishari melab, Daukeli lame, dume, ¢ iumali lebe, Shiho, Galla lama, Woloitsa nama, Wolaitsa, Woratte lala, Agan langa, linga (in 7 also la—ma), Malogasy rua or ru-wa [in Asonesin ru-a, du-a or du-wa, In-wa, nu—wa, du- ot The Begharmi sub, szab may, be of W. Nilotie origin through dum, lub [=dul)| &e. The anomalous Yan~ hep may be a modification of a térm having a similar ovigin ee leb |. The Galla—Malagasy terms lub, lume, dume, rua &e. resemble the In- do—European d-wa, d—va, d-no &¢e,, Lazian ku-wa, Uerian k-ay-to, “Turkish d-wa-t7 4 (i. e. 2 dual), and also those varieties of 2 and 4 which have a final r, Seythic, Caucasian, Indo-Duropean, The collocation of the * two elements and the yowel u distinguish the Galla—Malagasy from the Zimian forms, and might appearto connect them with the Indo—Hurope- an. But as the Bishari malub appears to be the full original form, these terms admit of a somewhat different analysis, and one that reduces them to an archaic variety of the common Africun numeral. In the archaic Semito—Libyan formation the Jabial was a detivitive as in Caucasian and Scythic, and it was alsoa numeral element, (See the remarks on the Semitic nunierals). In Semito—Libyan as in Caucasian it was used pre- fixually as well as postfixually, and Seythic, like Draviro-Australian, had archaic defimitives and units in which the labial was either initjal or pre - fixual. In Bishari, as in some other Setnito—Libyan languages, the labial is still common as a prefix, and its ancient prevalence in the Nilotie pro- vince is attested by its use as a postfix in Dankali and Shanyalla, ( corres» onding with the Hottentot usage), asa prefix and postfix in Berber &e. Pn Bishari it has the vowel o (wo, minke p and in Dankali, Shilo and Are kiko it sometimes takes the forms um as a postfix, Danakil an~ 13 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. ub “milk”, Galla an-an; gell-vb “the body”, Agaf@i-kel, Tire a—kal- at, Gatat a—Kal-at-an ; kull-wm “fish”, Adaiel kull-um, Hurrar tul—wm Arkiko mud-j' “sheep”, murroo; Adaiel kok-wb “ stars’; gel—wb “leit”, ra Shiko; ker-wb “near”, her-ud Shiho ; ruk—a) “carmel ’, Shiho ra-ku- ) Danakil, raki-ba Adaiel in-ob “teeth” Arkiko, il Saumali, iru Falasha; arrwb, “tongue” Saumali, ar—wda Galla, arr—at Hurrur. The Bishari 2 is probably therefore an archaic form of the numeral as it ex- isted in the Gul’a family, and the correct analysis would appear to be mal- uD. "This: is confirmed by sera-mia-) 7 (5, 2) in which the Inbial. keeps its place as the root, and by the 7 of Shiho, me:—hen, and of Dankali, mel- nene,in which the root has its full form, and which I have ulready idedtitied with the common African bar, bel &c. 2.'TheShiho and Danka 8 preserves the same root for 2, bahr, bahara, and itis also found i Mulagasy, val-u, The Shangalla metama 1 is a sitnilar example of the labial oveurring both as an intial and tinal, The Galla 7 also retains the labial posttix -de. The - Bishari mal may therefore be considered as simply a variution of the Ni- lotic form of the common Africin 2, bar &e. The terms under (¢) would thus appear to be mal—uh, lu-me, du-vm@, le-da, lama, na—mea, la-ha, la~ nga, li-riga, in which the root vowel sometimes takes the slender form a3 in the Dalla bee, Zimbian bi-li &c. The Malagasy term, in like man- ner, becomes rua, which approximates to the archaic Semitic form of the _ e term lost in 2 but preserved in 4 (2 dual) arta &e. The Tumali ar—wm 4 preserves o form of the original postfix similar to the Galla -wd, -ume. Amongst the current Semitie terms for 2 Hebrew retains a labial pestfix. In the historical condition of the Semnitic langnages the nume- rals, save 1, arg substantive not qualitive. They have cifferent forms for masculine and feminine, the fem. being, however, used for masculine words, and the other form, without the fem. final, for fem. words. As the numeral postfixes must have been originally qualitive, it is probable that in the era of the Semito—Libyan formation when the sexual variations, of the definitive were in full use, the unit took all the detinitive posttixea, and the higher numbers all those which had o plural npplieation. The Jabial was masculine and plural, although originally singular or indefinite; the liquid 1, n, r, d was also plural; i fem. and plural; u plorval &e. In, the most archaic period the unit probably itself varied with the sex, | The dental and sibilant def. may have been the fem. uot, and the labial ’ the mase., and in the higher numbers, which were but compounds of units, the distinction may have been maititained, Those terms in which the la- bial occurs as a root or-postiix were prubably the original masculine nus merals, ‘Their obscuration, concretion aud Jess is a phenomenon similar to the decay and loss of the masculine definitive in the. lossiries generally, save in Hottentot, and the persistence of the, eminine. It seems possible to expliin in this mode the original currency. of two terms for 2, ba-risba, ba-1-ba, m—al-ub, bi-ti-da &e. mase, and s-en-te, ith—-na—ni, o-il-el &e. fem, ‘The yowel may also have even in, these archaic form had a sexunl power bar, ba-tu &c. being a cominon and ba-ria fem. fivm. The mase. labial root they have been the principal - term when the Semitie system was first ¢erried to Africa, although the’ fem. root afterwards became the more important in the Seuiitic languages, : It is clear, on a comparison of all the Semitic and African terms, that. the ultimate archaic root is the liquid—na, he, in, nau, wu, roh, n, r Semi- te-Egyptian, li, 1i, le, ls, r, 1, de, te, dsl, 80 de. Airican—; that it wae. STHNOLOGY OF THE INO-PACIFIC ISLANDS. 19 early conjoined with tha@labial and with the sibilant (dent., ent. &e.) def. pre Bes postiixually, or both; that the form with the labial pretix bee came one of the most prevalent in Africa, the pretix early concreting with the root,—a secondary prefix, labial &c. | tog assumed 1 some lanpuayes, and the root itself being thrown off in several, With the aid of Koelle’s Voc.a list of the principal variations of the See mito-Afvican numeral may be given. . ‘The root has the variations na, ni, ne, nu; la, li, le, lu; ri, re, roh, ru; di, de, ndi, du; ta, ti, t; su, 80, si, dsi, se, dse; ka, gu, ‘Che uetual forms: are: LY a 1st, the pure foot, with the servile particles thrown off’ a3 ndi Bornul, le Gurma, so Grebo group, 2d, the root with serviles, as e-dsi Aka-Igala gr., -le, le-a Kasm ke é-le Kambali, ya-r Woloif, a-ro Nubian, rug Malugasy. 3d, the, plicated root, di-di Fulah, , 4th, the root with the Isbial pretix, bi-ni, Di-ri, va-ti, vi-di Zimbian, de- Je Dalla, fi-la Mandingo, me~dsi, me-lo Aka-igala gr; fhe same form with the labial postfix ma-lu-b (or meal) Bishan, brene-b Bute; with the guttural postf., ma-r-a Koldasi (7), with the dental postf. fe-l-ud Koldagi (7); wjth a secondary lalnal pretix, bi-wa-de Songo; with a se« eondary dental, sibilant, guttural or liquid pref, s/-oi-i Muntu, tem- be-re Nyamban, é-bi-en Ashanti, g-ba-11 Dsarawa; with the liquid final, ~ ma-ra-n Lando i-ri=- Mandingo, pe-ra-n Timani, with the rovt elided, be, i-be, a-fa, e-lh, mba, po, pi-pa, be-ba, be-fe, e-we, e-ve &e. very come mon in the N. W. Zimbiau pron. (iswwu or Cameroons-Gahun, the Calabar and Lower Nigerian groups, thence Inland over the Chadds basin to Sue dania including Hause, and westward in the Dahpm y group). —_» 5th, the root with the sibilant, dental or puttural pretix, sie-ne Hebrew, s-roh Mahrah, s-nau Re., sien Shillah, di-s-n Kandi, ki-le Arkike, ke-le Musantandu, zo-le Basuide, ka-t Angoli, di-el Gara; the sume formwith postfixes, tth-na-vt, ath-in-t-an Ar. she-ne-im Webrew, the-na-t Berbd., senou-s Eg., ke-ta-), ngi-ta-ba, fe-ta-w Bola gr, t-ri-n Mampa. 6th, the root with a labial postf., lu~me, du-me, le-ba, la-na, nama, Galla group Xe., su-) Beyharini; with the liquid postf. la-nga, li-aga Agou, In the Pilham kv-ga-avw the root is prutturaliseds The serviles connect the form with the adjacent Balla fe-ta-w. Uhe Pulup fu-ga-y-ten is the Pilham ga-w@ with ao superadded postfix and with the labisl ia place of- the guttural prefix. In the varittion fu-ten the root is ejected. In the Limba toe, kae the vurintion of the root consd- nant from the dental to the guttiral also takes place. The Bidsago mund-su-ive, i-so-be, Bulinda g-si-b-r are allied forms, and tie so of the Grebo group is connected with them. ‘The Bawhermi su-J, szas 6, sa-p is a Mid-African link between this detached aad peculiar Sene- gambian group and the Bishuri-Galla, with its lu-d, du-me &e. The line ° of diffusion thus indicated must have preefded the advance of the Zim- bian forms from the south into the Nigerian, Chadda and Sudanian pro- VIUCES~ - A. See Semitic. “ ak, B, The Draviro-Australian bar, bari, bula &e. are anologous terms, » In the other Asiutic systems the labial initial is not prevalent. The se- ~. coud and radical element li, le, mi, re, ui &e, is very common. With the — Foreign Affinities. * = ‘= ——e a 20 ETHNOLOGY OF THE INDO.PACIFIC: ISLANDS. sibilant definitive and unit prefixed, it occurs with the same slender vowel in the Semitic zs-in-, she-ne &c. (the in, ne generally changing to il, li, le, 1, ri &e, in African languages), in the Caucasian shi-ri, o-ri, Samoiede si-ri, si-de, Chinese il, ni &e. The’ Kongo terms favour the iaference that the African liquid element is the same us the. Scytho-Semitie and Can- casian. They even tend to show that the term was received from a Can- casinn language. But it is more probable that in the original Asiatic sys- tern, is in the Zimbian, the lahial detinitive and unit was preserved as the initial in sume varieties of 20s well asthe sibilant, guttural &e. If so, the African bili &c., like the Draviro-Anstralian bar, 5 bula &c., is a remnant of a form onee prevalent in Asia. ‘The Caucasian wi-ba@ (Abkhasian ), an Euskarian bi, are similar remnants, the former having the labial postfix usin the Malagasyterm. In many Scythie terms and in the Indo-European the labial element keeps its ground with more or less prominence and tenacity. But in thes® terma the labial has another definitive unit (dental, guttural &e.) prefixed a Japanese preserves the labial in 2 as in 1, 3. and 4(fi-to 1, u-ta 2, mi 3, mu 4). The closest forms to the Zimbian are preserved in the Seythie and. N. Vv. Asian 1, 5, 10 &e. bir, pir &e. 1 (Turkish ), mill &c. 5,10, Koriak, mer 10 Tongusign. The only N. E.derm that pre- serves this form in 2, and thus corresponds with the Draviro-Australian . and Zimbian, isthe Namallo( Esquimaux )mal, the broad yowel being Dra- viro-Anstralian and Nilo-Nigeriun but not Zimbian, «Hence it may be inferred that its dissemination as a term for 2 was very archaic. The Kongo-Anolan, so-li, ko-le, ya-ri Closely resemble the Semitic form, and still mgge closely the Georgian o-ri, shi ri, ye-rn, while all have Scythie alinities. They strangly*support the opinion that the liquid is the essen- tial element in the term, and that the contracted forms ni, li, ri, di, ti, si &e. area return to the ulthmate roct. The Zimbian aud other similar African terms appear to be- related to the current Semitic not directly, but. through their mutual derivation from an archaic Semitic or Semite-Li- byan mother-system, analogous to the Caucasian and Seythic, , The broad and widely spread Nilo-Nigerinn form, mal, bar, ba, ar &c. is mererly a variation of the slender iri &c., and itis probable, fron the Kuafi and Tembu terms, that the- full form was ba-ri, the liquid hav- ing the slender Asiutic-vowel as in Zimbian and the vowel of the labial cor- responding with thot of therWonge- e ‘form of Zim wa-li, ba- ni. But whether the broad form of the labial was a distinct Asiatic im- portation and directly connected with the archaic Namollo ma-l, Drayiro. Australian ba-ri, ba-r, or was merely an African modification, is not clear, In Asia the browd form appears to - the more archaic, the vowel in the élender forms having assimilated to that of the conjoined definitive in si, hi, ri, ti, ki &e. Itis probable that in Africa also the form bi &c, waa secondary and ba primary; in which case the eastern ba-ni, va-l would be the original Zimbian form; but the Caucaso-Euskarian bi suggests that bi may also have been imported from Asia. In the Semitic terms for 2 the sammie assimilation of the vowels of the two elements is found. In the Se- mitic 4 the broad vowel of bar, mar is preserved in the contracted form ar, while the tem. form of the labial 1 wali is identical with what appears to have been the older African form. | Obe. on the Distribution af the terms. Ast. The broad forms Va-ri, (ma-li) ma-lu, du-ma, ru-a, sa-D, su-d, ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Al so-ma Sc, ayypear to have been the first that were very tia 4 dissemi~- nated. They are the most prevalent in the Nilo-M and Nigerian provinces, and in the N. W. division of the Zimbian. e broad form was probably received from the Semitic province before it was replaced there by the present Semitic term. The slender variety bi-ri was probably diffused ata later period by an influential Zimbian dialect, as- it is the most prevalent in the east and south divisions of Zimbian. But the Dan- kali, Saumali, Dalla and Koldagi forms appear to show that it did not originate in the purely Zimbian province. In the west the later forms, and contractions of them, have spread northward, displacing the older varieties in most of the Nigerian groups. 2d. The*historical Semitic terms are saa comparatively recent im Africa and have made little- progress. The Abysinnian, Egyptian and Berber mark the oldest diffusion of the Semitic iorms. Three. » The African terms for 3 are remarkable for their adherence to one ultimate root, and for that root being the same as the Semitic. A. sho-meu-t, sha-me-¢, sho-m-t, sho-m;¢i, sho-m-te, Egyptian. eee see chy The Bishari mih, in 8 su-mhai (Semitic, Eg.), preserves ée labial. B. (a) tha-lé-th, sa-la-tha, sa-la-sa fem. tha-la-tha-ta . mase. Arabic, sha-lo-sh fem. she-l6-shah mase. Heb., si-la-sa (30) Babylonian, te-ln-ta mase. te-la-t fem, “haldee, se-le-ste Tigre. (6) ke-ra-d Berber, k-ra-t Shillah ka-rai-d Kandin, (See Semitic]. (6) is a variation of (a). . (¢) tha-th-i¢ Mabrah, tha-k-i¢ Gara, shi-sh-di, shi-sh-ef (sa-sa 30, si-ssa 60) Harragi, so-s-t Ambaric, to-s-k Nubian, (sa-sa 30 Gafat, ga-as0, sha-sha 80 Gonga,) The Mabrah, Gara and Harragi forms cor- Sag with 6 (3 dual) not only in these languages, byt in Arabic and Hebrew (which drop the la, lo of 3), The Babylonian 3 is not ascertain- ed, bat in 30it has the Arabic form. 1n60 however a more archaic form is preserved, su-su [Fleb, shish-shi 6, shish-shim 60], or su-si which corresponds with the Amharic so-s of 3, ( Harragi, Gronga ead Gafat 30), and indicates the former existeice ofa similar term in the Himyaritic province, The cefinitives and 3d pronouns in su, so, ha, ho, tu, to ke, reuder it probable that this was the oldest form of the Semito-Libyan unit, but those in a, i, e may have coexisted with it from an ancient period, Both are found in gp rican terms for 3 and other numerals. ee Semitic The Himyaritic double dental or sibilant, or dental followed by the guttural, is similar to the most widely prevalent African terma. Galla. family. su-ddeo Dannkil, su-d’de Saumali, se-dde, sa-di, za-di Gulla, a-dda Shiho, se-tte Dalla, From the Galla and Dalla forms and the absence of the intixed la, lo, these terms appear to be of Himyaritic origin. Zimbian family, ta-tu Suahelj, ¢-ta-tu Ki-Kambe, ha-hu Ki-Nika, ¢- ta-tu Makonde, ga-ta-tu Mudjana, ta-tu Masena, Sofala, Kosah, ma-tha- tu Zulu, ta-tu Benguela, Angola, Kongo, Kambinda, sa-tu Bondo, di-te- du Mundjola, ta-tie, tu-to, Fulah, nda-ta Tumali. These forms are allied to the Himyaritic tha-th. A: second variety suggests that both, although cognate with the historical Semitic, have an a common source—ma-~ ra-ru Makua, vi-ra-ru Takwani, tha-ra Mazambiki, ¢tri-ra-ru Delagoa ? 22 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Bay, ti-nha-ro, gi-na-ro Nyambana, ta-ru, tha-ro, da-ra-ro Sechuana, ?’-no-rra Hottentot, a-ru-se Hot., be-la-la Panwe, ba-la-li me-le-la Camaneons, Ji-ra Aku new la Binin, ra Bullom, ntsha-ra Rungo, te-re Calbra; ti-lu, ta-lu, te-lu Malagasy. The chief peculiarity of the Zimbian terms is the final vowel u. ‘In ta-tu it does not correspond with any Se- rnitic form, but as this form without the l, r is simply the double definitive and unit, it may be referred to an archaic form of it similar to the Mahrah and Gara ta-ut 1 and the Babylonian su-su 60, The u of ta-ru, tha-ro, ta-lu, te-lu &c. corresponds with the Hebrew lo in she-lo-sh, as the initial definitive and unit does with the Arabic tha-la-th. The same form of the li- ® quid element is found in the Mahrah and Gara roh of s-roh 2, and in the Ma- lapasy wa-lu 8 and fu-lu 10. [tis probable, from the variation of th, s, t, to r, l, occurring both in the Semitic and Zimbian terms, that the latter was not the plural def. n, &c. occuring in 2, but merely a phonetic modification of the former, as in the Panwe va-ta 1, which becomes va-la in 6 (6,1). _ SBu-su, tha-th, ra-ro, la-la &c. are forms aeunely being pis of the ori- - ginal term having been a reduplicated unit, for they are found in widely separated branches of the system. Such forms as tha-k, tha-ra, tha-la, ta-lu, na-ro, k-ra, &c. appear to be only variations of the original term. imilar variations occur iu the allied N. and E. Asian terms [See Semitic Numerals, 3 B.] The Gonga; Agau and Nubian terms appear to belong toa later era,—that of the extension of aay balm: to Abysinnia, The Nubian to-s-k, to-s-ho-ga (double postf. ), tow-s-/o, to-dje, are Himyaritic or Babylonian through Abysinman (Ambharic so-s-t, Gafat so-s-ta). The Tembu no-do-so is eyiae of Nubian derivation. The Tumali nda-ta, although resemblin; the Zi forms, appears from its final vowel to be Semitic. The l'ul ta-tie, tu-t, appears to be also Semitic through Galla sa-di, su-de, In Berber the Ist th, as we have seen, is hardened to k. Some of the Himyaritic dialects a to have hardened the 2nd th, and to have transmitted this form to Africa,« The Gara (Ekhili) tha-k-it pectin the Himyariti¢ source of the Agau sha-k-wa (wa is a Libyan definitive postfix, replacing the Reales —t,-d, asin the Agau lo-wal, ak-wa 5), umi- der shu-gha, Falasha si-gha, Shangalla w-ka-~ (both dentals hardened). The Gonga up hee Ke-s (2 for th, ¢, as in 2, and as in 3 of the Dalla # he-zza, Sodss: e-dja, se-ke-che, If the last form’ ( Katfa) be the one and the others contractions, the -che, —za, -dea is the def. postt., and se-ke the root, correspondmg with the u forms and with the Gara tha-k, the a of tha softend to e as in the Hebrew and Chaldee forms. The Shangallagiorm corresponds with the Kuafi o-ku-n: and represents the Nilotic parent of the Sudanian o-ku (Hausa), wu-ku, bu-ku, ya-sku (Bornui}, Pika &e. ku-nu, d-ko-an, ko, Buduma ka-ke-ne, ke-ne &c., as the u does that of the Mandingo sa-k-ma (Vei), sa-k-a ( Susu ), sa-g~? (Jullunkon, in 8), sa-ba, sau-wa, sai-bi, Serakoli si-k-a, and of the forms in sa,—Darfur is, Emghedesi ayin-su (ayin-ka’ 2), Ponti e-bi-sa (e-bi-en 2), Fetu a-die-san, Akim di-an-sang, Atutu a-ssah, Ave- kwom a-za, Amina é-sa, Akripon i-san. Dental forma of these sibilants also occur in Nigeria. Woloff ni-at, Kru ta, Tambu e-ti, Ibo a-tu, Papa, Whidah, Grebo tnoh, o-ton, Moko, Karnba i-ta, e-ta, Karapay e/-tonp, Panwe tuyh (Shiho a-d], Yoruba ma-i-ta, The Begharmi ma-ta isa simi- * lar variety. Koelle’s yocabulary gives numerous West and Mid African varieties of * ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 23 the Zimbian forms, tati, tat, ata, ita, eta, eto, ta, esa, cha, nta, bata, meta, guta, kotatu, bitate, butet, dsitadu, delaro, pelalo &c. &e, ’ Foreign. Ajfinities, A. The Egyptian term is Caucasian kc. (see Semitic), and its disuse not only in the Semitic but in the African 3,—although preserved in the a ie Yo 8 (5, 3),—is equally remarkable with the almost uni- versal prevalence of B. in the Semitic and African languages. B. The affinities of this double unit are indicated in the Semitie list. It has only one répresentative in Caucasian; but it is Indo-European in the t-r form, (whichis the Scythic k—-r, k-1), and Scythic m the t-t, t-k, st, s-s,@ k-r, k-l, h-r, gr torms, Its wide prevalence in the 5S. W. portion of the Old World, (Indo-European, Semitic, African), and the-circumstance of its having apparently supplanted the older Chino-Scythic term preserved in Caucasian and Keyptian, render it probable that it early became appro- riated to 3 in the numeral system of an influential and diffusive South Western race. In1 the same double form occurs in Caucasian systema, zi-3, ho-s &c. In Semitic it may have been fem., and sho-m masc. Obs. on the Distribution -of the terms. Following the analogy of the glossary generally, and on _histori- cal probabilities, it may be concluded that the terms similar to the Arabic, such as the Berber, are the latest Asiatic ithportations,— that those similar to the Himyaritic, such as the Nilo-Nigerian sha-k &c. and the Amharic and Nubian so-s, to-s &e., belong to the Him- yaritic eri,—and that the prevalent and widely diff ta—tu, ta-ru v. belong to more archaic ayes, like the common African terms for 2. They are probably of common origin with Semitic, rather than oF Semitic ort ‘he beyptian 3 appears toshow that when it was received from a Semitic language, the Semitic family had « greater variety of terms for 3 than it has had in later Gines.* In 4imbian the forms ta-tu and ta—ru pb have co-existed, and they probably did so in the earlier Semitic “eT he wide dissemination of the forms tatu, taru, rara &e. was probably effected by the great Zimbian movement. Pe That of the several Nilotic forims indicates important movements of Ni- lotic tribes to the westward, subsequently to the Himyaritic era. The Am- haric forms of Himyuaritic have carried to the Nubian tribes, and thence to one at least of the Nigerian tribes. The Agan forms—corres- ponding’ probably with the original full forms of Gonga, Shangalla and ‘uafi—must have been carried westward by an influential or dominant tribe, for they are more widely diffused in Nigeria than any others, The Mandingo tribes are probably the modern representatives of @his great Ni- lotic movement. The Fulah movement appears to have been a later one, and the term tor 3 concurs with other ee facts in indicating that the influence of the Galla migrations and conquests extended at one time into Sudania, and was thenve transmitted to Nigeria. * In the pre-historic era of the formation it is probable that Semitic had more dialects than in after ages. The more barbarous the tribes, the greater their segregation and the more numerous their dialects ; and the’ capacity of the family glossary for variations im the forms of rovts and their compounds, depends on the number of dialects, 24 ETHNOLOGY OF THE INDO PACIFIC IsLANDs. : The powerful diffusion of Himyaritic terms by different streams appears to show also that the Nilotic tribes received a t impetus from the ci- vilisation and energy of the Himyarites or an earlier Shemo—Hamitic race. The Galla, Fulah and Mandingian tribes probably derived from ther something of their physical superiority to the purer Negro races. Four. A, (a.) ar-ha-vd Mahrah, Gara, ar-ba fem. ar-ba-ta mase. Arabic, ar-hi-ch mase. areha fem. Hebrew, ru-bu Babylonian, » ar-bea mase. ar-ba fem, Chaldee, ar-ba-tt Tigre, ar-ba-ta Grafat, ar-at, aryut Amharic, ar-at Harragi, u-bah Arkiko, ar-um Turuli, {ra Bullom, e-re Yebu, per- af Ziphianl Egy fi mali, fe-re Dank i, b.) a-f-t, f-tu Egyptian, fon-so Tibbo, a-fu-r Saumali, fe-re Dankali (baler Shiho; baha-ra Dankali 8), hau-da, au-da, hoitda; oi-da Gonga, ~u, hu-du, fo-du, o-du Hausa; fu-lu Kalahi; e-fa-r, e-fa-tra, e-fa-d, e- fa-tu, e-fu-tsi ke. Malagasy (e-fa in 40), The Malagasy e-fa-r, with the allied African terms preserye the original full form of the Semitic ar and show that it was originally 2 repeated (2 dual). The same term is common in Africa as 2, bar, war, &c., contracting in Kuati and Tumali te ar, ‘The Semitic 4 (@) is also, inal] probability, a dual of this form of 2, with the archaic labial posttix,—ar-ba from wareba, It recurs in 7 (4, 2) and 8(4 tual), See the remarks on 2 above, and also Semitic Numer- als, 2,4, 7. ‘The contracted prefixual ¢ of Malagasy is found in some Nigerian terms, It is Semito-Libyan (a, e, i,&c.) and oceurs in the ni, Car Saumal terms under the form a. . {a,) zaa-cha, an-zaecha Shangulla, ach-ech Gonga, si-za, se-dza, sa- ja Agidu, ma-che-che Makua, mu-tye-tye Mudjana, se-se-s Mazambiki, si-ja Kongo, ‘These terms are similar to the double unit of 1, 3 and 5. They have no apparent connection with any of the Semito-African terms for 2, and may have been formed from 3 like the Kaffa4se-ke-che—se-koe (3,1), Similar terms recur in 9 (5,4), and in the Semitic 9, which not only resembles these African terms for 4 und 9, but the Semitic 6 and 3. All these affinities would be explained by the Semitic 3 being formed from 1, that is having originally been 2,1. In both its varieties tha-la &e. and tha-th, it is represented in current terms for 1. (See the remarks on the Semitic 9 ), - (b,) Berber ku-z, Shillah ko-s-t, is a similar term. The Nubian - ke-m-su, ke-n-ju, ke-m-so-ga may be connected with it. Both resemble terms for 5, and possibly 4 may have heen ‘1 from 5,” but it is much more probable that they were formed in the ordinary way from terms once cur- rent as 2* The Harragi ke-t, 2, (a contraction of ko-l-et), resembles the ee (ay Th Bishari u-ddi-g ( dig in 9}, Bornui di-gu, de-ku, Emghede- . (a.) The Bishari i-g (dig in 9 nui di-gu, de-ku, si a-ta-ki are probably 2 dual CBornui ea! 2, Fulah didi &e.); diis one of the variations of the chief Semito-African root for 2, na, in, il, li, di &c. In Bornui the guttural is postfixual in other nouns also, See (¢). (%.) se-lle —te-le in 9 (5, 4),—is probably from a similar term for 2, Leis the 2d element in} and2ulso, Ind its immediate origin is probably the Semitic 2, she-ne Hebrew, s-en Egyptian, si-l-i/ Kallahi, (c.) The Zimbian term is the nasal Seiuito-Atfrican root for 2, ne, ni, na, nai ke, It prevailsin the Nigerian as well. as the: Zimbian proviree. Buaheli ne, Kikamba tna, Kinika w-ne,Tukwani vi-nai, Masena /i-na, Sofa- ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 25 la nai, Kosah ne, Sech. i-ni, Nyambana gi-mu-ne, Zulu i-ne, ma-ne, Benguela kwa-na, Angola wa-na, Kongo kwa-na, ya-en, mi-na, Sonho, wa-na, Embo- na m-na, Mp enfi. Compare with the Southern Makua {Takwani] and Mpongwe form the Gabun terms,—Datanga de-nai, Bongo ba-nai, Pan- we be-ne, Akuongo mi-nu, Camancons me-ley, Kerapay e-nay, Karaba ina, e-nang,, Runge, Calbra ni, Moko i-nan, Kaylee di-nan; the Binin nin, [bo ano, Papah ene, Akripon ne, Whidahe, Efik inan, Yoruba mene, Fanti, Akim anan, Amina anani, Avekwon ana, Grebo hanh, Kru nie Bullom nen- ol, Timmani pan-li ( pan, pur &e. is a_pref.), Kissi iol, Woloff ni-an-e?, Mandi nani, na, Fulah ni, nai. The Darfur ong-al, (in 40 on- ge-val ) is probably connected with this Zimho-Nigerian form. The Yangaro nan of nan-giri 8 steers to the same term. e In the Zimbian 4 the most radical and persistent element in the Semi- to-African 2 appears as the original term. In 2 it has the forms in, na, roh, nra Semitic, il, li, le, ri, re, ne, nu, ni, di &e. African. Some the amplified forms, na-i, nani, nan, i-nan &c. involve the Semitic postfix as well asthe radical element ( ith-na—ni Arabic), or, as is more probable, they are the original dual form in full, 2, 2. . These double forms are not found in the E. and 8. W. Zimbian dialects. They occurin N, W. Zimbian— be-ni-n Melon, e¢-ni-n Negoten, dé-ne-n Isuwu,; in the adjacent Chadda. v. na-n Ham, mi-nyi-n, nyi-n Tiisi, a-na-r Koro, a-nye-ra, nye-ra uku; and in the N. W. Nigerian na-ra-to Gadsaga, pa-ne-re pa-n-le Timani, na-ni, na-n Mandingian gr. The term is probably equally ancient with 3 and 2, and referable to an era of the Semitic system when the liquid was the proper root of 2, and had not been conereted with the labial masc. or sibilant fem. definitive used asa prefix or initial. To this period its acquisition of a dual and thence of a plural power is probably to be referred: ; Foreiqn Affinities If the African terms for 4 are all Semitic of different periods, and based on Semitic terms for 2, their foreign affinities can only be considered through Semitic, and as illustrating its arehaic condition. The only term of interest in this respect is the Zimbian. In several N. and E. Asian systems the pure liquid definitive is found as 2and 4, Aino i-ni, Korean nai, Ugrian ni-la, mi-] &e.,4. (See Semitic 2). These forms, with those in which it occurs a8 2, render it probable that it was used as ‘a numeral element in Semitic prior to the concretionary era,—a conclusion that is supported by the history of the language generally, which carries back the numerals to the period when the detinitives were free, and capable of being used as wnits. e Zimbian 4 appears to belong to that era of the Semitic system when the liquid root had not become agglutinated with the initial definitives. At the same time it must be recollected that a contracted term existing in one dialect may obtain a wide currency through the spread of an influential race. The history of the Zimbian 4 must be considered in connection with that ef the other numerals, which certainly fayours an archaic, and not a recent, derivation from the Semitic system. Obs, on the Distribution of the terms. 1. The contracted historical Semitic forms, including the Himyariti have made little progress. ’ Mg TNS? / 2. The Egypto-! alagesy terms appear to preserve the full form of the Semitic, and are probably pre-Himyuritic or archaic Himyaritic, and of ~ > 26 ETHNOLOGY oF THE INDO-PACIFIC ISLANDS the same era as the similar and most prevalent African terms for 2, now 3. The Zimbian term appears to be equally ancient. Although the broad ar (from war, bar wey is the prevalent Semitic form in 4, it is pro bable that slender forms werealso once prevalent in 4 as in 2.* | The broad Zimbian terms with the fobial prefix which prevail in. the Western group, wana, wan, /?-wana, bi-wana, buna, with the Felup Ju-bare-gen, have the seme form as the Malagasy e—far &e. 4. The other terms, with the exception of C.(a), which is probably of ual antiquity with the Zimbian, appear to belong to that era of Semitie when fem. forms bad begun to replace masc. TVrony the distribution of these terms they appear to be of later introduction into Africa than the Egyptian, Galla and Mulagasy form. The prevalent Semitic 4 may be considered.as exceptional, because it preserves the same archaic mase. form. + The double sthilant &c. of Shangalla, Agau &c. was probably that of a Semitic dialect which had replaced it by the fem. form, although all the Semitic languages afterwards assimilated im their use of the masc. form; under the influence, if may be surmised, of that single dominant la which has produced so large and remarkable a uniformity in the Semitie numerals and pronouns, and in much of the general glossary also. Five. The terms for 5 are similar to forms of the unit 2 found in lower num- bers, 1, 2, 3, and the most common, as well as the closest, affinities are with forms that are used for 3,—as has already been noticed with reference: to the Egyptian 3 and Semitic 5. In the prevalent African systems & is the hi hers number in the first series, six being 5, 1, seven 5,2, &e. It might therefore be well expressed by one of the names for the unit, 1 hand, or 1 tale, and the term would naturally have a plural or collective form ag in 2, 3,4. Nour having been exprsssed by 2,2, the collective unit for 5 would probably be taken from forms used as 3. But from some of the terms itmayra- ther be inferred that 5 was originally 3, 2 or 2, 3, as in some other formations, The Semitic kha-m-sa, kha-me-sh, Berber su-mo-s &c. is the form of 3° (1, 2) preserved in Egyptian, sho-m, sha-me &c., followed by the principal or sibilant numeral root, which may haye represented 2, or a remnant of 2, she-ne, ath-in &c. The Gallo-Zimbian form—which only differs from the Semitic in having the@liqnid in place of the labial second element— elosely resembles Semitic and Zinobian forms of 3. Comp. sha~n, sa-nu, ta-nu, ta-ui, ta-ru, so-lu, lan, lo-lu, &e. 5, with she-ne, ath-in, sa—ni, su-nu, zo-le, ra-n, ta-la, sa-la, tsa-la, tea—ra, sa-ra, ka-ra, ka-ru, ku-na ta-ru; Jato, la-ro, lal, la-n 2, (i. e. 1, 2)« * Koelle gives wer-be as the Arabic of Beran. Possibly thew is archaic. + From some of the forms in higher numbers given by Koelle, it appears probuble that the second labial is radical also, and that the original emitic term was bar~bar or war-bar, i, e,2, 2. t The Mandingian group preserves several of the variations,—so-lu, so-li, su-ti, lo-lu, no-lu, ndo-lu, du-ln in 6; su-n, se—ni, su-ra in 6; su-nr sti-ln, so-lo, sora. in 7; su-n, so-lo &e, n 8 &c. In 3 it has distinct Semitic mid Nubian forms, sa-ra, St-Ta-n, sa-@-wa, sa-wa,—sa-g-wa being Himya- ritic. through Afau. Theo, of the higher numbers is the archaic Semuti¢ form found in Egyptian, Berber, Bishari, Galla&e, « ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 27. Some specific instances of the reapp ee of forms for 3 in 5 will place thefact beyond doubt. Gadsago 5 a o, (Nubian and Bornui nit postf. ); Kandin ka-ra-d 3. Banyun moto-ki-la, ki-la—/ 5,—the same vari- ation of the Semitic 3, but preserving the slender vowel, as in ke-ra-d Berb.; the Banyun 3, ha-la-/, has the Kandin a, asin the Zimbian la-la, la-n, la-/, ra—ro, ta-ro, ta-ru, tsa-la, Mandingian sa-ra-n (preserving the Semi- tic s form of the ‘at inthe 8. E. Zimlian tse-ra-ro—), 8. E. Zimbian k-la-n 5 (Nyamban), similar to a N. W. Zimbian 3, Negoten be-la-n, Isuwu de-la-ro, Calabar-Chadda e¢-la-ro- Nalu te-du_5, Funtian te-ty ta-tu &c. 3. Timbuktu os be 5, Bornni u-gu, u-ru, Hausa u-ku, o-ku , Pika gr. ko, ku-nu &c., Kuafi o-ku-ni. Mandara i-li-ve 5 ( 3, 2), Ndob (S. of Mandara, apparently in the Chadda basin, whence probably acquired its Zimbian character) le 8, be 2. Ba- pina mi, Hausa bi-al, bi-ar; this may be either 3, or 2, closely allied ms occurring for both numbers; the Hausa bi-u 2 ay bi-ul &¢.) is in favor of its being’ 2, but similar forms are common for 3 and 5. Fulah dso-wi (5), dso-we, dse in 6 (5, 1), Mandingian so-lu, Bo-ko so-lo, Mbarike i-tso-n Dsuku tso-ana, Woloff dsu-dom Nuti, gr. gu-tsu; these are typical of the most common forms, save that the sibilant has generally a and the liquid u,—ta—nu, sa-nu &e,, Zimbian, The u of nu &c. as well as the prefix, shows eae forms to have been derived from 3, ta-ru, ta-ro, ta-tu &e., and not from 2, in which the liquid root has generally slender forms, li, ni, ri&c, and takes the Jabial prefix Forms in o and u are also found in 3, e-to, ¢-to Isouma gr, age ae gr., to-re Bagbalan, ne-o-do-so Kiamba, ma-dsou Padsade, bi-dso Biatada, ’ A (a) tu, tiu, tie (in 50 tain, teui) Egypticn,—the unit in the archaic Semito-Libyan 3 form, as in the older African forms of (b) su, tu, and without the labial.* . The sibilant unit is found as 5 ( 3 for 3, 2) in the adjacent Bishari 6, 7, 8 and 9, in the forms su, she, se. In 6 it has the form se-ra (1 A. hk), the Semito-Zimlian form of 3. 4 | The u, 0, form of the dental unit (varinble to the sihlant &c.) is pre- servedin the 8 of Gara, thu-m (3, or 5}, for 5,3), corresponding with the3 of yptian sho-m, with the 1 of a and Malagasy t-so, Egyptian uot, =wo-t), Nkele, Bongo wo-to, Undaza mo-ko, Murundu ¢o-ko, Grebo gr. , Boko do, Afudu ka-do, do, Mbarike n-dso, n-dzo, Veido-ndo, Kauro ku-dum, Kiamba u-dom, ka-lum. — (b.) kha-m-sa, kha-m-s fem., kha-m-sa—-ta mase. Ar. kha-m-ish- shah m., kha-m-—esh f. Heb., khe-m-as Mahrah, kh-ish Gara, kha-m-is- ti Babylonian (kha-n-sa 50) kha-m-sha m., kha-m-esh,f. Chald., a-m- is—t, au-m-is-t Amharic, (ka-m-sa 50), au-m-ish-te, ha~m-ish-ti Tigre, a-m-is—t, ha—min—is-t Harragi, hu-m-is-ta Gafat, ha~m-za African Ara- bic and Emghedesi, a-m-us Arkiko, su-m-us Berber, su-m-os-t, Shillah, tu-m-at Timmani, o-ma Tumali, bu-su-me Dalla, [See Semitic Nume- rals]. In these forms, which are radically 3 (i. e. 1,2) or 3, 2, the initial _ anit varies to kh, k, h, s and t, and its vowel to a, e, and a, The u, o form—probably the archaic Himyaritic—appears to have * Ihave not seen Lepsins’s paper on the Pgyptian numerals, but it appears from rie reference to it in his Egypt, that Lepsius considered tuto be? (for 2 + 3. + 6 Ke the same form in the tu-m, su-m of Berber, Dalla Ke. (see 5), . 28 BTHNOLOGY OF THE INDO-PACIFIC ISLANDS. been early and widel apt over North and Middle Africa—pre- sedi the a form of Zim ian (¢). It is found in Darfur os,*—with m (for 2) in the Dalla, Berber, Timmani and some other archaic forms of the Semitic 5 su-m, tu-m &e.—with » (for 2) in the ole tn, sped Sgeatita in several emis vocabularies. : O-wt -g0, dso-nwe-go 6,i, e. 5,1), Dsuku tso-ana, a-tso- ana, Dsarawa to-nan, Mbarike. i-ts0-n; Abed mbe-dsu-en, be-tu-en, Okulomo so-ne, Mandingo gr. so-lu, su-di ke., Woloff dsu-dom, dsi-rom in 6, 7 &c., [Kiamba hv-dom, hu-lum 1], Felup .fu-to-gen, Filham fu-toh, Dahomey er. a-to, Nufl gr. a-tu, gu-tsu &e., Ibo gr. i-so, n-to, Ham to, MepTeete a-ru, é-ru &c. Ereg’ba i-thu &e., See b, ¢ aud d, for other examples of the wide currency of o and u forms. To the Babylo- nian su-su of 6O Colonel Rawlinson has now enabled us to add su-—nnu 2 (identical with the 3d pron, pl,), tsu-du 6, ru-bu 4, su-ma-na 8. * (c.) The Galla and other forms with the final liquid n, ana, lu &c. have so wide a mange that they must he considered separately and more ful- ly. Sha-n, shu-n, za-n Saumali, Galla, ko—n Shiho, ko-no—yon Dan- kali, ta-no Suaheli, i-da—-no Ki-Kamba (de-n in 6), za-no Ki-Nika (ha-n, ta—n in 6), ma-ta—nu Makua, mn-sa—nu Mudjana, x-hya—nu Makonde, vi-ta-nu Takwani, si-nu Masena, (a-n in §), sha—nu Sofala (ta-n in 6), tha-na Masambiki, tha-nu Dalavoa Bay, n-ka-no Nyambana, txla-nu Zulu, txla-rn, cha—nu Sechuana, hla-nu Kosah, ta-nf Benguela, Angola, Kongo, Kambinda, bi-ta-nu Mundjola, sa—nu Sonho, Bunda, ma—ta—n Camancons, al-to—ng Kerapay, ¢-tu-ne, i-ti-en Karaba, o-ta-ni R ‘ so-ni Calbra, ve—ti-n Moko, bi-ta-n Kaylee, ba-ta-n Bongo, ta-ng Binin, a-to-ng Papah (Dahomey), u-nu Akripon, num Akkim, a-nu-m Amina, e—nu—m Fanti, -u-nu Ashanti, { probably frora mi~u-nu, mi, bi &e, being refixual in other mimefals in this group), m—u-—n Bullom, m—-u Kru; so- fi Vai (the Ist element with o as in Papah, Calbra, Karaba, Kerapay, Shiho and Dankali), su-li Susi, (the 2d element with i as in some of the lower Nigerian and Gabun terms), du-li Sokko, do-lu, lou Mendi, In lu Mandingo, na—lu Pessa, Ja-lu Kosse, lotlu Jullunkan, ngue-nu Kissi, ma—lu, a—ro Yoruba, ; The Hausa bi-at, bi-ar, Kashna bi-et, Kallahi vy—die, Mallowa ‘he-a strongly resemble some of the Galun forms of the Zimbian }i-ta—nu, vi~ to—nu. Kwallalifa has ba-k-wi. In the Calla-Zimbiwn term the root varies from the sibilant and dintal to the guttural, and in Mandingo the common interchange of the sibilant and liquid takes place, solu, lelu. Other liquid varieties are given by Koelle, e-lun Penin, ili-ve Mandara, elon Eafen, be-ron Mbofon, {Hausa diar, bial], ‘lan Nyambana, ba-di Pika, fa-di Bode, ma-ru, a-ru, e-ru, e-ri Yoruba pr., e-lu Igala, ba-nu Guresa. Koelle pra gives be-tane Nki, mi-tan Undaza, be-ta Murundo, be-tui, we-taiy sLOnguan, The Galla—Zimbian term is evidently cognate with the Semito-Li- byan sibilant 5, with the m2replaced by the n 2. Comp. Berber su-—m-ws, Kandin su-m-¢s, Shillah su-m-os-t, Timani tu-m-a?, to-m-at, tsa-m- at, Dalla by-su-me, Landomage-tsa-m-ot, Limba ka-so-f &c, The Galla shuen, the Shiho and Danakil ko-n, ko-no, with some of the Zimbian and west_ * But this may be the remnimt.of a term like sum-os (see ¢), + On the Fibers of u, 0, forms m the earlier Semitic, and their we to Africa by the Babylouo-Uimiazitic colonists, seep. 4. ETHNOLOGY OF THB INDO-PACIFIC ISLANDS. 49° ern.terms, tso-n, dsu-en, tu-en &c. have the same archaic yowel. The Ba= bylonian kha-nsa 60 is an Asiatic example of ener coee of m ton. In Galla-Zimbian terms for 10 the Semitic compound occurs in both the m and n frysns tu-m, tu-n, du-m, tu=m—un— &e, Walla fam ku-mi, ku-ni &e. Zimbian; ta-n, ta-muj dso-n, dso-b in other vocabulurica, The Gulla forins suggest.that tu-n, su-n &c, may have been coutractions of lost Semitic forms of tu-m, su-m, which took the nasal or plural posttix as in 2, in place of the sibilant (generally fem.). In the other Semitic and Beyptian terms in which the combination kha—m appears with the initial unit in the sibilant form, the labial has a final n (8 Sem.—Kex,, 3,10 Eg.), Kambinds has sa~mba—nu 6, sa-mba-ida 7,. Kongo ma-su-toba-iu 6, @-tsa-mbp—— adi,7, Angola sa—ma-nu 6, sa—mbo-adi va Benguela panda 6, pa—ndu-ali 7. Phe ids adi, ali, of 7 (5, 2)is2- But the tinal nu, ndu of 6 is not. the urrent 1. elt may be said that both 5 and 6 wre varigtions of archaic forms for 1. The Zimbian sa-ma-nu, sa-mba-nu contains Ue same liquid root 6 and the same pretix, and micht appear to be only wii additional pretix, The liquid oc- curs With both prefixes in 2, sa-niysa-nu, s-roh &¢. mbua-ni, mba-n; and if - it has the power of 1 here 1t may indicate that the second series has eom- menced. . There are rempants, however, of similar forms of 1 and the analogy of the other African systems, with the use of the 6 form for 4 in 7, * mightseem to support the conclusion that both 5 and 6 were variations of a term for 1. Comp. gt-en, /u-n, gu-n, dsi-ni, n-ne, n-n0, we-ne, ke-bo-ne, ba-ne,. va-ne, pa-le, wi-an, mo-re, bala, ha-lo, hele, ya-i-ni, pu-lo-lo, fé- no-d &c, with the fiqnid forms t.r 2. ‘The Shiho vurtation of the Gulla 3, ko-n, is identical with the Akurakura 1, ey The prevalence of the liquid unit as the radical element in 2 and its dual. and plural power have been noticed. ‘The terms for 5 more often resemble those for 2 und especially 3 than for 1, and/it may well be thatin 5 the wuite~ also had the form appropriated to 2, or 3, because in the unit had neces+» sarily # pluralor collective meaning, aud in its original condition indeed was probably 3,2. The Galla-Zimbian sha-n, sa-n, sa-nu, tu-nu; ta—ru, , ta-n; 80-h1y 80-10; so-li, lo-lu,-no-no,* e-la-n, k-la-n, are similar to the 2 of Semitie’she-ne, ath-in, su-nu, tu-r, te-r, s-roh; and the cognite African terms zo-le (Zimb.,)ma-ra-n &c. but the same forms the common 3, (1,2), + Semitic, Zimbian ke., and a comparison of all the dinects renders it clear that 5 was 3,2, and that the terms now current generally contain 3 only. (.d), tu-pa of isi-tu=pa 6 Zulu, and -to—ka 6 Gatangu, has the same form of 5 with the labial for 2, In 8 and 9 it is preserved as 10 in the form to-ba,. ‘The form is similar to tu-=m of ()-) and (¢.), and to the Babylond- Layptian 3d pron. mase,, su-va, tu-t. Thal (e). TheAgau ak—wa, ank—va (-wa postf. as in lo-wal, sha-k-wa 3)" is an analogous termto(a@.) It is probably a contraction of shak—-wa 4. ). ma-ku-8 Shangilla, hn-ch, j-chi-sha, hu-che-sa, hu-cha, Ganga gr, te-su, di-dju, di-dja, di-ka Nubian Ceomp. di-k, 1, Darfor). "These are: examples ofthe comma double form of the nuit, or anit iad fen postilx, used for 3 in Mahrah, Gates Nubiin, wid Abysinnian languages &e. (82 B, c.). Ln. the aspirate form hu-eha ke. the initial unit has the same form asin the Arabic and Hebrew Sd proa. (ho, hu), In the Gonga hu-u-pona corresponding with su~su 0 of Babylonian, it retains its radical value 6 as 1 (1-5, so la=pona, 7, i. ¢. 2,6, hos—pona 48, i: e, 3, 5,—3 hetng the same Semito-African sibilant unit), ‘he term reeurs in the Shabbe or Kakauda 30 ETHNOLOGY OF THE INDO.PACIFIC ISLANDS. of the Lower Niger as 1, in the compound hoos—war- of 2, 3 and 4, B. ib Bishan. This may bea form of the labial unit as in 3 mih, but se prohekty only the archaic labial postfix as in 3 and 8, the root being e C. po-na, pu-na, Woratta, Wolaitsa in 6 (1,5), 7 (2,5), 8 (3,5), and? (4,5), fu-n Yangaro iné. ‘This is the labial unit, but it may have been more immediately derived from the she-men, she-mon, form of 3 by contraction. In the Kissi 6 ngom—pu-n (5, 1, ); it has the same form. So-ba-1,1, Bullum. It also oceurs in 7 of Ki-Kamba mon-sa (5, 1 ), and Ki-Nika fung-ahe (5,1), and, with a form similar to the latter, in the Tumali ‘fun-as—an (10, 1) and fung-en 10. The Darfur wi-ng of 10 Cin 20, 30 &c.) is the same term. D. lima Malagasy. This term does not appear to be a Semito—Afri- ean unit, In Asonesia it is used bothas 5 and as aterm for “hand”, and the latter was probably its primary meaning in Malagasy also. Its aN, E. Asian term; and in Zimbian itis still current for “tinger” and “toe” Jemi, liemi, liam, lembu Ke. ; Foreiqn Affinities. The various forms of the most common Semito—African term, kha-m, su—m, tu-m, sa—ma, sa-mbo &c, are N. Hi. Asian ko-m, Samoiede so—mb, s0—bo, sa—ba, sa—m, and Mongolian ta-bun, ta—bu, in which the labial is an archaic qualitive postfix. Forms with final n also occur, tung, sun &c. In the archaic Scytho—Chinese or FE. Asian system a conneetion is also observable between these names for 5 and some of those for 3. The Ramoiede and aap ees sa—m, sa—ba, ta—bun of 5, agrees with the 3-of Chinese sa—m, and Caucasian sa-mi, su~mi; andthe Tungusian tung-ya, tong-na, Kamschatkan ton-o/. &c.{in 5,8, 9) &c. with the 3 of Chinese sa-n, Yeniseian tong-y¢, dong-em, and Mongolian kol-m, kor-da &e, The Uerian and Turkish 5, vi-s, vi-t, vii-si, bias, biesh, bi-l-7/ &c. is 3 in Turkish wi-sse &c., Japanese mi-tsu. The Caucasian chu-ba, chu-thi, chu-t 5, is similar to the 3 of Caucasian chi-ba, su-mi,and of Ugrian chu-d- em, (in 6 chu-t), ku-m &e. The sibilant unit of 5 in Japanese, is a like form to the Egyptian &e, (a) and the tlouble form (7:). ‘The labial unit (C) is used as 5in Ugrian and Euskarian; and the Indo-European term may possibly be the same, The adjacent Caucasian chu-thi, chu-t, wo-chu-si Georg., t’chu Circ., chu-ba Awar are analogous to Semito-African forms, Obs, on the distribution of the terms, 5 presents a repetition of two of the roots for 3, the sibilant, guttural &c, with the labial, and with the liquid, -s-m: s-n, The form of the sibilant &e. with the labial postfix is evidently very archaic. It is preserved in other numerals also, and the cognate Caucasian, Scythic and N. E. Asian terms are proofs of its high antiquity. The Semitic final s (kha-m-sa &e. ), like the final n of 2 (p.16), appears to have been added in the Asiatic branch after the archaic Semitic system had been carried to Africa. Thearchaic or pre-historic u, o, form of the Semitic unit appears to have been conveyed by a Himyaritie dialect to Africa, and to have been spread from the Nilotic province to the Atlantic, over the Northern and Middle regions. It probably preceded the later Semitic a, i, forms in Gal- Ja and Zimbian also. See the remarks on the Ngyptian 3, ante p. 4, Since that page was printed the full*Babylonian series of numerals, as given by ETHNOLOGY OF THE INDO-PACIFIC ISLANDs. al Colonel Rawlinson in the last no. of the Journ. of the R. Asiatic Society, has corroborated the inference that the detinitive in the u, o, form was the ee unit of the Semitic system both in 1 and higher compound num- ers, It is found in the Babylonian su of 2, 8 and 60, tsu of 6, and ru of 4. The Babylonian 5 has a later form, and probably replaced an archaic one similar to that preserved in African vocabularies, sa-mu, tu-mu &c., to the Egyptian sho-m 3, and to the Gara thu-m, Bab. su-ma 8, The Himyaritic and Agau-Mandingian sha-k-a 3 is not found in names of 5 although preserved in Mandingian terms for 8 and iy the Fulah: 10. The current Semitic forms for 3, tha-la, su-la, sha-lo &e. Arab., Heb., Chald.; tha-th, shi-sh, so-s Himyaro-Nilotic, are similar to the most pre- valent African terms for 3 and 6. They have evidently spread from the upper Nilotic province over the rest of Africa, protanls replacing the more ancient or northern Nilotic form with the labial postfix, in most of the vocabularies which had previously received it. Their use probably dates from the time when the liquid postfix superseded the labial in the Semitic 3. The dialect, perhaps a west Himyaritic one, from which they spread as a centre, must haye used this form both in 3 and 5, The principal ‘rican diffusion of the t-t, t-n, t-1, t-r, -r &c. forms is distinctly referable, first to their prevalence on the upper Nile, and then to the great Zimbian dispersion over southern Africa and Nigeria, including the Chadda province and part of Mid—Africa conterminous with the Niger and Chadda basins. The Semitic 5 in its-archaic uw form ei cae to have previously spread over northern Africa. Its presence in Dalla, ( Egyptian), Berber, Shillah, ‘Kandin, Falah, Gadsaga (in 6) and in several of the sangunges of the Senegambian coast (Landoma, Timani &ce.) is thus explained. It is found along the Semito—Libyan band from the Persian Gulf to the Atlantic. Sz. Tn the Semito-African systems, as in most others, the terms above 5 are repetitions or compounds of those of the lower series, Six is 5, 1, or simply 1, or it is 3 (dual); seven is 2; eight 3; nine 4; and ten the second 5, or 6 dual, (2, 5; 5,5; 255), The Semitic 6 is probably 3 dual. Nearly all the African terms are 5,1,— the 5 or the 1 being sometimes ¢lided, Some of the Eastern and Southern Zimbian terms are 3 ( dual); the others are 5, I, or, by contraction, 5, or 1. A. (@) s-ou, s-ov, 8-00 (in 60 se) Keyptian. This is the sibilant detini- tive and unit, with aremnant of the labial postfix it has in 8 and 8, and probably had in 5, See C, B. si-t, si-te, se-te fem., si-ta-ta masc. Arabic, sli-sh-shah mase., she- sh fem., Heb, shi-tta masc. she-t fem. Chaldee, tsu-du, (su-su 60) Babylo- mian ; sha-t-id Gara, ha-t-id Mahrah, se-d-es¢ Amh., se-d-ishte Tigre, se- d-is herber, shi-tta, shi-dda Hausa, su-th Shillah, zu-du Kalahi, { = teu-du Bab.}, z-du Bode, si-tta Emehedesi, se-tta Suaheli. The unit occurs here in its double or fem. form, as in the Himyaritie and several African terms for 3 and 30, It is probably therefore to be considered - net as a quinary term, butas the second 3(3 dual), The Gonga 6 shi-r-¢a, shi-ri-ta appears to preserve the liquid element of most of the Semitic forms of 3 (se-le-ste Tigre, sa-la-the Ar. &c.) The Darfur sitta sun-dik appears to be a double term; sitta is the Arabic 6; sun-dik is the Gallo- Zimbian 5 followed by the Darfur 1. ; | -. The Babylonian tsu-du, Kalahi zu-du, is a link between the dento-sibi- 33 ETHNOLOGY OF TUE INDO-PACIFIC ISLANDS. lant and the liquid forms of the definitive, the former brangs se imyari- tic (Mahrah, Gara) and the latter the common Semitic n of 3. The Sof Galla su-de, the Bab. d of 6 as well gs the u of the first element, The liquid form is also found in some names for 6, Ndob so-lu, so-la, so-ro. This is the 3, sa-la, tha-la, sha-lo, she-lo, of Sem., with the Him, 30; of Zimbian, ta-ro, ta-ru &c.; and of Mandingian sa-ra, It is also similar to forms of 3 found in 5, so-lo Boko, so-lu &c. Mandin- gian gr., and to the Semitic 10 @-sha-ra, a-sa-ta &c. ‘The connection be- tween 1, 3, (5), 6, 9 and 10 in the Semitic system is well illustrated by some of the Mid-African heat ya which also preserve the Semitie form of 10 in 3and6. Bornui gr.1 las-ga, las-ge, la«kn, 3 yas-ge, yas-qu, 6 a-ras-ge, a-ras-gu; Calabar gr. 1 e-dsi, dsi-dsi, dse-t, 3 ¢-sa, eon 6 sa-ga-tsa, a-sa-ra—sa, be-sa—ra—le-sa, e-sa-r—e-sa; Dsukul a-tsu, 3 a@-tsa-la, a-tsa- ra. In some of these forms 6 is obviously 3 dual. dv-sa-ra, ¢-sa-ra, a- tsa-ra is the form of the unit preserved in the Semitic 3 and 10, Asonesian | sa—ra 1 &c. C. (a) su-geoor Bishari, 5,1. The sibilant unit su, she, se, represents on ee and ) [See 5, A(a@)]; goorisa term for 1 found in nu-gpir di. @Seel, A. ¢:, 7.). dse-x0, (lso-we-00 Fulah ( go, go 1 ). ; ()). gor-ju, gor-go &c, Nubian; the same form of 1 as the Bishari oor, c.) ogu-su-ee, gu-tso-ai, to-ai Nufi gr.; 5, su, tse, to, as inthe Eeyp- tian and Bishari 6 (see 5, A. a); ai, ei, 1, for wai, wei (as in tnu-wel, another form ) a contraction of weni. f D. dja-ha, dya, ja Galla, Dja-ha &e. may be a variation of the double sibilant of Semitic &c. asin 3, See also 5 A. f. Sr E.. (a.) Veh Saumali, leheye Dankali, leh Shiho, er-de Dalla, el-el Tumali, ille Kuati, hm-le-do, hm-le-@-le, me-le-do Grebo gr. (5, 1). Thia is the liquid form of the unit as in Dalla ille, Agau lo-wa |1, B. es). Comp. 3, le Ndob, ra Mfut &c. ili-ve Mandara, (see 3 and 5). (). wa-l-ta, wo-l-ta Agan, fa-r-schu Nub., be-li-wa-l Dselana, va-la Panwe. The liquid is the root for 1 in Agan lo-ma, la-gha, Nubian wa-r-um &c.; far-shu muy be 1, 5, as di-suis 5. The most prevalent Nigerian 1 has the form of Dalla (er-de 6, ille 1), 'Tumali el-el 6 &e. Comp. 1 pu-lo-le Kan- yop, ke-le-n, ke-le Mandingian, e-li, e-ni Aka-lgula, la-ku Bornui (Agau —gha) &e., ke-de-n Bode &e, The Kambali to-li, Hum to-ni, Yola e-ri Penin ¢-len-daro are cognate terms (Bornui ti-lo 1, Shangalla me-te-1 1 : (c). e-ni-na, e-ne, e-n, a-i-ne, u-ne &e. Malagasy, ene, ini, uno, ono, ani-m, inu-m &e. &e. in Asonesia [See 1. A. 7f:]. Aka-Igala e-ni &e. 1. Jt may be from the liquid 3 (5). : F. (a.) wo-ra, uo-ra Mandingo, wo-ro Jullunkon, Sokko, ro-wa Mpong- we. Woro appears to have originally represented 5 in this numeral, 5, 1, 23 it is found joined with 2 in? (Mandingian). (See 5 C.) (). ma-i-fa Yoruba, e-va Yebu, a-wa Vanti {1, A.] Koelle gives ¢-fa, me-fa, as the form in all the dialects of this group—Aka -Jgala. It isnot the current 1 of] the group, which is me-ne, e-ni, é-li, i-ne in 1, and mo-ko, wo-ko &c. in 11 (Idsesa has o-ko 1). It is the dou- ble labial prefix as in some of the Zimbian forme of 1, and indicates the archaic carrency of forms like me-fa—ni. Comp. bd-ne’ Gadsaga, ke-do-ne Nki, fa-no-d Felup &c, The labial prefix is found as 1,2 in several dia- lects, The double labial occurs in Zimbian terms for 2, The Calabar and - ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 33 adjacent inland dialects have be-ba, e-ba, e-we, and some of the W. Zim- biani bi-ba &e. 2 (for be-a-li &e.), and this form of the pretix is, by its vowels, the same as the me-fa of the adjacent Aka-Igala 6. pay Ge husu-pona, hosu-puna Gonga (Woratte, Woluitsn).1, 5. (1, A. f.). ar hosu, hu-su is 3 (he-zza), it here clearly retains its primary. v power, for pona, pun represents 5 in hiwher numbers also, 7 (2, 5), 8 Bae (45 ). In 8 the same forngis used as 3 (hos-pona, 3, 5). » den-tatu Ki-Kamba (i-tatu 3), han-dabu, tan-daln WKi-Nika (ha- hu 3), an-hatu Masena, tan-hatu Sotnln, ya-taru Sechuans, daa-tanta Gu- resa, These terms are 3, 3, asin Semitic. I, _sa-ma-nu Angola, Sonho, Bonida, [See 5, A. ¢.], pa-nda Benguera a elided), m#-sa-mba-nu Kongo, /i-sien-mu Mundjola, mchasan Raa inda, ye-vala Panwe (vala for vata 1). Koelle gives several additional examples, and all in the W. Zimbian group, to which the term appears to be contined,—/i-sa-ma-na, di-sa-ma-n, gée-s8ema-n, mi-sa-mu-nu, ba-se-mi-n, bi-se-mi-ni, p-sa-mba-nu. These variations are the%same as in the Semito-Libyan 8, shu-ma-n, tha-ma-ni &e. The Amharic and Gonga form, se-mi-n, is identical with the Mutsaya ba- se-mi-n, Nteredi-se-mi-ni. Both are radically 3, the Zimbian dual based on the form ef 3 preservedin Eeyptian sho-me-n-t ec, Gadsaya has also tu-nn. At the same ini the sa, ta isthe unit found in Sand the second element ma-n, ma-nu, tabi-nu is similar to the Semito-Zimbian yu-li, wa-l, ba-ne, mbo, &e. 2. K. (a). ma-tanu na ui mo-dya Muakua, m-zauayi mo medi Mudjana, #i-tanu no mo-si Makonde, thana mo-asa Masambiki, abyanu na moji Takwani, nkana-yana Nyambanua, &e, &e. (5 and 1 ). ; ). i-ti-a-ket Karaba (5,1), i-si [bo, 1 eel a ndday Kerapay, é- * gi-a Fanti, in-si-a Ashanti, e-schi Amina, e-schen Akripon, m-ischiang Akkim, at-ugo ne ac! of Dahomey (at, is from attong 5, Zimbiun,—ugo is the guttural 1 of Yoruba &c.), ta-hu Binin (tan 5, bo 1). (¢), mu-i-edu Kru (mu 5, dn 1), mein-bal Bullom (5,1), ngom-pum Kissi (5, 1), diu-rom-ben Woloff (5, 1). (d). ue-itah Kissa (itah 1, we-tah, we-tah Mendi, ai-tah Pessa, dso-we- go, iowi-ero, ie-fa Fulah 5, 1, | ; The Isuwu or Moko group, to the N. of the Kongo-Angola, has the eastern Zimbian form 5, 1, or adual form, wenerally much contracted. Ji-ta- na-i wote Baseke (comp. the* contracted 8. 6. Zimbian na-i-modsi for ta-na ni modsi), The other forms may have 3, ora remnant of it, eto wa, pi-nto, nto, tu, nto-), ntu-ru (Oomp. Mand, ndo-lu 5, i. e. 3, Zunb, ta- yu 3). Butit is move probable that mu—to-we, n—-to—). is the dental form of the Semito—Libyan 1 (Comp. the Shangalla me-ta—jm, mi-ta-! 1), The other western and middle groups bave generally the form 5, 1. e. g. owa-ra-zon, Akurakura, ton-sar-mo Caetano. e-len-daro Penin, e-ta-tio Murundn d¢-ta-ndat Konguhn, solo-do Boko, tsof-unte Lamba, kilan-ga- duk Banyun, basj-modi Pika, hinon Dgbira, ban-tanta Guresa, hm-ledo Grebo, sun-dondo Vei, men—bul Mampa. | mint L. tsi-tupa Zulu; tupa here represents 5 apparently (5 A.d.); isi is the substantival prefix. 2-toba Batanga is the saine term. M. wa-ta Shangalla, 1, (me-ta-me 1, mi—ta-t, Calla ti-ka, Himyaritic “tii-ut, tant, Genes i-ta ke. jy fa-d, g-fa-d Bulanda (fo-du 1.) : N. bali wal Dselana (2, 31. e. 2d 3, bali 2, ba-ta 3), 84 ETHNOLOGY OF THE 1N80-PACIFIC ISLANDS. Obs. on the Distribution of the terms. : The Semitic 6 is exceptional in its formation as 3,3. It is Irano~Scy- thic and not African. Of the three Semito-Libyan terms for 3,—s-m,; s-]; and 5-9, s-t, t-t &c.5 the first is obsolete in the Semitic 3, but is preserved in the Egyptian 3, the Semitic 5, and the Semitic and Egyptian 8. The Eeyptan 6 may bea remnant of it, and from the close resemblapce of the Kongo-Angola forms for 6 to the Semito—Nilotic for 8, it appears certain that the former is the archaic Semito—Libyan s~m form of 3. TheGadsaga tumu a to be a distinct remnant, From the position of this dialect on the Senegal, ite proximity tothe North-African linguistie province, the North-African affinities of its other numerals, and the Semito—Berber form of tumu, it is evidently of the same era as the Egyptian 3, sho-m, and Berber—-Tim- mani 5 su-mu, tu-m. The second form s-l, s-r &c. is the current Semitic 5, and one of the Zimbian forms. As 3 it is found in Gonga and a fey of the Chadda~ iverian vocabularies. e third form is the Himyaro—Nilotie 3, common also in Zimbian, It is the current Semitic 6. In Africa it is not common, Some of the Northern forms are Arabic and evidently modern. But the Kalahi, Shillah and Bode mark the ancient presence of the Babylono—Himyaritic form of Semitic in N. Africa. Some of the E, Zimbian terms are also formed in the Semitic mode. _ With hardly any other exceptions the African terms are quinary.’ inary terms are found in most of the proper Nilotic languages, and similar forms are, as usual, traceable in the West Nilotic or Nubian voca- bularies. The Nigerian 1 in most of the groupe is the Nilotic liquid; and the 6 of the Grebo group as of Dalla has the Dalla—Nigerian form of 1; Bornui has the form in 6 as in 1; and the Bullum gr. has the Nubian. The Fulah guttural 1 of 1 and 6—preserved in 11 of the Aku- ala group mo-ko, wo-ko, mo-ka (Idsesa has it in 1 also o-ka), 1 of Ashanti e—ko, and in some of the Calabar-Gaboon dialects pa-ka, mo-ko, o-ko &c.—is the Semitic a—~kha{=ma-kha] in one of its archaic African rms. The E. Zimbian quinary terms are formed from the current 5 and 1, and do not appear to have spread. The W. Zimbian 5,1, and 3 dual, are also confined to that group. = Seven. The terms are quinary (5, 2, or simply 5 or 2). . The Semito-Egyptian 7 is the unit in the 3 and 6 form with the labial postiix, and at p. 7, I considered it to be 6 for 6, 1,—there being no traces of the Zimbian subtractive naming in Semitic to admit of its being ex- . plained as 3 (from 10). Mr. Koelle’s Zimbian vocabularies, and the iden tification they have enabled me to make of the names for 5 and 3, show that the Semitic 7 is not an exceptional trinal term, but quinary like the African ones. It is a contraction of 5, 2; and as 5 was itself 3, 2, and is represented in 7 by its first element, 3, the term is identical with forms of 5. Further, 3 was radically 1, 2, so that in the fall original form the root for 2 must have occurred thrice |(1 X 2) +2] X2. The first element representing 5 is the ordinary 1 in its sibilant form sa, the second element, the lubial ba representing 2, is from the labio-nasal 2 common in Africa sud preserved in the 4. | . RTHNOLOGY OF THE INDO-PACIFIC ISLANDS. 35 A. sa-be, sa-ba, fem., sa-ba-fa masc. Arabic, shi-bea mase., she-ba fem, Chaldee, ha-ba-id Mahrah, shu-a’ Gara, shu-ba-te Tigre, su-bha-¢, se-ba-t Amharic, su-bha Darfur, sa—ba-ta, she—ba-ta Gonga, sa-ba Suaheli, se-t, aa—d Berber, sha~sh-f, se-fe-ch Egyptian, [z0-s—pi Euskarian]. The Zimbian forms render it certain that the Semitic are 5, 2. " Moko gr. (Gaboon-Cameroons) sa-mba, s¢-mbe, et a (i. @, sa, se, sia from ta-ni, tea &0.5; mba, mbedrom mba, mbe, ba2); Mfut ta-be (ta from ta-n 5 as in’ ta-fay 6 i, e. 5, 1, ta-ra 8 i. e.5,3: be 2); Ndob sa-mbe {sa-n 5, mbe, be 2); Bute ta-bamn; Yasgua to-mva, (nto 5, mva 2); Dahome gr. te-we (a-to 5, owe, eve 2); Kongo-Angola gr. sa-mboadi, sa-mboa sa-mbat, su~mbdal, tsa-mbodia, sa-mboali, tsa-m, ndza-ini, sa-mbids biele, tsa-mbodia ta-nu, sa-nu, ta-n Ke. 5: biwade, woale, baol, biele, tu-wids, bol, miol &c. 2. These terms are composed of the first or more radical element of 5, and of 2 in its full form. ) _ A still more prevalent Gallo-Zimbian form preserves 5 uncontracted. _mi-tann na ui medi (6 and2) Makus, m-zona-zi-viri Mudjana, vi-tanu na viri Makonde, mhyanu na i-vidi Takwani, thana pili Mazam- biki, n-Kana-ti-vere Nyambana, kam-bini Zulu, tan-na-peli Matatan, tanu na beli Kiriman.” Mr. Koelle gives for Nzamban #-lan ni zimbede (t¢-mbe- re 2), In the Moko gr. Baseke hes bt-tane-ba. The form is also com= mon in the adjacent inland or Chadda-Nigerian tongues,—ten-sa-bari, n-send-fa, ¢sinedlele, e-dsan-efe, d¢-tann-iwa, tanda-ra-mot, tson-ifa, o-ton~" afa, &e. Nufiin the same proy. has gu-to-aba, tnu-aba, ato-aba, Okuloma sono-ma, Ham tor-fo [to-ro-ba Gallal; Mandingian so-lo-fere, sulu-firin, Ashanti solo-pla. -The Nilotio languages preserve similar forms. t sera—mna—-), Bishari, 5, 2; the Semito-Zimbian se-ra 3 for 5, as in 6, 8, 9; ma-)2, trom mal-u+ Comp Mandingian sora-pere. , to-r—ba, te-r-b Gulla, vdu-bah “Dalla, ko-lo-du Nubian, tu-l-ur- Bornui;-to-r, ko-lo is the Semito-Zimbian 3 (1, 2); ba and du 2. umel-neneDankali, mel-hen Shiho, bar-de’Dalla; the 1st element is 2” fell-ad Nubian, mar-‘o Tumali; the 2d is merely the nasal and dental poste” x. : “The Zimbian sa, ta, of sa—mba, tsa-m, ta-be é&c. is the sa, ta of ta-n, sa-n &¢. 5. But the Semitic sa— ha-, ‘su-, she-, she~, se- differs in form though not in root, from the first element in the current Semitic 5, kha-, - khe-. But the African forms in 5 ha--, su-, tu-, preserve examples of the forms used in 7. ” me fa) li-nye-ta, la-ngi-ta, Shangalla, .la-nga-ta, la-m-ta Agau. The Agau2. _ The iid Wigetian lo-ba is a contraction ef the Boko solo-p-la 5, 2, so-lo being a Chadda-Nigerian and Mandingion form of the Semito-Zimbian 2 . and Zimbian 5. ndingian has solo-fere é&c. (). la-pona, Ja-puna,na-fun (2,5) Genga; the Gonga, Agau and Galla la of 2 (5 G.). yiys _G. mon-sa Ki-Kamba, fung-ahe Ki-Nika, The labial 5 as ind[5. G.j with the sibilant 2. — 4. Bulom gr. mena-tsin, Mandinge woro-fela, woro-nlo, wo-fela, Grebo gra _ mu-leso, hm-leso &c., Mpongwe ora-genu. The sibilant 5 (3) with the labial prefix occurs in-some of the Rigeriea vocabularies, ba-dai-belu 2) Karekare (ba-ulo Pika), me-dse, e-dse Igala (5), pa-dsi-na palo Bola gr, yl . $6 ETHNOLOGY OF THE INDO.PACIFIC ISbANDSy _ nabi-shana Kunafi, 2, 5; shana is the Galla-Zimbian 5, and na—bia eon— tracted form of the Zimbian 2. D. bok-woi, bak-si, boek-wa, bek-ai, wok-oi Hausa (5,2)., Kwalalifa has bakwi 5, (3,2) and Hausa bu-ku, u-ku, ku-an &c. 3, (Bornuiu-gu, U-rit ). s > weer E. fi-tu Malagasy ; aform of 2 similar to several occurring in African ‘languages for 2 and 4 (2dual ), and therefore cognate with the precedin terms in which the labial 2 is used for 7, The 4 of Malagasy e-fa-t, 0 Eeyptian f-tu, and the allied African terms, are similar forms. It may cana 5 in a Zimbian form, bi-tan, fu-toh, (3 bi-tatu, ba-ta, be-sa, -ten éc. . F. g-fad gi foda (6 and 1). : Obs. on the distribution of the terms. The full forms current in most of the Zimbian vocrbularies cannot have been derived from the contracted Semitic ones. They represent the ar- chaic Semitic forms, and preserve both the Semito-African forms of 3, s-m &e. and s-n &e. They former keeps its place in the common Semitic 5 as wellasin 8, and belongsto the era of the Egyption 3, The latter he- longs tothe era of the current Semito-Zimbian 3 and Zimbian 5, and ap- pears to have spread from the upper Nile (Bishari, Galla-) Bight. _ Most of the terms are quinary (5,3). Some of the Zimbian are qui- nary, some are 4 (dual), and a few are 2,10, ) A. tha-manj sha-man jém., thn—ma—ni-ta mase, Arabic, she-mon-ah -masc. she-mon-vh fem. Hebrew, te-man-—ja mas¢., ta-mn-e fem. Chaldee, tha-man-id Mahrah, thu—m Garo, she-mun-te Tigre, se-min-* Amharic, se-min-t@, shi-min-fa Gonga, the-man-ia Darfur, sh-men, sh mun, sh- mun-¢, Heyptian, te-m Berber, te-mp-@Shillah, du-ba Tumali [du =su, tu], mai-sn-wa, do-sa-wa, mai-sag-wa Mandingien gr.; 3 or 5, 3. . sumbhai Bishari, 5, 8 (su for 5, mi 3). Probably the term. is the same as A; mih 3 may. be a contrattion of a term like su-mbai, a yariation of the Egyptian, : = C. sa-di-et, za-di-et, se—le-id Galla, Saumali, (3 of Galla &e., Seni- tie), shi-dan-fa Kallahi, si-dda Kwollalifa, se-re Masena, Sofala [{ Galla se-ori, se-de), se-tti Soko, se-ri Mandingo. Zambian and Mandingian forms preserve the preposed 5, /-lan ni ze-ra- _ro Nyamban, tim-na-taro Mataten, In-tane-bi-tads Baseke, dsowe-tati Fu- lah, son-tatat Limba, sun-sag-ba, solo-ma-sare &e. Mandingo er. D. ba-lia-ra Danokil, ba-h-r Shihe, This is 4 dual, 4 itself being 2 dual. Itia the common labial term for 2 found in the Gallo family in 2 and4, The Malacasy valu 8 is the same term, The final u is that of the Bishari malu-b. [See 2, B. ¢.}. IS, su-gua-ta Shangalla, sa-gho ta, sa-gha-ta, so-ho-ta, so-ta Agan, The Agau 3. ; ’ See C. The Nubian form of sa-ra is also fourid without the preposed 6 in Gadsaga, se-Eu, I’, qnon-que-dah, Dalla, 5, 3; quon, which recurs in § and 9, is the form of 5 in Shiho &c. kon. In queduh for sette 3 the sibilant becomes guttural, as in the Gonga 3 ke-#lja, ke-s &e. G, hos-pona, hos-puna Gonga (8,5), The same form of 3 oecurs,as the unit in 6 husu. 3 is heza ke, © . ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 37 H, nan-giri Y : giri probably corresponds with gin of iz-pin 8 which is the os cake the Galla- family: mai ratiionse oS be the Chadd. ah oy and Zimbian 8 and 4. . . e-du, i-du, Nubian. The root e, i, is probably a contraction of is, 3, Darfur, 9 is u-e-du. L. wu-sku Bornui ( wu-ku 3 Hausa, yassea 3 Born.), _ N. ta-ko-as, to-ko-s, to-ku-s, to-gu-s Hausa ; ta—ko 3 for 5; as, 3, as in a—to-asa Nuti. i, 0. (a). ma-tunu na ni ma-raru Makua, m-zana-zi-tatu Mudjana, vi- tanu na vi-raru Makonde, nhyanu na i-tatu ‘Takwani, n—ka-na-ti-naro Nyambana (5 and 3). (%.) The following are contracted (rallo-Zimbian quinary terms, ye-ve- Oh Panwe, is-ieti Kuati, e-tye Fanti, iti-ita, ete-atah Karaba, ttte-iata loko, as-atu Ibo atija-tong, Papah, enni Kerrapay, mui-eta Kru, auo-tui Ashanti, de-sas Timmani, dui-rom ni-at Woloff, ngom-mag Kissi, me-ni-ra Bullom, ta-ra Binin, suli-ma-saku, sagi &c, Mandingo group, no-a-gui, anguee, uguee Amina ke,, ggietu-tar Nufi, e-yo Yebu, ma-jo Yoru- ‘P. na-ne Suaheli, Ki-Nika, nin-nia Ki-Kamlg, bg-na-ni Zulu, kie-na- na Benguera, a-na-na Kongo, na-na Kambinda, ‘na—ne Sonho, na-nai Miongw es (4 dual; see 4 EF). gives several didditional examples of this form, e-ne-ne Alege, ne- ni Ndob, na-na Basunde &e, ¢-re-ri Egbele &c. &c. — toba meni-mbina Zolu (10 less 2). This isa second Zulu form in which toba represents 10 as in 9, toba no monyi, 10 less 91. Bina is pro- perly 4 (Takwani, Mandjola &¢.), but ibis here evidently a corruption of lini 2, R. The Sechuana hera menoana me-beri is formed in the same « way, hera being the term for 10 as in 9, ne g (a), mboxo, bozo, po-hlongo Kosah, mpuo Mundjola, paamo Mut- saya, woam Ngoten, fomo, fama, fomi, fam, fom, fan, lo-mbe &c. Moko gr. 10 (a contraction of 2 from 10). : b.) owa-mbi Batanga( 10,2). See 10K. . ongo-Angola nake, nak, Ji-nak, di-nake. This appears to be a con- traction of 2,10, Bangun has ha-nak 2, kila-ga-nak 7, ha-lak 10. T. danda-fudu Ngodsin (fudu 4), fi-tedu Karekare, pordo Pika (4). U. dogar-so Baghermi; 80 4 (dogar-mi is 9,1. e. 5, 4). Obs. on the distribution of the terms. Considerable variety is manifest in the modes of forming 8. The most common is the quinary. The Semitic term preserves the sibilo—labial 3 of Egyptian. ‘The ‘Nilo-Nigerian band has both this form and the s-r, gd, or current Semito—Zimbian, 3. The Mandingian forms are of Nubian dnd not of Zimbian derivation. The cognate Zimbian,3 (of Nubian origin) is also found in some of the Zimbian terms for 8. * Dual forms are found in Hast Nilotic vocabularies, in Mid-African and in West Zimbian. The last are referable, not to the eurrent 4, but to the double form found in some Chadda—Nigerian und in the Mandingo voca- bularies, and preserved also in the Yangaro nan-giri, It is the Semito- Nigerian form of ai * ss The, forms 10 less 2, and 10, are only found in a few Zimbian dialects: They are too fare to indicate any specific affinity with Asiatic languages which possess such forms, But they render it probable that the archaic —_ , a TS 2 _" a — a ——E——— << — —— * 38 " ETHNOLOGY oF TUE INDO PACIFIC TLANDS. | Semito-Libran system, like the cornate Svythic, nsed different modes of expressing the numbers above 5, before the terms for them became con- creted aud contracted. ‘Thus six was probably 5, 1 and 8, 3;—seven 5, 2; 3 trom 10; 6, 1;—eight 5, 3; 4, 4; 6, 2; 2 from 10, In gome of the vocabularies two forms ave still current, e.g. ‘Tiwi has for 6 witan karmon 5, 1, aud @-tari-ta 3, 3. ; Nine. Most of the terms are quinary (5,4). Some ure 10, 1, or simply 1, i. e, 1 short of 10. ‘ A, ties’, ti-sa’ fem. ti-sa-ta mase, Arabic, ti-sh-ch muse., te-sha fem. Heb., ti-sha’ muse.; te-sh’a fem, Chuldee, sa-id Malowh, Gora,’ ze-tti Ambiaric, ze-te-in Harragi, za-te—va Gofit, ti-shu-te Tigre, dje-fa, yi-dea Gongs, ti-ai Dariuy, t-ga Kandin, p-si-6, p-si-s “Eeyptins, pit i980. sa-sa Shangalia, se-s-ta, se-ssa, tani-eha, si-cha Agau (se-lza, si-za Xe. 1), All the forms of A appear to be 3 dual, 4, or 4,5. The Bey ptian term, os Lepsius has indicated, is 4, 5, p-si4,s 5 (YO pes-tuin). The upper Nilotie se-3, Sa-su, are foul in 4 ond the form is the Vhmnyastiib v she-die Bishori 5,4 (uddie 4), — t sa—ye-al, su-ve-al Shiho, Saumali, Galla, se-ea-la Danakil, sa-1 Kuufi, ta-ra Housa, iz-ci-n Yayiguro, is-ko-du Nubian; $1, su,- se, iz, the common Semito-African contracted form of 65 gu-l, ga-la, pi-n, ko-du is not a current E. Nilotie 4, but it is the commom Semito- ‘African root of 2, 4, with the outtural prefix as in the Abysiminn” form of Remitig ki-le-te &., in the Aim!ino ko-le, kua-rv &e. In sone of the remote Nigerian languages it is retuined both in 4, Sand ®; 4 ku-ba-he-l — Vilbam, —he-du-k-+ Tila, he-ea-e-r Seceres, hu-la-Ke-r Kanyop; 8 bu ki-rei Bola, ba-gi-rei Severes, ba-ka-ri Peel; 9 kanyen- ga-lo (5,4) Bola, Sereres. quon-telle Dalla | 5,4, selle 4, See & F. ). hodu—popa, hodu-puns Gonga 4 4, 5 ). u-e-/u Nubian ; uncertain; e-du is 8. fan-dsano ‘lumali; 5, 4, a5 in be; or 10, 1, The KE. Zimbiaa, like the Semitic and Nilotic, terms, are 5, <4, and most of ‘the Mid-Africun are also quinary, kea-da Savhek, Ki- ika, i- ken-da Ki-Kamim, toba no monyi Zula, f-'an ni move Nyamban, we-tann ne ui ma-sheshe Makun, m-zaiia, tyetye Mudjana, ei-tauunma vi-rn-iuei ~Ma- konile, whyanu na t-iyetye Tikwani, ogi-t-war-ne Nauti, e-mu-na Fanti, jasois Bisin (nin 4), atj-eene Papuh (cone 4), me-io-nehol Iejhum (ne- nol 4), de-ai:-li Timmani (pau-li 4), din-rom anet Woloff, suli-ma-nani Kissi, dse-nai, dso-we-nti Fulah, ta-ra Hansa, solv-mia-nani, su-nasi, mai-on, ko-non-to Mandingo gr., Bulum, mena-hiol, ku-dsidsi Kambaii, #-si-ti 8600, ehi-rs Bipio, i-te-vena lsonma., si-vi \/alagasy ; si for 5 as in maoy African langrinires (si, se &e.), vi the Zimbinn coutracted 4 unt 2 (as in be, 2, Ndob, Mfut). The most common West Zimbas term is 1 (for 1 short of 10), di- VOR, bi-vO, Yi-VO, i-¥y %-VOa, (you, vou, bua, va, abe, d=bog, si~po, é-fi, ve, de-bun&e, Mall torms are found in Secluana, hera mono-ana monya-hela (1\) less 1), Mpongwe, ina-gomi, (/-pomi 10), and Orang feni-homi, Genyueis has hie-kii (kiwi ia 10, the term for 1 dropped), Okam sorv-wone (10, 1), Akuraucu osuie-asob (1, 10, )jlum mbon-kb “1, 10) &e, =D «. o ape > ETHNOLOGY OF THE-INDO-PACIFIC ISLANDS, — 39 The Buduma heli-gar, Bornuile~rar appears to be 4, 5'(Buduma herai 4), Asimiiur collocation and the sume voohtor 5 occur in the Ekumtu lufa ele-gale, Udom ale-gale 8,71. e. 3, 5, ene, be-le 3. nae Ten The archaic full form of the Semito- African 17 was2, 5 (i. e. 2d 5.— 5, 2, being 7. A tew examples are still extaut, e, ¢. mar—kum Shillah, bi~n-bad Pika, d@i-sin-yane Pepel, bela bue, bla-bue Grebo gr.; but the common forms are contractions, 5, or 2. A, (a,) a-sha-r fem., o-shin-Fa-ta mase. Arabic; a-3a-ra-h masc., e-se-r fem. Helrew, es-ra-¢ Bubylonian, 9-s-@ mase., a-sa-r fem. Chaldee, a-ish-ri-d Mahrah, ish-ri-d Gara, as-ur-te Tiere, as-ra, 08-ir Amburic, a-shi-va, a-s1-4r Gonga, “ha-la-& Banyun. This is the Semitic 3, but as 10 cannot be from 3, and as the common African 10 is (dual), and Sis 3 (for 3, 2 it restilts that ia the archaic Semitic system, both forms of 3, s-m an 8-11, Sel, 8-T, Were current for, as in Arica, For the Atrican forms of (a) See (band a). / (4 and @) men-t, mn-t, me-t, fem, me-t-i, me-t-e Egyptian, 5 dual. ‘This is the other Semito-African 5 and’ 3> in a contracted form (sho-men—t, sho-me-ti ke. 3, Me., su-mu-s 5 Berber &p., tha-man, she- moon &, 8 Semitic, sh-men “Ey.). The tull form is retamed by puper Nilotic languages, and as both the m and n forms occur inthe same dine lect, 1 place them towether. ~ -ta-mu-nt Bishari, tu-ba-n Dankali, tu-be-n Shiho, to-mo-n Kuafi, ta- ma Woratta, Wolaitsa, di-ma-ga, di-m-1u Nubian, tu-n (in 40, 50 &e,), zu—n (in {H)) Shiho, tu-n, du-n Saumali (2Y, 30 &e.), tu-na (im20), du-m om ), tuem (40, 50 ), tu-ma (60, 70, 80.90) Diiukali, ta-mu Gadsaga, edu-ob Anan, dso-b Akurakura, Okam, di-om, do-m, Moko vr., ta-mon Serakoli, te-ng Mandingo, ta-u Vei, Jullunkon, (ta-u 5. Zimbian). ti-ng Sokko, to Kissi, to-fa-t Timmani, (to-mut 5), i-du Ashanti, Fanti, e-du ‘Amina, Akim, u-e-du at ee te Biniti, The Fulah sak-po, sa-po, Kambali hok-pa, is the same form of 3—Him- _ yaro-Nubian—that 1s found in the Mandingo gr, _ Tt is remarkable that while the forms of the Southern Zimbian dialects adhere to the Galli variety, both in the initial element and post- fix, the most prevalent West Zimbian and Sudanian term returns to the Semitic form of Dinboth. The cull fortn is preserved in mar-kian Tibbo (2,5), mulo-po, le-guini 8. &. Zimbinn, kn mi Suaheli, Ki-Sika, Taic- want, Masena, Sotala, Anvolo Mundjula, Kumbinda, #-ku-mi Ki-Kemba, ~ di-ku-mi Mudjava, ma-ku-mi Makonde, ¢-ku-me, ko-mi Nyambana, d-shu-mi Kosah, shu-me Sechuwna, kn-i Benyuera, kwu-ni Argola, shu- mi Sonho, sha-nhi Bunda, i-to-mi Mpongwe, e-ka-me Kongo. In the Moko group, the initial particle has the variations d, lyr, eh, di-om, ¢-do-mi, i-lo-mi, kv-ro-m, p-rim, gum, t-ho-ani. The same forms occur in Sudanian vocabularies, mar-ku-m Tibbo, gu-ma,co-ma B de, o-ma Hausa. They occur alay in the Chaddu-Nigerprov., lu-m Dsariwa, du-ob, Anan, o-za-he, o-ta-be, za-be Koro, ko-b oem v—ko-b, w—ho-b Yasena, dsu-m Mandara (in 11 &c.), misku-m* Unduza. They are also found in the Senegambian prov., Limba k» f [=ko-b Ham], Baga to- fa:t, to-fa-ts, Ashanti gr. o-ku-ru, ku-li. aly > The guttural forms depart not only from the Semitic 10, but from the Zimbian 5, aud approach vo the Sewitic guttural 5, kha-ui, of kha~m-ish, ‘40 ETHNOLOGY OF TAB INDO-PACIFIC ISLANDS, kha-m-is &c. But the yowel of the initial is the archaic Semito-African u of the definitive and unit (hu, su, tu &c.), as in the N. African form of the Semitic 5, su-mu &c. The original of the m form, in all its varieties, is the Semito—Libyan guttural, dental or sibilant definitive and unit as apphed to 5, followed by the labial or masculine postfix, and in some cases with superadded postfixes of number or gender. (e.) ku-dh-an, ku-du-n Galla; this appears to be du-n &c. of } (the Zimbian 5), with the guttural prefixed. The Dalla quila kudde conjpins the Galla term with another. cing B. (a@.) man-ku-s Shangalla (5 in the common Himyaro-Nilotic 3 (b.) chi-ka Shangalla, Agan, tsi-ka, su-ka u, ta onga (ib.). G The root for 2 pe Bas anne in a large number of the voca- bularies, bure Nubian, fung-en Tumali, wing Darfur (in 20, 30 &c,), mer- aun Berber, mar-ow Shillah, mar-an Kandin, [mar-kum, 2, 5 Tibbo); Landoma maran, Moko gr. u-wom |e-vuo Ngoala must be, de-bua, bo, e-voa, 7-vua, /i-vo, di-yoa Kc,, ntsere is given as 11, but it is used as 10 in higher nombers|, Ndob wum,-wuom, owum, Mbe wum, Nso yum, Tiwi uo, Konguan bin (2 in Hausa), Atam (Chadda-Niger) womo, ewuwo, bofo, ee opt, ubo, bo, Biafada wa-popo, Padsade pare Nalnu ¢e—bele, W. or Upper Nigerian fe-r, hi-ru, fu-ra, pi, ii, Lower Ny, i-li, i-ri, ig-be, Nufti e—wo, a-wo, &e, PS hae Aku-Igala me-wa, e-wa appears to be the double prefix of the mbian 2, ; The Mala vu-lu, pu-lu, Moko bu-! is a full form of the Zimbian 2, similar tothe adjacent 8. BH. Zimbian mu-lo, W. Zimbian bu-ol, bo-l &e. 3 In the Grebo gr. vu, pu, pe appears os oadingian and not 2, ie, ue (2, 5) occurring in Ghe. Hence the ingian pu, pu-go, fu A also stand for 5. (The other current term tan is 5). athe same western range the Mampa wan is 5 (man). : Sudan bim-bad Pika, bum-bad Karekare 2,5, (bad 6, bi 2 is Zimb.). The Bulanda g-fad nign ta-sila is “6 and 4,” The Bideogo woru ago is probably 2, 5, i Wolof fuk appears to be referable not to fog 1 (Nfut &c.) butte The Zimbian term is of upper Nilotic origin. The full form mulo- Matatan, mar-ku-m Tibbo, oa a Nilotie form of 2, and the 5 isa Himyaro-Nilotic in its vowel, tu-m &c. When the forms tu-mun, ku-mi, were adopted by the Nilotic vocabularies, it is probable that a similar form was current in that of the Semitic colonists from whom it was derived. The vowel of the common Zimbian form ku-mi is that of the Hebrew and Babylonian 5, kha-mi, The Dallasu-me5 and Egyptian sho-me & has both vowels. Shihohasthem in 10 also, tu-be-n, and itis probable that Egyptian had both sho and sha in 10, as in 5 and 3. From the persistence of the s-m, t-m, k-m form in the Nilotie vocabularies, and particularly in Egyptian, for 3,5, 8 and 10, it may be coneluded that it was the prin- cipal form of the oldest Semitic colonists of the Nile. The m-r, m-l, b-| &e. form of 2 is probably of the same age, as it is the common upper Nilotic and Zimbian torm for 2, and appears to have been associated with the s-m, k—-m form of 5. } The s—n, t-n, k-n, s-], s-r &e. form of 3, 5, and 10 probably became the principal term of a leading Semitic nation ata later period, As 3 it ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 41 appears to have been communicated by a Semitic tribe to all the Nilotic . languages save Vishari and Heyptian, and to have spread over all Africa, chiefly through the Zimbian movements. As 5 it must at one time have been used by the leadine Semitic nation, otherwise it would not now be found in 10, In 6 it became the principal Nile-Zimbian term. For 8 the Semitic race adhered to the m form, but the some African tribes that had adopted the n, r form for 5, used it in 8 also, In 10 while the Semites have reserved the r form, most of the Africans have adhered to the older m orm of the Egyptian 10 and Semitic 5, SumMARY OF INFERENCES. 1, The Semitic and the African numeral systems are connected by a common archaic basis and by the wide diffusion of later dialectic names and forms in different eras, They are thus, in a large sense, dialects of one mother system. The dialectic. modifications have been great and re- peated, and the result is that each of the existing systems is very irregular, 2. In the mother system the current definitives were the numeral ele- . ments, several definitives were used, and in their numeral applications they probably retained that power of distinguishing the. genus of the substantive which they possessed as definitives. The system originated in an eva when the monosyllabic definitives had not become concreted, and when they migh freely compounded. The basis was binary. The naines of higher numbers were obtained not only by addition but by multiplication, and, when the denary scale was assumed, by subtraction also. The mode of indicating the higher numbers appears to have lo remained arbitrary—so long probably, as the different elements retain their identity with the common definitives, aud were not concreted in the compounds and phonetically unitised and modified so as to become separate wards. In this stage each number admitted of being ex- pressed in several modes,—the unit, for example, varying with the class of the object, while the definifives preserved their original appli- cotions, and afterwards arbitrarily, while several generic detinitives con- tinued current,—and six being three-three, twice three, jor five-one. The full double series was (T.) 1; 2; 2, 1, for 3; 2,2, for 4; 3,2, for 6; (II) 5, 1, or 3, 3, for 6; 5, 2, tor 7; 5, 3, or 4, 4, or 3, 10,dor 8; 5, 4, or 4, 5, or 3, 3, or 1, 10, for 0; 2, 5, for 10. In the concreted systems the name for 3—itself penerally a form of 2—hecame the most important, as a form of it also represented 5,-the radix of the quinary names in the second series. The process of dialectic change consisted in a gradual impoverishment of the archai- abundance of roots and names, in the concretion of compounds into separate or independent words with the loss of one or more of the component roots, and in the substitution for the native dialectic names of others borrowed from the languages of influential tribes. This process destroyed the original homogeneity of the system in every dialect, and reduced each to a series of terms having only an obscure or a slight ety- mological consistency, save in those African dialects in- which the names above 5 are still undisguisedly quinary. Some dialects still porseas more than one name for the same number. Thus Bornuwi has three words for* 1,—1ti-lo or fu-lo, Ja-s-ye and pa-l (bu-r in bu-r-go-be, first), and Tiwi has two for 6, witan—karmon, 6, 1, and @-tar-tar 3, 3 ; 3. When the numeral names became concreted and “ab hot they took the current definitive prefixes and postfixes of each dialect, and a new source of diversity and of phonetic change was thus introduced a2 ETHNOLOGY OF THE INDO.PACIFIC ISLANDS, These seconda y elements, in their turn, became subject to concretion ; and the contraction of the compound has sometimes merged them in the root, find sometimes substituted them for it. New definitives have in certain dialects been superadded. . 4, In their present condition the forms of the numerals vary greatly. In general they are polysyllabic compounds, in which the servile defini» tives are distinguishable. The archaic root compound, or compound of root and servile, is dissyllabic in many of the groups, but in some it has contracted into a monosyllable, Its form has suffered great and various modifications, rendering the correct analysis of the names difficult and at times uncertain. i 5. ‘The Semitic system is the last remnant of a dialectic group. It could not have attained its present form without changes having taken place in different dialects which influenced each other. In Africa several of such groups are still preserved, and the mutual borrowings are obvious, Yn addition to the changes thus induced, the Semitic system has, in different eras, affected the African, and some evidence is thus found of the existence of dialects of the former differing im certain names from the resent, E Two well marked stages of the ancient Semitic system are disclosed by the African numerals. ‘The second or later appears to have been an ar- chaic form of.the Babylonian and Himyaritic. ts ygptiges are oe found in Northern Africa including the Nilotie province on the one side and the Senegambian on the other. Some of its forms penetrate deeply into Mid-Africa and are even found to the south of the Delta of the Niger. The first or older stage appears to have been that of the mother formation both of the Semitic ann African systems. Hoth the archaic Semitic and the pre-Himyaric or pre-Babylonic African are referable to it. The influence of the later Himyaric is chiefly marked in the Abyasinian langu- ages. That of Arabic is slight and evidently modern. . ; 6. The detinitives entering into the Africo-Semitie numeral systems ‘are the same that form the pronouns, pretixes and postfixes &c. 1. The sibilant and aspirate, varying to the guttural, dental and even to the liquid, fem, in the Semitico-Libyan branch—with the exception of Borne archaic nfasc. ’ forms—and having, as a root, fem, applications in the Zimbian branch, in which it is one of the principal detinitives. As a 3d pron. this particle occurs in the Semito—Libyan languages in various simple forms, su, shi, sa, i-sa, es,-se, ~is,-s, hu, hi, ha—ha, -ah, —ha, ta, ti, i-ta, -ti, ti-, eth-, th-, -at, ~it, -ith, -ta ; in compound forms, fem. su-at, er-su-a, en-tu-s, en-te-s, n-tho-s &e., mase, su-va, hu-wa, hu-ma, hu-mu, hhu-m, he-m, ho-mu, en-tu-f, n—tho-f, &e., pl. su-nu, su-na, su—n, ho-n, he-n, se-n &c. kc. Dankali, the Galla fam., Bornui, Zimbian and Mala- sy have i-sa, i-si, i-ta, i-zi. Similar forms occur in Berber and Egyp- def. is a demonstrative and dem. element, and a prefix and - tian. The same Pee ile ct 1, Tn Phe SheuitosTabyes it hie anquited : e liquid n, 1, r. In the Semito-Libyan system it has i a plural force. But it also occurs as a prefix and posttix to roots used in the singular number, and it is still current as a sing. def. and demonstrative, Tn those African lon that have the largest basis of archaic Semitic it is a 3d. pron.—Galla, es ages , Malagasy. In Egyptian it is the ‘Sd. person assertive. In the Zimbian system it is always singular, 3 Iu &c.; and in the archaic Semitic system it was prob, sing. ori ‘originally, ETHNOLOGY OF THE INDO-PACIFIC ISLANDs, 43 8. The labial,—a 3rd pron, in Egyptian and Gonga; demons. and @ef. in all the branches, but a pref. or postf. only in Semitic. (save as interrogative and relative). Ithas a mase. and mase. pl. force in Semito« Libyan. It has a pl. force in Zimbian, but it ia also an important sing. def, In Zimbian the liquid @) and the sibilant (1) sounds interchange di- alectically to a considerable extent. Thus the def. pref. si, zi, of Kosah become re and liin Seelraina. In the Semito-Libyan system also 1 and 2 appear to have originally passed into each other. t In the numeral systems the same def. occur, and with similar variations. Semitic has in 1 the forms ha, kha, bi and ta (Him.), im its African forms also hha, a,t; in-2 the forms su, ta, te, ath, ith, js, she, 4#—African the, si, ti; khe, ki, he, hu; in 3 the forms tha, sha, sa, she, shi, si, te, Africa ke, ka, k, sho, so, to; in 5 the forms kha, kho, kh, Africa su, tu, tin, tie, wu, a; in 6 the forms si, shi, she, tsu, sha, ha,— Africa also se, su, zu, %, initial element,—ta, te, sh, t, du, Africa also da, final el.; in 7 the forms sa, shi, she, ha, shu, Afr, ‘also su, se, sha; in 8 tha, sha, she, te, fa, thu, Afr. shu, se, sh, t; in 9 ti, sha, sa, za, sh, ke. 5 in 10 sha, sa, se, ish, s,in Af, si, sli, ha é&e. » In the African systems the same definitive occurs, but less frequently. The variations are similar to the Semitic, buf broad forms are more com- mon than slender ones and contractions and inversions are rarer. Broad forms are also preserved in some of the Semitic dialects, and they appear to have been those of the archaic Semitic and of the primary Semito-Atrican 2. The liquid 1, n, r oceurs in the Semitic 2 in the forms na, ne, in roh, r, Af. nau, nu, ng, li, il, le; in 3, la, lo, le, Af. ra: in 4 ar, ru; an 10 ra, ri, r. In the African langnages it is much more common. 3," The labial occurs as the Ist eletnent in the com. Semitic 1 wa, 8; as the 2d element in the Eeyptian 3; as the 2d element in the Semito-He. 4 (archaically in the Ist also, it is probable); as the 2d element in the Se- mitic 5; as the 2d in the Semitic and Ee. 7 and 8; and as the 1st in the ' Eg. 10. It appears to have been secondary or servile only in the original gystem,—oecurring both prefixually and postfixually. 8. One, The radical detinitives of the unit in Semitic and all the Afri- ean languages are the aspirate &c. and the liquid; and from the forms and distribution itis probable that they were originally variations of each other. The former, in its asp,, gut. and dental forms, ia the Semitic cardinal. In Africa it is also the com. Nilotic and Zimbian root, chiefly in sibilant forms, but dental and guttural also occur. The liquid is preserved in the Arabic ordinal, and it 1s found in Africa in the Agau gr., Bornui and most of the Mid-African and Nigerian ah Suh . The labial pref., full and contracted to vowels, is found in Semitic, Egyp- tian, Berber, Nubian, Bornui and most of the Zimbian tongues. Some of the latter and some of the Niloticand N. Africanlanguages take other pre= fixes, dental, guttural &e. It is probable that in the oldest form of the Semi- ‘to-African systems the prefix or separate def. varied with the substantive. Postfixes also occur in several of the systems. An example of the use of tthe labial both as a pref. and postf. is preserved in tha Sh nte-ta-ma. Jno. The primary root of 2 is the liquid. In the Semitic system it has only the sibilant prefix in 2. But the labial is preserved in higher ~ numbers, 4,7 &c. In African systems it has the full range of pretixes, In the Zimbiun fam. the Jabial is the most com., as it arene to have been in the earlier Semitic. The plural application of the liquid appears 44 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, to have arisen from its use as the numeral. Three, The navw.es for 3 are compounded of those for 1 and 2, and as a ccmequence, were capable of being yaried toa considerable extent. Tlie Semitic svstem had several terms, (A.) A prevalent one or one that characterised the earliest Babylono—Himyarie dialect that influenced the Eevptinn—was the sibilant in its archaie broad form sho, sha with the labial postf, The final n of the labial may be the plural postf. In the Semitic and Eeyptian 8 it is also retained. The Semitic 5 takes final s, The radical elements however were the sibilont, dental &c. with the labial postfix, as this compound is a common representative of 3 in higher numbers 5,6 (8 dual in W. Zimbian) 8 (5,3), 10 (245), It resembles some of the archaic forms of 2 more than any that are extant for 1. Lu-nie, lu-b, su-b, su-we of Galla, Bezharmi &e. are strongly in favour of the sho-m, tu-m, &c. of 3 having been 2, but the puttural forms of the Se- mitic 4 are in favour of its having heen 1. (B.) The second extant Semi-- tio 3 isthe sibilant followed by the liquid ; aud the third, (C,) the sibilant reduplicated, but both appear to be variations of one archaic form, the sibilant passing into the liquid. In Africa these forms are very prevalent, the first element also sometimes changing to the liquid. All these varia- tions are referable to that form of 2 w ich has the sibilant, prefix. Fur is also 2 in different forms in nearly all the systems. The Semitic arba, Nubian arum appear to be contractions of forms similar to ma-ln-d, ma-lu-me &c. 2. The Bab. ru-bu preserves an archaic form similar to the Galla. The Bute d/-ne-b 2 (com. Zimbian :-ri &c.) is an example of a slender form similar to that of the Hebrew ordinal re-vi-/y (re-uSne-b). Thea-j-t, f-tu of Ee, fou-s0 Tibbo, a-fu-r Saumali (fv-n Tumali in, i. e. 4, 5), fu-lu Kalahi, e-fu-tsi, e-fa-r ke. of Malagasy are less con- tracted forms. The Zimbian 4 is afSo the liquid 2 with different prefixes. As ba represents 2 in 7 itis probuble that itdoes so in 4 also, and if so the original of ar-ba must hove been a reduplicated 2 in the form bur-bar, Five is 3 (for 3,2). Semitic has the A form of 3, but with the unit in its guttuial form. Ee. has the dentel form, but without the labiol postf., 10 (2,5) having the prstfix without the root. The full form is found in 6, 7,8 and 10 of various dialects. The com. Gallo-Zimbian t rm is the B form,:and it is also preserved by Semitic in 10 (5 for 2, i. e. twice 5). Sia, The Semitic 6 is 3 (dal) in the C form. Itis found ina few African tongues, variable to the B form, but with these exceptions the African 6 is quinary, 5,1 orl. In Gadsaga and the West Zimbian group 3 has the A form in fi, Seven, 5,2, Semitic has the A form of 3 (for 5) with the sibilant as in 3, and not the guttural asin 5. The Zimbian terms are the same, but = are less contraeted, and- preserve both the A and Dy forms, ight, 5, 3 in Semitic (A) and most of the African languages. Dana- kil, Shihe and several Zimbian dialects have 4 (dual); and a few of the latter have 10 (for 2,10). Nine. Semitic 3 (dual), C form, The African forms are 4,5; 5, 45 and 1 (i. e. 1 from 11), Ten. ~ Semitic 5 (in the B form of 8) for 2,5. Nilotic and N. African generally have the Aform. Many voeabalaries only retaiti 2. The general Asiatic athnities of the mumeréls,the relations which they indicate amongst the different groups of the Bemite-African alliance, and the light thrown by them on ita history, will be considered at the conclu- sion of the glossarial illustrations. ETANOLoGY ‘OF THE INDO-PACIPIC ISLANDS. ~ 1” PART I, CHAP. Vf, sc. 5 (Continued). COMPARATIVE VOCABULARY OF HORPA: a, Bhotian. YE. Air, pu ryo, Thochw mo-zyu, Bhotian lung ma. 2, Ant s-khro, Thocha tu khra, Manysk bara h, Gyarung Xo-rok, Bhot. g-rog ma. 3. Arvow. Lda, B. m dah, Mi T. jeh, Tokpa m Ja, ® 6. Boat gra, B g ra. . Bone. réra, ‘TY ri-pat. The others are broad, rus,ru,rhu. Lea gian re, ray du, ro, Drag, Cite &e. 12 Day. nye le,G@ nye, Pakpa nyen di, B, nyin mo, nyi mo. 14. Ear. nyo, T. nukh, B, 7 oa, Chepang no, Kar nho. 5. Barth. bcha, Bisa. 16. Eog sganga, 8. s gonga. 17 £Llepiant. \amo cones G, lang. Seb ett, Sokpa lhabo-che, B. g« Jang eh (Ch. ) m. 19, Father. apa, M. Takps apa, B. pits. (20. Fire uias’, B. me, Aka w mah (Ch.) 21. Fish. hya, B nya, Lhop neya. 22 Flower. meto, “i. Takpa mento, B metag. 23. Foot ka, B. rekiing pa, Manip a kha, Yoma ka-kong &e. (Ch,) 25. Heir spu, B. spo, Takpa po, M, moi, Dhimal muni fu (Ch) 26. Hand tha, B. lag-pd, Takpa la, MIs p che’ Lhop. la-pa, G. ta yak, Naga yak. ' 27. Head, gho, B. mgo, Gyar,, Naga fa ko; ‘Takpa gok fi, Manip: ‘ole 28. Hog er M. wah, Takpa nbn: B. phag, phak. 29 Horn. k vuw-bo, B. ra, T. ak, M ro bu, Takpa ru ba,G ta ru; Garo ko reng Sunw gts re &e. 32, Tron. chu, G. sho m,'T. sor mo, Sokpa tha mar, B /chage, chhya,: Yeu. ta p, tip, Mong, to m ur, te m-or, thu mar, Soaheli chu-m ar, Korea eoy, Chin chiar, chi. 33. Leaf. ba la’, Tokna 5 la p, B. lo ma, 36, Monkey 20m de’ B en. tyo. Gr. ti. 37. Moon. s \ik- no, G. tsi Ve, chi te’, T. le’, M the’, B. els va, dae wa. The Worna form appears to preserve an archaie cons final softened inG, T- and M_ intothe abropt accent Comp. Chia. ngiat, Samoiede j-ret, irda, Chukchi tra lnk, Pagani lago, Meri leka. 38. Mother ama. B.,M.T. ama. 39. Mountain, r-rep, BY, Tom (Chy 40. Mouth ya. B®, T. kha, G ti khe, M- ye-ba, Ch. (Ch,) 42, Name smen, B,.G,M mine; Takp- myeng. 44° Oil war.nak, B A-bra mar, T. kya mar. 45, Soll chha’, B chha. 51. Snake phri, G Aha bri, T. hri ge, Bs bral, 52, Star sqre, B.s kar ma, M, kra’, Burm. kre- 54, Suu yna, B nyi-ma. 55. Tiger s tak, B. s tag, tak. 56, Tooth. «yo, B Rn, 60. Yam, zo, B, do va, \homea, 2 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. ‘ db. Chinese. 4, Bird gyo,C. chio, 5. Blood. wye, wh, C. aye. (B., T., and M, have the broad form thak, sah &c.) : 10, Cow ngav-meh, © ngoi, also Lau, Mon,; G. nye nyi, Burm ngi, 18. Eye me, Chi mek, Mon mot, Li7, 20,23, 25, 39 and 40 are Bhout.-Chinese, making the entire nume ber of Chinese words 10.) ~ c. Non- Photian. (9, 11, 13 and 34 are Bhotian in row ), . 9 Cat. cho la’ M. machen ‘The la" ia found in T. lo chi, B, byte la, N. Tangkal la me. ‘he cho, chev appears to be a broad furm of the Thocho chi, Bhor, si of simi ‘The Maring tong, Kivibu tong-kan, ie the same root, Probably +! khar Sindh, Hind, khal, Ostiak kor parga, Lesg kuli, Malay &, wlit. 40. “ky ko, 53. Stone r game, The gatt. root is common, but with | as the finaL Fin haw kiw, 57 ‘Tree. nah, Mon 4a non, Aino nyh, Pashtu ona, "58. Village rhava. 59, Water. hea’, Chepang lang, Navkowry rak, Newar Yenis, la, dok, ur, ul, Turk, yrratseh, ir-mak, COMPARATIVE VOCABULARY OF THOCHU.s - a. Bhotian, A . Air. mo zyu, H. pu ryo, B. lung ma; Manipuri ma au &e, » dAnf, STHNOLOGY OF THE INDO-PACIFIC ISLANDS. 3 3, Arrow. : 5. isluod. oa’, M, aha’, B. thak, (Tungas. shok sho, eug al, sho.ma), Bingfu say Bova zs, asim Lembo jek shuk-pa(Coup. Water dok Ye pie. dak. tak &ce. Kawbojao group, 7. Bone «pat. B, roe-pa, Hor, rera. The slender form is not found io wner PL—U. vocanularies, aave Lepcha @ che tf, bu: it ie See moiede ly, Usrian ly, Caucasian li ka, and it ales occurs in Anunevia, ri Tarawa, lid Boul. ru’ Solor, Dhe double poacf resembles shat of ibe Galla la fa ti ‘The d- uble Horpa re-ra resemble the Caucavian In-], and Koreans lu-lor, the original of che Jast being probably the T.—U. long, (Ts. 9. Cat. lo-chi («ee Marpa). 12. Crow. nyag-wo, Bh. sp. ab-lak, Champhung chag-hak, Rae khviug tchas-in, Mishoitssk-la, Singphu takha, Garo dakha, Bodo doukha, tauka. : 13, Day kbwa’, Bh. khyi. The slender form is the most common in the south. Ihe uw or w is preserved in the Burman khwe; Singph. kwi. Whe Cauc, gwai resem'les the Tbochu torm. 15. Earth. zi-p, Bh. sa, G. se’; Jap zi, tsi, Ch, ti &c, Sam, tociia, ja, Turk sir. : 20. Fire. me’, Bh ma, me, . 25, Hair, grong, Bb. kra 26. Hog. pi, Bh. ; hug, Chepang piak. wil 7 eal rak, Bh. ra, Chauglo wa-rong, Uraon ma-rag, Ultr, ral Ye 3i. House, ki’, Bh. khyim, Sunw. khi, Kar. hi, 34, Light. vik, Bh, hwe. 42 Name. r-ma’, bh. ming, Naga-Manip. ming, mang, man. 43. Night. a-sia, Bu. m-tshan-me, 47. Road g-ri’, Gyar. tri, Hh lam, M. ra’. 48. Salt. che’, G. chhe, M. che’, Bh tstia, chha, Horp. chha’, Take tsa : P49. Skin, ra-pi, Bh. pag-pa. (The ra may be the Horpa g-la, M. g-'a'), Singpho, Karen phi, Marmi di-bhi, Mozome Angami bi-kbar, ». Boe bi-gur, Garo bi-gil, Yama moe-pik, Ch, phi. . 61. Snake. bri-yi, H,, G. also slender’ Bh, a-brul 66. Tooth. swe’, G. ti-swe, Bh so. Burm., Murmi swa, Ch, ch’hui khi, Ugr sho-pan, : 67. Tree. gwo-z0si, Gr. shi, M, sa-poh, Bh. /-jon-shing, shin-dong, Ch. shi. 69, Water. chah. Bh.chhu. | b. Chasese, 18. Eye. kan, Ch. gan, Drav, kan. ; v: 22. Fowist lame’, Chia, low, Ker. kha, . 35, Man. mi’, Anam ngoe, Ch. lang, male, ang, nan, husband (See 10, 13, 15, 44, 55>) » ce, Non-Bhotian. 4. Bird mar-wo, Drav. par-va, Angami Naga para, Samoiede miari-nane, Osetic mar-g, Pashtu mir-ge, Hind. mur-gh. 6. Boat. pbyu, 2 p-la Tibeto-Ultr. Car Nicobar la. ap. . 10. Cow gwa’, Pashtu ghwa, Lau nguw, Mon nua, Lhop. pgo, Chin, ngui, go &t, | 4e ETHNOLOGY OF THE: INDO-PACIFIC: ISLANDS: ® 12. Day. styak-lo, Lepcha sak-ni, (sun, sa-chak, Limbu, sky, dant sak-pa,) * 14. Ear. nukh, Karen neko, N@Tangh nakor,-Mishmi nakru, Limbu neko, Ch Tibeto-Ult, na. [Fin, face, ak, nyako] fe y 16. Egq. kiwost @ ki-wost “egg-of-bird”’, bird Singh» wu, Naga vo, egg Newar khyen, Ugrian ai, Siem khoi, . 19. Father ai Uenan ai, * . 21. Fish izhn’, Yenis. visya, isse, Aino zis-!, Ugr zen, Mong. sa- ga-enn ; Nias, Paser isa’, Philip isda, sida, sira. Y 22, Flower. lom-pe’, Cheta, bua, Kar, pha, Garo, Yume par &e. nie Fovt, jako, Turk ajuk, Garo foot jak, Kasia ka-jat, Simang cliat, ‘25 ‘94, Goat. taah, M. tsah (see Horp.) 25a. Hair, hom-pa, Ugr. yop, Samviede hopt, Turk. mui, Ultr. som, ~ Newar song, Lepcha achom. ; 956, Hair. kachu, Sok. kechi-ge, Koriak ketsohu-gui, Sanskrit” kesha; Naga kocho, Kar: khosu. : - 96. Hand. jipe’, Magar bat piak, Turkish finger shar-bag, Manip. eD. pang, pan, ban. ‘ 27, Head. kapat; Comp hair Sam, bopt, Ugr. upat, opta &e , head Indo-Eur. kopf, hotd, caput, kapala &e. 30. Horse. ro’, G,M, bo-ro’, H. rlii (Ugr. lo, see Horpa ) 32, Jron. sormo, Garo shiur, Bod. shor, chur, Dhim chir; Korea eny, Tangus solu, zhila &c., Sam. sommaya, Suaheli chumar (see Hors* a. v 3 8, Leaf. thrompi, ? Tib.loma, Ugr. lop, lopta &e. way 36, Monkey, wui-si, Sokpa me-chu, Singpho we (si is man in Horpi v-zil, and ti is monkey in Gyar.) Comp. Naga si-mui, mui-nak, Abor, aie heh &e,. 9 “87. Joon. chho’, Sokpa sara, Yenis. chaip, Manip. kachang, Milch ga-tchang, Garo ja (“star” Chin. ch’he, Karen sha &c.) ; 38. Mother: ou, Osm. Turk. ava, Ugr. awai, Sam. eo, Manip. D. noa, onu &e, ~ “"”A 39, Mountain, spya’, (? s-pya), Sokp. taya, Turk tapa, tuba, nba, oba, tope. Manip daplung, bom, : 40, Mouth; daukh, Yenis. b-yuk-kon, Sam. hek, Cauc. leku, Kamsch, shakesha, tsch-na, Koriak shek-shen, Ugr. sbus, &c, Fin su,” gon &e, (hin. sui &c 41, Mosquilo, beup. (‘The labial root, single or reduplicated, is mosquito, fly, bee &c. in many languages. 44, Oil, ching-yu, Ch yu. 45, Plaintain, sarmi. 46, River, cha-bra’, H. bra, Kiranti, Sam, tscha-ga, ja-cha, cha- . wa; Tung. amar; bera &c; Sambawa brang. - se 50, Sky, mahto, M. ma’, G /u-mon, teu-meun’ ™ 62, Star, ghada, Mishmi-kadang (? Bhot. pe. kar.) 68, Stone, ghol-opi; Sokpa chhilo, Takpa gor, Many. wobi. 54, Sun, mun; Sky G. mon, Gurung mon, Singpho &e mn, Mirl do-mur; Fin poi-wa, pew, Sad. fi . : 55, Tiger, kho, (h. hu, Gyami khu, G. kong, Dltr. kya, &o. 58, Village, wekha, G. wokhyu. 60, Yam, jyab, ETHNOLOGY OF THE INDO-PACIFIG ISLANDS. 4 GOMPARATIVE VOCABULARY OF GYARUNG, - a. Bhotian. 2 Ant, Bhot. grogma, Gyar. horok ; Takpa rok-po. 4. Bird, B. byu, G. pyé&pye, Takpa ge st 6. Boat, B.gru, G. bru. 7. Bone, B. ruspa, G. sya-rhu, (Manip. sa-ru). rl, ig B. ablak, G. she 12; ay, B.nyinmo, nyimo, G. nye, pish-ne [pish Chin | 13. Dog, B. khyi, &. khi (Oh. bia. cp ] 14. ur, roa, na, G. firne (Angami Nag. anye) [Ch. nai, li]. 15. Earth, sa, Gr. se’. i ; en BiB] 17. Elephant, B. sapeghe, G. lang-ehhen [Ch. chhiang]. 18. Eye, B. mig, mik, (Chinese mok], G tai-myek, tam-myek [the G, form is also Burm. | 2 * 19. Father, B, pha, pala, G. fape Burm. phae [Ch. pe, be}. 90. Fire, B. ma, me, G. timi [Chin. we, Aino abe, Fin, bi, com], 91, fish, B. nya, nga, G. chu-ngyo- . 26. Hand, B. lagpa, G. tayak, Naga dak, yak. 27. Head, mgo, go, G. lako, 29°" Horn, B. ra, G. taru, 31. Hotwxe, B. khyim, G chbhem. 35: Man, B. mi, G. tirmi [Fin mis, Cauc. mi, me, ma, Galla mi], 36. Monkey, B. sprebu, G. she-pri. 37. Moon, B. zlava, dawa, G. tsile, chileh, M. lheh. 38. Mother. B. ama, G, tomo No Spas &e]. 40. Mouth, B. kha, G: tikhe [Ch. khau, Yenis, ko, Semit, kho &c]. 42. Name, wing, G. tirmitig. an ees B. lam, lani, G, tri, Thocha grih, (Karen Ale, Khari aga mdi). * 48. Salt, B. tsha, chha, G. chhe. (Sam. si, sak, Ugr. sow, sal]. 61. Snake, B. sbrul, deu, (M. bru, Takpa mrui), G. kdabri H. phri, T. brig#. ._ 63, Stone, B. rdo, do; G. ragu, (Ultraind. lung, long &c). 54. Sun, B.nyima, G. Fini, 56. Tooth, B. so, G. tiswe (Burm). 57. Tree, B. tjon-shing, G. shi, Chinese shi, chang &c. Kamch, 69, Water, B. chhu, G. tichi, Takpa shhi, (Chin, chui]. 60, Yam, B, thoma, G, seten, b, Chinese. B, 6, 10, 18, 15, 17, 18, 20, 26, 47, 52, 55, c. Non-Bhotian. ° 1. Air, fali; Burman gr. Zali, kali, khli, le, Turk ¢yel, Ugr. il, lil, ima &e, {Wind” Turk. il, dil, yil, sel, Ugr. tal, éul, fil, tauli &e.) 3,° Arrow, kipi; Garo phi, as 5. Blood, tashi, Gyami sye, Nag. New. si; Chin, chiue,[Tibthak], = ®. Cat, tarhu (? a misprint), 10. Cow, nye-nye’, Gyami neu, nyeu, Ch. ngui, (Turk, ona, ina], 16. g, kitan, Gyami chi-tun, Chin, tan, Anam, N-Ultr. ting. 29, , tau-den; Kas, sin-tin. : 23, oot. tami, Takpa femi, Changlo bi, Naga uphi, Manip, chapi, 5 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, be wang, Milch. bang, Nipal Dial. pali, bele &e., Agones, val, wai &c., alag. fe, pe, Indo-Eur, pad, pes, foot &c., Ugrian pilga, Sam. tovol, tobo, 24, Goat, kaso (see Manyak). - 25. Hair, tarni, alae! inang, Agon. oli, nirl a. 26. Hag. ki, pets Naga ak, Gyami tiko, Ch, ti, chi. 33. Leuf, tai mek 39. Mountatu, tavet. Monegal dybe, Turk, taw &c (see Manyak), 43. Night, todi, Angami Naga ti z!, asang dai, Turk. tin mae al, auni &c, 46 River, tichi, Takpa chhi (see Pater). if Skin, fidri, Dhimal dole, Torkish diri, dari, tire &c., Ugr. toul &e. a Sky, tumoe, ten mean, T’, mahto, M. mab, Borm, group mo, mo, mi &c 52 = Star, tsini, Nave le tsi, pe-ti &c, Chinese sim, Yuma er kasi, a ahi &c. Torkiah syltis, ildie &e Uer sili &c. 55. Tiger, kong, T. khoh, Gyami kha, Nag, tokbn, kho, Chin. hu, ho. 58, Fillage, wo khyu, tu khya, COMPARATIVE VOCABULARY OF MANYAK: r a. Bhotian. The Manyak forms, it will be remarked, generally resemble the more slender of the Ultraindian and not the Bhotian, 2, Ant, B. grogma [G. korok] M. barch, (Abor marang, Magar mahr, Aka farak, Manip. group laugza, chaling, miling &c. Khomi palong, Silong hedam). 3. Arrow, B. miah, da, M. ma (Burm., Magar mya, from ymra Burm., Kiranti me). 4, Bird, B. chya, M. ha (Nag. ou-ha, au-ha &e, Mrung ho, haw), 5. Blood, B. thok, M. shah (Thochu suh, Horpasye, seh, Gyami aye, Gyarung tashi) 6. Boat, gru, M. gu (Gyarung bru, Ultr, rua, rung &e ) 7. Bone, B.s. ruko, M, ruklu. 10. Cow, B. ba; M. wo-mi (Anam bo, Siam woa. » In Manipuri, Yuma Xe. woi is used generally with names of quadrupeds, as mi is with those of the cow and buffaloe in Manyak. The Anam and Lau names of the cow appear to have been derived trom the Tibeto-Ugrian ba, mus—whence bos—through the Mauyak form). 14, Ear, B. rnawa, na, M, nepi, (Naga, Burm. &c.) BY Father, B. pha, M. apa com. . Fire, B, mg, me, M. same’ nee mi &e.] 22, Flower, B. metog, mentok, Takpa chin 26, Hand, B. lag-pa, lango, M. lapec heh, Tekpala. 28. Hog, 'B. phag, phak, Horpa vah, M. wah, (Angami Nag. the- . Horn, B. va, rajo, Horpa krumba, M. robu, Takpa ruba. 31. House, B. nang, M. nyeb, (Deor, Cb. nya, Bodo noo, Naga oe 4, Iron, B. chhya, M. shi. 33. Leaf, B. loma, Horpa balab, Dhimal lhava; Takpa blap, M. nipcheh, Naga nyap, ponye, % ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 6 84, Light, B. hod, hwe, eu; Thochu vik, M, wub, Takpa wot. 37. Moon, B. zlava, M. lheh (Naga le), , 38. Mother, B. M. ama. (com,) | 42, Neme, B.M ming, 47. Road, B, lam, lant, M. rah. 48. Salt, B. tsha, chha, Thochu, Gyar. Many. cheh, 51. Snake, j= 8 sbrul, M. bra. Star, B skarma, karma, M. krah, Horpa sgre. Sun, B. M. nyi-ma, i fae 5. Chinese. e. Non- Bhotian, 1, Air, merdah, Naga ra, rang &c.: Dophla dori, Burm, li, le, la &e. Drav. la. ‘The broad forma ot Mois Negis Dat: and Aino. The slen« der form of Gyarung and the Burmese group is Turkish, Yukahiri &e, The Tibetan form of the common root is distinct, rlungma, lhakpa, the latter being connected with the Naga-Manyak form. 8. Buffaloe, ding-mi, wo-mi, ‘‘cow’; Kar. pi ‘‘cow’’, Dhim. pia, Newar &c, me “buftaloe’’. Dips is peculiar, unless it be a misprint ce bing. : ea wage . Cat, macheu; Dophla ac aga mochi, Bodo mouji, Mong. michoi, Korea koi, Ugr. mishok ce 2s i $ 1l. Crow, kali; Horpa kale, Sokpa khere ; Champh. khala, Mish- mi tsa-kla, Garo koura ( Asam , Beng.) The term is Scythic, Indo- Eur., Somitico-Afr., and Asonesian, 12, Day, nashcbah; Bodo shyan, Garo, Naga, Yuma san, Tiberk. ghangnia [‘‘Sun” in other dialects] 13, Dog, kshah, Horpa f4atuk, Nago fasu, az &c., Nipal dialects hushu, kochu &c. [Yenisei. il-tscha, Kamch ko-sha, Root N. E. Asian, Scythic, Caue., Indo-Eur., Semitivo-Atr,, Ason | 5. Earth, mali, m'i; Naga ali, Manip malai &c., Kyo ni, Burm, mre, Mish tari (‘Turk. yir, er, Korea chili. } 19. qq, racha (? che “bird”, ‘lib. chya), Korea ar, al. 18. Eye, mni, Mru. min (? from Tib. mik, like the Dhim mi, or frow the Mong. nidu by contraction, Kamch nanin, Jap. mey, mamige ='ib., Korea nun, Mong. nudun, nidu &e, 21. Fisk, yu, Gyami yue, Chin, bu. 93. Foot, lin-cheh (see “band’’), Garo, chap-lap. : 24, Goat, tsah, T. tsab, H. chhe, kuso, Alor sa-ben, Dhim. eecha, Anain, Kamboj. sha-bain, Kashm. shawal, Burm, sheik ; si. 25 Hair, mai, Dhim mui tu, Morpa spa, Takpa bo, Torkish mui. 27. Head, wuli, Ditim. puring, Khari Noga felim, Manip. iu, Ahom ru, Yukahiri ulu, monoli &e, Ugr. yor, yir &c, 30, Horse, boruh, broh, G. boroh, T. roh, H. rbi, ryi, Mong. mo-= ri &c, Tung. moron &e , Korea mol, 35, Man, chhoh,-Changlo songo, Naga saun-yak, mesung; Ugr. ° chu, choi, Afho chogu. . : 36, Monkey, miyahah, ? Dhim nhoya. 39. Mouniain, mbi, Khari Naga apih (T. spya’, Sokpa tava, Q@. tavet, Mong. dybe. ‘Turk. taw, tau, uba, Yuk h. pea, Sam. bija 40. Mouth, yeba, Sokpa ama, Mong. ama, aman, Tung. amga, Ugr,- um, om, im, wom &c, Naga amu, ¢abang, ¢ebaun, rs 4, 10, 20, 21,-47, 55, v | ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 41. Mosquito, bimo, T. beup, ? Angami Nag. viru, 43. Night, kweakah, Kir. khakwe, Gyami khe-lo. 44, Oil, ichira, itira, Dhim. chuiti. 46. River; dyah, Bodo doi (see Waiter), = 49. Skin, grah, Horpa gla, Thochu rapi, Soka saru, Mozomi Ang. Nag. bikhar, Bodo digur. 50. Sky, mah, Thochu mah-fo, Gyar tu-mon,’ Turk. awa, 53. Stone, wobi, T. ghol-opi (Takpa gorr ; gol &c. is Tatar, Korea, Kameh , Yakahiri and Ugrian in different forms; pi, pe &e. is Samoiede, and Aino), 65, Tiger, lephe. . 66. Tooth, phwih (? Gyar. tiswe, Thochusweh), Burm, thwa, Tak~- pa woh, Kami afba, [Tungus. wei-che, Jap. fa, Ugr. pui, pu &c. he. 57 Tree, sapoh, Nag peh, pan, Burm, apen &c., | Turk. twos, Tung.mo, Kamch uo, Sam. po, pu, pe, Ugr, pu, eu, fa &c. 68. Village, hu, Takpa yu 59. Water, dyah, [? G. ti-chi, Takpa shhi, B, chhu Chin, sui, cheu &v., Bodo doi, Yuma tui, Nag, tu, ti, si &c,, Sam. tui, Tartar su, zu, dsu, she &c.] 60, Yam, zgwah, : roy ETHNOLOGY OF THE INDO-PACIFIC ISUANDS. a Sec, 6, THE GLOSSARIAL CONNECTION BETWEEN ULTRAINDO- GANGETIC AND TIBETAN. 1. General remarks on the Bhotian affinities of the . Gangetic and Ultraindian languages. The Ultraindo-Gangetic vovabularies present two classea of Bhotian affinities, each of which haa two branches, . The Ist class consists of words, or forma of words,.immediately deriv- ed from Bhotian, and at least two branches may be distinguished, viz, vo- aables derived from the modern Bhotian, and vocables derived from the ancient or written Bhotian. Ifsas is probable, the etrong phonology atill prevailed throughout the Bholian province atthe time of the first great irruptions into India, it is not necessary to assign an older date to the yovables of the second branch than the earlier centuries of the Chris. tianera. Indeed they may have continved to be imported to a mach more recant period, and may possibly de still received into some of the Hima. layan dialects ifthe old phonology beretained by any of the Ram pas who migrate to this side of the snows atthe presentday. Whether Western Tibet directly sent vocables to the southward before \he age ef the Himalayan conquests is a question thal cannot yel be answered, The 2d and most important class of Bhotisn affinities are those which exist ia the Ultraindo-Gangelic vocabularies not becanse they were received from Bhotian, but because the Northern linguistic itael of the tribes which use them was closely allied to the Bhotian, both having for basi4# a Common formation. They may be nowconsidered as Sifan, These archaic ‘Tibeto Ultraindian or Sifan vocables possess two forma, one characteristic of that modification of the formation which it had when it first came in contact with the prior Mon-Anam formation of Ultraindia, and the other peculiar to the Borman branch, which appears to have spread to the southward and westward at a more recent period, after having long remainid secluded and comparatively pure in the North Eastern part of Ultraindia or the adjacent Sifan mountains. The older diffused forms are generally full and dissyllabic, and the first syllable is frequently a definitive prefix. The later forms are remarkably curt, and in this respect contrast not only with the older, but with the Bhotian, the latter having prefixed consonants and frequently adding & postfix to mo- nosyllabic rootey Iwill proceed to consider each of these varieties of the Bhotian affinities more particolarly. The absence in Indian history of any notice of the modern irroption of the Bhotians into the Himalayas and the plain of the Ganges, of which positive but faint historical evidence exists in Chinese books, exhibits its partial and untrustworthy character in a strong light. It has preserved no distinct secord of an event of 80 much importance that it pave to the Himalayas a new people and new dialects, subverted the ancient dynasties of the plain,—Arian, Draviro-Ultraindian or Arianised,—and Jed to the eatablishment of a Ti- betan dominion, which lasted so many. centuries in Bengalasto siirct ~ not only the ruder lanvvages nesrthe mountains, hut, in a very slight de. gree, Bengali itself. Ifa revolotion of this kind, that began some c+ntue vies later than the commencementof our own and of the prevalent Ja. 16 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, dian eras, has been suppressed, how much caution is needed in making any historical nse of the Hindu chronicles. The degree in which the proper Bhotian glossarial forms of the Vibeto Uliraindjan formation have been diffased tothe southward of the Himalayas sufficiently appears from other Sections, The ancient or written forme are frequently found in Lhopa and Serpa and sometimes in the less Bhotised Janguages of the Himalayas. The modero forms of Lhasea and Digatchi bave been pare tially spread by Bhotians among the Himalayan languages, bot very few examples are found io the Middle and South Gangetic or in the Ultrsindian Janguages. ‘Tbe ancient Bhotian forms have been sparingly diffused from Bhutan among th» middle Gangetic and the adjacent North Ultraindian languages... The prevalent Ultraindian forms of the Tibeto Ultraindian roots sometimes gree with the Bhotian but mora frequently differ from them, and most of those which agree with it are too widely diffused, and are, in many cases, of too essential a character and too intimately blended with the Ultraindian glossarial systems, to have been recent derivations from Tibet or Bhutan, Many have now been found in Sifan vocabularies also, and it t¢ clear that such forms were car. ried southward by the oldest migrations from Tibet, which must have long preceded the Bhotian irroptions of our era. Every great glossarial formation exhjbits a propurtion of roots which preserve an identity in form in separate provinces and after an extent of diffosion which it muat have taken many thousands of years to effect, The Asonesian languages, in their archaic affinities with African, N. Asiatic, N. E. Asiatic and other remote languages, afford striking illustrations of this, The glossaries of the Ultraindian and the connected lodian languages are exceedingly mixed, a necessary result of the single introsive Ultras indian race having partially blended at leakt three distinet lingnistie formations, the archaic Draviro-Australian with ites modern N, E, Dravi- rian branch, and the imported Mon-Anam, and Tibeto Burman, Every attempt at an exact separation of the roots belonging to these several formations must prove to a certain extent a failure, because all had are chaic affinities. Thus the Dravirian bad Tibetan and Chinese offinities, and the Mon-Anoam languages must have had archaic connections with the adjacent languages ofthe Tibetan family before either of these branches of the Vhino-Tibetan or Himalaic stem was carried over the Mountains into Ultrsindia and India. But it ie possible to make a roogh approximation to such a separation, owing tothe circnmsiance of the Tibetan vocabolaries still extantin Vibet and the Dravrian vocabnlaries of Southern India having preserved certain portions of the ancient gloa. garies of two of the formations comparatively free from Ultraindian intermixtare. In the cave of Tibet it ig not probable that its archaie vocabularies have been affected by the non-Sanekrit languages oo thie side of the Himalayas, and although Southern India ismuch more ex- posed, the general effect of all the ethnic evidence is against the Uliua- indien tribes and languages having influenced the peninsula beyowd the Vindyae to any notable extent. Elaving already partielly traced the Dravirian voesbles in Ulttaindia, the first step towards ascertaining tha probable extent aod diffusion of the Mun-Anam glossarial remnenta will be to separate from the Gangeto-Ultraindian vocalilaries those words which clearly or probably belong to the Tibeta-Burman formation. The Bhotian affinities of the various vocabularics of the Borman fami- * BTENOLOGY OF THE INDO-PACIFIC ISLANDS, at ly and of Ultvaindia generally, are, with a very few exceptions, archaic. Takeo with the large amount of disagreement that remains amonget” these vocabularies, after excluding the Ultraindian words of probable Dravirian, Chinese, and Asonesian origin, they prove that the ae Ultraindian formation embraced several jangoages-possessed of vocabula~ riea that differed considerably. This indeed might have been anticipat~ ed. Atthe remota era when Tibeto-Chinese or Scythoid tribes began to descend into Ultraindia, it is not at all probable that any civilisation prevailed immediately to the north of the Himalayas sufficiently advanc~- ed to bave establiehed one nation and one Janguage over a region go cold, arid and mountainous, Tht western progress of the Chinese may in time bring about.such an event, but it may be considered as certain that it has never hitherto existed. At present there are several dialecta in Tibet itself, and, according to Chinese anthorities, eeveral aleoin the adjacent provinces now embraced in western China. Where there are now five distinct vocebularies there may have been more than double that number when the tribes of this region first began thefr movement into. Ultraindia. In estimating the amount of the archaic glossarial #ffi- nity between the Ultraindian and the Tibetan languages, we must allow something for the words that may have been conveyed by Bhotians into Ultraindia since they beeame so civilised as to carry on a traffic with the upper tribes of the eae, such as the Mishmi. 2, Zhe General Connection between the Gangeto-Ultraindian and the Sifan Languages as Dialects of the same variety of Tibetan. With our present imperfect information respecting the East Tibetan and Gangeto-Uliraindian languages, a detailed grammatical compari- ‘son is impossible, In Secs, |, 2, 3, 4, 5 and 6 of chap, 1V. I have ebewn that the North Ultraindian and the Gangetic languayes are inti- mately connected in structure, so far as their strocture is known, ae well as in pronoons, numerals and ether particles ; and that they all belong to the Tibeto Barman fawily, although a variable but slight are ehaic Indian or Dravirian ingredient is found in most of them, and some have been influenced by tha Mun-Anam formation. The preseoce of a large Bhotian element was indicated, but many common traits were found to connect the Uliraindian with the Gangetic languages which could not be referred to Bhotian, and which pointed ata derivation of the primary Ultraindo-Gangetic variety of Tibetan not from Bhotian but from seme archaic eastern branch of the Tibetan formation, Referring to chap: 1V. for an examination of the structure of the Gangeto. Ultraindian dialects, it is only necessary to add here that the Sifan languages that have since been brought to light by Mr. Hodg. gon prove te be representagives of that. Eastern branch of Tibetan from which the Ultraindo.Gangetic dialects were mainly derived. At present very little is known of their grammare, and it would be premature to corclude that any of these Sifan languages was the immediate parent of all the allied Gangeto-Ultraindian, Ic is clear that the latier are primas rily and principally dialect# of the Sifan and not of the Bhotian branch of ‘Yibetan, bot there most have been a great lapse of time since the Sifan . tribes first began to cross the mounlaing; dialects may have existed then in Eastern Tibet which are @t now; and the surviving divlects have pros bably been modified by internal change, by movements amongst-the nr» 12. ETINOLOGY OF THE INDO-PACIFIC ISLANDS. tive hordes and by the inflaence of the surrounding Bhotians, Tartars and Chinese, For the present we must be satisfied with the conclosion that the Sifan and the Ultraindo-Gangetie dialects are al) sub-varieties of one East Tibetan dialect, but that many of the Ultraindo-Gangetic have been moM@or less modified by the infloence of Bhotiun since they first spread into the basina of the [rawadi,the Brahmaputra, the Ganges and the ups per Sutlej. Some light will bethrown on the sveceysive phases and mi- grations of the southern dialects by our glossarial comparisons, The little that is known of the etrocture of the Sifan dislecta has been given inthe preceding Sections All the traits that diatinguieh them from Bhotiap are found in the Ultraindo-Gan. getic languages. The principal is the large use of vocalised pre. fixes, ‘The identity in these prefixes, in the non-Bhotian pronouns and in the perticlea generally, belongs to the glossarial compori. gon, The somewhat more Scythic character of the phonology also con. necte the Sifan, or at least the Gyarung, with the more harinonic of the southern languages, as the Bodo and Dhimal, ‘The postfixoal aggluti. nation of the pronouns is a Scythic trait, which muat ‘now be considered as of Sifan, acd not of Dravirian, origin, in the Dhimal and Naga dialects in which it hae been found, The existence of a dual or inclosive plural of the Ist pron. ia Manyak and Thochu connects the Bifan idiom with the Soythic on the one side and the Draviru-Australian on the other, Mr. Hodgson has found it in the Himalayan dialects of Kuswar, Hayu and Kiranti (Jonrn, As, Soc, Beng, 1853 p, 62), ao thatit may prove to have been earried by the Sifan tribes to the southward. The Ho, San. thal and Uraon forms to which Mr. Hodgsdn also refers, are Dravirian, and not Manyak, Thochn or Seythic. [n referring to chap, TY, it will be born in mind that the Sifan vocabu- Jaries have now greatly increased the ascertained ibetan element in Gangeto-Ultraindian, and, as a consequence, diminished what I had considered the Dravirian. The phonology and pronoans may now be held as mainly Sifan, although some Dravirian ingredients are still res cognizable (see chap, V. sec. 11). ~ he general ethnological inferences may be briefly adverted to ia this place. i | The first conclusion to be drawn from the ascertained facts is that during an era subsequent to that in which the Mon-Anam formation became predominant in Uliraindia, Tibetana crossed the Himalayas in large numbers and acquired an ethnic position and influence in Northern Ultraindia and the Gangetic basin. The Tibetan language in ite inleprily was traneported to this side of the snows, and, through the dispersion of the Tibetan tribes, gave rise to new dialects, and deeply ard variovely affected the prior Gangetc-UJtraindian languages, In many of the existing cit-Himalayan dialects we find Tibe®an pronouns, perticles and ideologic usages, while the miscellaneous Tibetan vocables form an ine gredient, generally very considerable, in the cloxsaries of all the U)train- do-Gangetic tongues, Although no single mixed vocabulary appears to be more than one half Tibetan, it is probable that the greater part of the Tibetan gloseary was at one time current in the southern dialects or was interfosed amongst the different native langusges which came pnder their influence. G | An influence so great, and embracing” so many languages from the dd ETUNOLOGY OF THE IN00 PACIFIC ISLANDS. 13 Milchanang to the Singpho and Lau, could not bave been exerted by a Tibetan tribe which was confined to themountains, like the present western Bhatiangand eastern Liopas. Tibetans or Tibetanised Himala- yans most have descenged into Ultraindia or India and acquired a puli- tical and social predominance over a considerable area. ~ The second inference is that the diffusion of Tibetan elements on thia side of the Himalayas bas not been caused byaeingle movement of Tibetan tribe confined to one era. These Bhotian irroptione into the enh Himalayas and India which may be regarded as historical bave pro- diced a comparatively emallinfluence. Although it has been continued until the present time, or for ahout twelve centories at Jeast, owing to the permanent advanee of the Bhotian ethnic frontier into the sub- Himalayas, it hae failed to assimilate the conterminous langnages of that tract. From the Tibarkhad to the Abor a nearly uninterrupted band of languages is preserved, which retain non-Bhotian forme of pronouns and particles, and two thirds of the vecables of which appear to be non Bho- tran. Even the most eastern of these languages, aa the Daphia and Abor, which are spoken by bighly Bhotoid tribes, have a very considerable ba- sia of non-Bhotian trails in phonology, glossary and ideology. In the Gangetic plain the influence of Bhotian has been still leas, It ie obvious, fram these facts, that the pure Bhotian tribes and languages of the sub: Himalayas have always been separated from thoseof the plain by a barrier of only partially Bhotised langasges. In Bhatan the influence of the bistorical Bhotian alvaace ta the southward has been more pow~ erful and extensive than in Sikim and Nipal. bot the physical and lin- guistic character of the Bodo and Dhimal shew that beyond the mouns tains it was comparatively feeble and superficial. In the basin of the Trawadi and the connected western territory as far #6 Bongal and the Bay,—the tribes of which are entirely separated from the Bhotians by intervening ones and are less Bhotian in person and customs than some of the Himalayan tribes,—we find that the Naga and Yuma vocabularies are twice as Bhotian as most of the Nipalese. The Miehmi, Jili, Sing. pho, Naga, Yuma and Garo appear to bave a direct glossarical connece tion with Bhotian—whatever may be the chronological and ethnic relation of the Tibetan movement which induced it—distinct from that which Tihetised the more western languages. But to ascertain this relation satisfactorily it is necessary to advert ta the non-Bhotian traits of the Nags, Yoma and of the Gangetic languages, and these will be discuased in a separate Section. In the following details my principal object will be toshew the extent to which Tibetan entera glossarially into the languages of Ultraindia and (India, In the comparative jists in the Appendix, compiled before the publication of Mr. Haodgson’a Sifan vocabularies, I bad indiseri- minately entered all words that have Bhotian affinities. A few are thas incladed of which the derivation from Tibet may be doubted, Mid- Asian and other remote languages having forms that are nearer the cis Himalayan, while others are given which now appear to oe Sifan and pot Bhatian. But making every allowance for these, this Section with the Appendix will sfford a general view of the influence which the Sifans and Bhotians have exerted on the vocabularies of Ultraindia and India from the era when they first found their way across the snowy barrier. 14 ETHNOLOGY OF THE INDO-PACTIFIC ISLANDS. 3. Pronouns, The Bhotian pronoun ofthe Ist person, nga, ma, prevails in most of the Himalayan languages and in Ultraindia, bnt as it is not common in the Naga vialects, itis improkable that the Ultraifian nga is of im- mediate Bhotian origin. I should rather have supposed that in some of the eastern Himalayan dialects it was of Ultraindian and not of Bho- tian derivation, The influence of the Gangetic Bhotian of the Pal era on the Burman family was far too slight to haye extirpated the ancient Burman pronoun of the 1st person throughout all the Burman dialects, The Burman pronoun agrees with the Bhotian and Chinese (ngo) be- cause all have ultimately derived it from the same mother formation. The origin of the Ultraindo-Gangetic nga, na &c. has now been set at rest by the Sifan vocabularies, abit . The Bhotian 2od pronoun, Ahyod, hhyo, khe, has been carried by the Bhotians into most of the Himalayan Janguages, but not beyond them. ‘This pronoun is of itself almost decisive as to the relation of the Ultraindian to the Bhotian, and of each .o the Gangetic languages, The Burman nang, na is found in several of the Naga languages, in Garo, Bodo, Dhimal, Abor, Miri (no, nan), Daphla (no), and even in Magar (many). one of many proofs of the connection between the pre- Bhotian Himalayans aud the Burman family. The Mon and Kambo- jan pronouns, the former of which have spread into the Malay penin- sula, are distinct. . The Bhotian 3rd pronoun, hho, Ahu, is found im several of the Hima- layan vocabularies, but not in Dhimal, Bodo, Garo, Naga or Burman, the prevalent forms being Sifan and Draviroid. The Singha hhi is te Bhotian, but as a similar particle is a common definitive (¢. g. asia ka singular, ki plural )- this is uncertain. : The Bhotian plural particles nam, chag, dag do not appear to have made much progress in India, unless the Bengali dig isa derivative from the last,* The postfixed definitives po, mo &c. are found in several of the Himalayan and Gangetic languages but not in the Ultraindian. Most of the Himalayan possessives are apparently modifications of the Bhotian, but some may be Dravirian, the latter having a wide range of possessive particles. The extent to which Bhotian forms and parti- cles have been engrafted on the Gangetic languages appears sufficiently from chap. IV., so that it ia needless to pursue the subject here. Prom the evidence of the pronouns it may be inferred that the Bho- tian dialect intruded on achain of Gangeto-Ultraindian dialects which possessed the Sifan forins of the Chinese. Save in the vicinity of the southern Bhotian dialects, the Bhotian pronouns have made little pro- gress. Even the highly Bhotien Takpa retains the Sifan-U/traindian 2nd pronoun and bas not borrowed the Bhotian one from Lhopa, Changlo, like Bodo, Dhimal, Abor and all the proper Ultraindian dia- lects, has the Sifen pronoun, and in the sub-Himalayan band the Bho- tian eile not to be found to the eastward of Nipal. That the Sitan branch preceded the’ Bhotian even there and further to the westward, appears from Tibarkhad preserving the Sifan pronoun, A full comparative list of the Sifan and of the allied Ultraindo-Gan- getic pronouns has already been given in chap. V,sec. 11. Among ° Magar ku-rik, (See chap. V, sec, 1). . ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 15 the correspondences there noted will be found the contracted form a of the Ist pron. common to Manyak, Angami Naga and Mikir; ang, the Gyaruug postfixaal form, found as a postiix in Naga and @s a separate form in’ Bodo, Garo and Kiranti; nge Takpa and Singto; ka Thocha, Dhiimal, Lepeha, Lau (kha, kau, ku), Toung hoo, &. Most of the Ultraindo-Gangetie forms of the 2d pron, are Silan, nan, na The Manyak variation of the vowel too is found in Daphla, Abor, Deoria Chutia, Angami, Mozome Angamiand Namsan- gya. ‘The other Sifan particles are also Gangeto-Uliraindian, I add a few examples. yd i The Gyarung particles occur in U)traindian languages. Ma, m &c. is common asa negative and caritive postfix or prefix (Abor, Dophla, Dhi- mal, Mikir, Garo, Burman &c,), ‘he Khamti ma-, mo-, and Chinese pre- posed m &e. is the’same partigle. Da denotes the present in Abor aa in Gyarung, and a Gyarung-like combination of it with /a, another form of the same Bhotian definitive, renders it emphatic, Jada, 1n another form, éa, it is completive, corresponding with the ee yarmng 1, past. In @ Dophla pertect pana a combination occurs similar to Gyarung, by itsell’ being tuture in Abor (in Dopbla bo). Changlo has -/e pre- sent, -ba past (Burm. byi, Bodo bai), ada future (Burm tho, Kham- ti ta—). Bodo has -dang present (useil as a verb yg bai, nai, impertect, dang-man pertect, nise, gan, tut. Dhimal has At past, hi, mhi, mhi present, aro has -na, -enga present, enga-chim impert-, a, ~da pert, -chim perf. def, esa, kheng, fut (Bodo). Naga has -~? pertect, la-prefixed, with -¢ postfixed, as a second perfect (Abor, Gyarung, so in Tibetand-&c. pref. with-s postf.), tis tuture, Mikir has -loh past (Naga, Kas,), -ye future (i Naga), -bo, -bang empha- tic futures (Abor, Daphla, so Burm mi), -si participial (Gyarung). Garo has -na present (Gyarung na-). Kasia hus Ja-past (Mikir, Naga, &c.), n-future. In Singpho -Aa is past (Dhim.- Ai, Bhot.-s also ha-, h- -ha-dai perlect, -a future (Bhot -a). ‘Takpa beiog at present the only known language that appears to be conterminons with the proper Tibetan dialects on the one side and with the Ultraindo-Gangetic on the other, it will be useful to advert to the affinities of its pronouns and particles, I add the numerals to give great. er breadth to the compariaon, As Takpa is the language of the Towang raj, it must be conterminous with some of the dialects of the Bor and Abortribes,* At present we are only partially acquainted with those of the southern Aka, Daphla and Abor- Miri, The two first appear to be the same and to be closely akin to the last. How far they are spoken to the north, and whether any other dia. lecta exist between them and the Takpa, ia not known. * Arethe Tog ab or Tag-ab, one of the tribes of Bhutan who in« habit the district of Tog-na, or Tag na, Tagana or Doka, not a section of the Tak-pa who happen to be under the dominion of the Deb Raja? (As. Res. XV, 146, 140 Pemberton 111). The Pilo of Tag-na’s territory ‘Jies between Bakshaand Cherang. He has two Dwars or passes, and the Refu Jadu and two Tuomas are under his orders, Hie territory ia. eight days journey long and four days from east to west, He pays al- together annually in two instalmenta about 3000 yupees and rules about 8-16:hs of the country” (Ag. R, XV, 139): > . 16 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. The pronouna of all these dialertsare Sifan-Ultraindian, The Tat is oge, nye, in T'akpa, the vowel being exceptional, The Eust Mi thong. Komi tum, Kami ka-ton, Mro shon, Tonglbho thong, Abor-Mirt a-am-ke, a-om-o0, ang-om, a-om-a, have aleo the Bhotian vowel, but ge vo wide a diffusion jn Ultraindia of the Bhotian form cf the numeral ewould be exceptional, and as Horpa bas aleo w (90), it ia probable that u, o forma at one time existed 1¢ Eastern Vibet alan, and were thence trans. ‘mitted to the Burman branch of the Ultraindo=Gangetic family. It ie “probable that che i of Thocha and Manyak has been anbstitated for an archaic u, (Bhotian and Horps), breause the interposed Gysrong has adopted or retained the current Chinese form sem, and the i form hae ~maile little progress in Ultraindia. ‘The change is similar to that of brol, snake, to brie The Gyarung must have bad the proper Chinese form when its glossary was carried to Ultraindia, The Nijpal terme in um, om, may be of Ultraindian and not of Bhotian derivation, - » The Chinese vowel ais retained in Gyarung ka-sam, and in the Ultraine d)-Gangetic Mijhu Mishmi Aa-cham, Mikir Aa-tham, Garo gi-tbam, Naga asam, e-tim, ean-ram, Kachari Bodo tham, Dupbia a-am, Changlo and Lepcha eam and Sunwar sang. The Thochu and Manyak slender variety & shi ri, si-bi is only repree _ sented tothe south by the Sak thin, bot the coincidence appears to be accidental ae the other Sak oumerale have no special egreement with Mapyak. ‘ 4. The Bhotian dshi, shyi ies Serpa end Lhopa. All the other Himalayan terms and all or nearly al) the Ultroindian have the Sifan form, Gyarung Aa-di, p-li, Manyek re-d), Horpa bia, lex The form pli ieremarkable. kt is only foundin the Gyarung ka pli-si 40, where it appears ae aroot with the ordinary Gyarung jrrfix Aa-. Ka-di, 4, iathe troe Gyarung form, and pli must bavebeen borrowed from a Sifan dialect in which p. and not & was the prefix. The archaic prevelepce of such a dialeet is supported by the corrency of the labial, jat ae the qualitive postfix in Bhotian (.po, -b0, -mo),—2d, es thenumeral postfix in Manyak -bi,—dd, a» Sareks in the Bhotian 4 7, 8S and 10 and in Bhotian verhs,—and by the prefiraal position of the qualitive def. in Horpa (ka, ga &c.) and Manyak (de, da. &c ), and of the quali- tive and numeral def in Gyarong (ka). The labial is one of the archsie attribotive defioitivee of the Tibetan formation (qualitive, nomeral, se- gertive), atld the regular archaic posilion of euch definitives “as. prefixual, The dialect from which Gyarung borrowed pli, must have beena very influ. ential one, a8 a similar form hae been widely dispersed on the southern side of the mountains, It appears to be now represented by Takpa in which p-li te 4, and in which it ie explained as the Bhotian prefix (.) joined to the Sifan liquid form of the root, di, ri, re. Takpa pli, Abor a p'-ko, Taying Mish) ka-prei (ak in the Gyarung 8), Mijha Mishm 6 ri a: 40, Garp bri ' Bodo bre, Dophla a pli, Mikir phill, Sin. pho me!i, Naga bell, pi'i, pha’), a-li, Karn mali, Sak pri, Changlo phi, Chepang ploi zho, Lepcha phall, Murmi, bli, Magar bull, Newar pi, Gurnng pli, The Nogeung Naga pa-2 isan example of a similar form in which the root hae the broad form of Thochu gha, Angami N, da, &c The Manyak variety re is Moz, Naga deh (comp. Gyar, di), Burman and Suowar le, . It ig also found with the pretix in the Naga phgle. Bodo bre 4, + Gurung and Murmi pre 8, Kirantire-yaS8 These forme Me exanipies of the:operatvion of aeimilar phonetic tendency, | %, The a form of Thochu, g-zhare 4, khtare 8, and Horpa, his, jd not WYHNOLOGY OF THE (NDO-PACIFIC ISLANDS, 19 “bound to the sou h, save in Ang, Naga da, and a few forms for 8,—Singpho “ma teat, Bodo jat, Dophla pla-genayg, Kasia prah, There is no southern dialect in which p.ia found regularly prefixed to the other numeral roote as well as to 4 and ¢ It-is probable therefore thatit ‘was not carried across the Himalayas by a dialect like Manyak in whichdt War curren! ae the regular numeral! servile, but by one in which it had be- come restricted io 4, &c , or which had borrowed it from a sysiem io which it was reguiariv used. Shendu bas me as ile prefix throughout, bur as4 ia me pu'l, puli must have been received ly it asa concrete vecable or root, in like manner as Gyarung received the pli of ka-plt. So alao Bodo has man-throaghout, and 4 is man-b-re. But in Singpho and some of the Naga dialects the use of the labial i 4 corresponda with ite use in several of the other numerals,—3 masum, 4 mei (40 mili.ei), 5 manga,8 maisat. The ehange of the vowel in 4, is explained by ita arsimilation to that of the root, Io the Kam mali, Naga phale, Lepcha péali, the primary vowel of the pref, Femaina, The west Himalayan (Nipal) forme are evidently of Uitraindian—chiefly ~ Naga—derivaiion, ; 5. ‘pe Bhotian and Manyak forms are the same, ngo, nga, and Gya« yong ie only distinguished from them hy the vowel, o, which is Chinese. The a form is the most commun in the Himalayas and it prevaile almost exclusively in Uliraindia, Io general itis probably of Sifan (Manyak) and ' pot of Bhotian derivation. The Chinese and Gyarong ngois found in Lep- cha pha-ngim and Sanwar ngo. It was probably a North Uitraindian form also before it epreadto Nipal. It has now been found in Taying Mieh. mi ma ngo. The Takpa lia-nge repeats the root for 4, as a prefix to that for 5, and the same asage ie found in Miri pili ngo-ko, Bongja rai oga-kar and Mijhu Mishmi ka lei (with the root fur Selided). These terms appear to explain the Kambojan p-ra-m, Anam la m, oa-m, Nancowry la-m 6, 6 Tie Bhotianu and Sifan forms sre similar. Bhotian bse w, Gyarang ‘and Horpae. But Manyak has wand Takpao. ‘The Bhotian tho, dha, tuk, is probably the original of the Marmi dyo, Newar khu, and Chane glo khun. See App. - co The Bhotian wr, druk is similar to he Manyak ¢ro-bi. This variety and another with thé k- prefix appear to have been the originale of the come mon southern variety. Tukpa kro(Gyarong kutok, Thocbo khatare) Sings kra, Garo Arok, Taying Mishmi tha-1o, Mikir thorok, Naga tagok, thes ok, arok, érok, sora, Borman khrauk, khyauk, Sak khyouk, Kumi tara, Kami tan, Shendu me-churu, Kaki ruta, Tooglho ther, Chepang kroke aho, Lepeba farok, Sunwar rok. Thea form of Thochu kha-ta-reia not found in the svoth) The Barman amplified kh-raok is the original of the Mon Fa-rao, Ka trav, Kbyeng shauk, Anam sau’, The form that has ins troded into the Vindyan system turo, tur and been received by it as a root fo which a native poss, and qasl. definitive bas been postfixed (1uru-éa, tur sa, toro-i&c.), resembles the Bhotian d-rok, Takpa k-ro, Mikir thae rok, Angam: Naga soru, Shinda choru. The Gond sa-rong resembles the Naga tarok, sora. The Mijho Mishmi ka-tham isthe Gyarung 3, kac gam (i ©. 3 dua)). 7. The exceptional Bhotian &- dan, dan is only found in Serpa dyun, Lhopa dun and Changlo sum. The Gyarang quioary ku-sh-nes, Horpa s-ne (2 for 5, 2) are the Tibetan 30 STUNOLOGY OF THE INDO-PACIFIC ISLANDS. representatives of the prevalent Uliraindo Gangetic term, The Gyarung ‘prefix occurs in Abor-Miri and Burinan ky nit &e. The Tibetansh, « ie Mona in Singpho, Garo, Karen &c, The curt Horpa zne resembles the Kode and Garo sni, Bongjv sre, Kuki sti. The Naga and Yuma tani, thanyot, #anet, anath, sarika, sati, Burman éunaik, &c. appear to be connected with the Mijhu Mishmi nun (ning in 2), Abor Ao-nange, Daphla 4a nag, Chepank cha.na-zho, Sunwar cha ni 8. The Bhotian drgyud is not found tothe sonth, The gp. form gye ie cane. and Lhopa. The Gangeto-Uliraindian forme generally aré Sifan.— See 4). ‘ ‘ The west Himalayan terma are of eastern derivation, Chepang prap 2ho, \ Dophla plag-nag. (Thochu khra-re, 8, gzha-re, 4, Horpa hla 4); Lepeha- kakev,—Kami kaya; Kiranti reya, Mormi, Gurung pre,—bre, 4, Bodo, phale 4 Naga, (rebi 4 Manyak, leska 40 Horpa) pre 2 Mrn. The Gyarung or-yet has an egceptional prefix, but it is found in Ultrein- dia as a variation of t,e, Mru hagitin Sri-yat and 7 ra-nhit, Taying Mishmi hae el-yem, 9. The Bhotian and Sifan terms are the same’ The Lepcha ka-kyot, Chepang taku, resemble the Takpa du gu, Dophla kayo, ‘Taying Mishmi konyong, Naga taka, Kuki xkoka, ‘Yunghiu kut. wy 10. The Bhotian broad form bchu, chuh is found in Gyarung and Ti- barkad only. The Bhotian varieties of the Chino.Tibetan numerals have |herefore made as litile progress as the pronouns, They are hardly found beyond the southbera Bhotian dialects,—Serpa and Lhepa—eave in the Bhoto-Sifan forme of Takpa. But there are a few examples of a very archaic exis. tence of Bhotian forme in Sifan-Uliraindian eyetems or of a special connection between such systems and Bhotian in one of its older stages. The Bhotian Jabial nomers! prefix appears at some remote period to have been used in Sifan and Ultraindian dialects, In some itis now more segularly nosed them in Bhotian, ‘The UOltraindo-Gangetic varieties gre either current Sifan, or are cons nected in such a niode with the Sifan ae to show that they must have been derived from syelems that once existed in Rastern Tibet, alihough they are now represented only by remnants that have been adopted into the eure viving systems, The most prevalent Uliraindian systems appear to have 20 Mipal Nega—Yuma which spread westward along the sub-Himalayas fo Dial. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 21 &. Miscellaneous vocables. The ethnic place and influence of the Tibetan glossaries ean only be pro- perly shown by means of general comparative tables of all the known forms of South East Asiaz: roots. These tables must embrace, Ist, the Chinese, 2d, the Scythic, with the allied N, E. Asian, Caucasian, Indo-Europe Semite-Afmcan and Malagaso-Polynesian forms, 3d, the Dravirian Druviro-Asonesian, 4th, the ‘Tibeto-Ultraindiun and derivative Himalayo~ Asonesian, and, Sth, the Mon-Anam and derivative Himalayo-Asonesian, The comparative vocabularies of this kind which-I have compiled are not complete enough for publication; and, for the present, [ must refer the reader to the appended vocabularies, although, from the time that has yew since they were prepared, they are, to a large extent, out of date, shall here examine some groups of roots with more exactness and fullness, The general result of the comparisons I have hitherto been able to make, may first be briefly stated. The verious forms and applications of almost any single root, and the manner in which they are now found dispersed amongst the Tiheto-Ul- traindian and Mon-Anara vovabularies, justify the following inferences. 1st. The Himalaic glossaries have an archaic radical connection with the Chinese. 24, They have bothnn archaic radical and an intimate secondary con- noction with the Scythic plossaries, Not only the root, but various forms god epplications of it, are often common to the two provinces. Of these Seythis forms some have been retaine! in Tilct, while others are now obsolete there, but current in Gangeto-Ultraindian vocabularies. The ‘Tibeto- Butroan and Mon-Anew gloaseries possess many of the normal veriationa to which roots ere liable in Scythio from the change of the vowel; from the assumption or disetrding of a final consonant, umtable from a dental, sibilant or guttural te 2 quid ; ond from the presence or absence ofa servile definitive. In the ancient Bhotion and the allied southern formes, and in the less ernssculated Mon-Anato forms, the Seythie consonantal finals ere found much more frequently than in Chinese. Even the ancient Chinese forms are frequently less consonantal than the Mon-Anam, the Seythie and taany of theenvient Bhotian. The influence of tho modern emasculated Chinese is strongly marked in oll the Tibeto-Ultraindian phonologies, including the brocdest and most consonantal. Sd. There are specie] Ugro-Turkish and Turkish affinities. 4th. Vurious forms of the roots must hive been carried by diferent routes and migrations, and by different tribes, from Tibet across the Tli- malayas. 5th. From the variety of these cis-Himalayan forms, the mode of their Cistribution, and the preservation of several that have been lost in Tibet, it is cartain thatthe Pbetas migrat.ous to the southward commenced at@ wery remote period, éth. Inthe Gengetc—Ultraintion province these forme were further Cispersed and raodified ; and distinct lines of diffusion are recognizable. 7th. Itis probable that from esch of the southern ethnic districts of Tibet, migrations have taken place in different ages, end that the limita - ‘gnd mutual relations of the tribes have varied. At present the tribes ia cofitact with the sub-Hitmalayans, and possessing all the knowa passes, are the Bhotizs and the Takpas, 32 ETHNOLOGY OF THE INDO.PACIPIC ISLANDS. The Bhotias are conterminous with the Gangetic tribes of the Hima- layas, from the Tiberkhad to the Mishmi. The Takpa appear to marc with the Dophla and Abor, They are succeeded again by the Bhotias Kham, who possess the passes at the head of the Assam valley, descend aa tradersinto the Mishmi country, and probably march with the eastern ae The Bhotian dialect appears to march on the N. E. with the Mongo lian Sokpa, which, on the south, is separated by the Amdoan dialect of Bhotian from the Thochn,. To the south of the last, the Gyarung appears to march with Bhotian throughout the rest of its eastern limit. On the western 2% of the northern boundary, Bhotian is succeeded by the Horpa,—which has Bhotian on the south, Mongolian on the east, arid Turkish on the north-west. Its thus, like Sokpa, widely separated from the sonth Himalayan dinlects, but there are numerous scattered Horpas ga well as Sokpas in Tibet proper. The Thochu appears to have the Bhotian on its north and west, Gyarung on its south, and Chinese on its east, F Gyarung has on the N., Thochu,—W., Bhotian,—S., Manyak,—and E., _hinese. Whether it approaches any of the passes of the Irawady basin d not appear. It probably embraces a portion of the basin of the Me-nara ann Yang-tse-kiang, and marches with the other dialects of south western Sze-chuen. The Manyak is probably interposed between it and the northern dialects of the Burman and Law families. at The Manyak appears to lie to the southward of the line formed by southern Gyarung, Khampe Bhotian and Takpa, It is probably therefore placed on the Irawady passes, and may be in contact with some of the undeseribed dialects to the north ofthe Singpho. ‘On the east, and south the Manyak are aly conterminous with some of the tribes of 8. W. Sze-chuen, and N. Yun-nan, if indeed they are not themselyes the Mong- ~ fan of Sze-chuen. ! * From this distribution of the Tibetan dialects we should infer that the pan of the Bhotian vocabularies would enable them to affect the whole ine of the Gangetic ones,—that the influence of the Takpa would be contined to the Abor group,—and that the Manyak and perhaps the Gyarung might affect the Irawady vocabularies, ‘ We find, however, that many of the vocables that are distinctly Bhotian, i.e. hoth in form and meaning, have a very limited by no means commensurate with the present influential position of the dialect, and irreconcileable with an exclusive possession, for any long period, of such a position. : Many of the most widely diffused Ultraindo-Gangetie roots and fornia are common to Bhotian with Sifan vocabularies. Others aré exclusively Bhotian, and others sgein are exclusively Sifan. The broad and frequently consonantal forms prevailed in Tibet when the southern migrations com- menced, for they are the most common in the southern vocabularies. These archaic forms are frequently still retained in Bhotian, where the Si- fan forms haye become slender or vocalised, Manyak sometimes retaing broad vowels where they have been lost in the other Sifan vocabularies and especially in Gyaruug. The Sifan vocnbularies haye some non—Bhotian roots and forms in common with Mon-Anam, as might have been antici- pated from the northern origin of the latter formation. ‘The slender and attenuated forms of the Sifim vocabularies, and particularly of Gyaraug, have spread to the south ata comparatively late period, ‘ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 23 . here have been two well marked periodsofSifan and Bhotian influence {n the south. The first was when the southern migrations commenced, and when the Sifan forms of common roota were probably the same as the Bhotian. © It is difficult therefore to ascertain what common roots of this iod are to be considered as of Bhotian or of Sifin origin exclusively. he prefixes afford some clue. ‘The second period is a very modern one. he Bhotian forms referable to it are in general ‘confined te the southern Bhotian (dialects, to the adjaceut Nipal dialects, and to Takpa, but they are also partially found in more southern dialects. The spread of the later or clendne Ota forms to the southward indicates a distinct movement from the archaic Tibetan and the modern Bhotian migrations, The Chinese influence on the Tibeto-—Ultraindian dialects has been of the highest importance, and very complex. There is a radical comrounity of roots. Chinese has at later periods given numerous vocables to all the Ti- betan vocabularies, and many of these have been carried to the southward. Chinese has also directly influenced all the southern = pasar and vo- ¢abularies, Mon-Anam, Naga-Manipurian, Karen and . From these, and especially from Naga-Manipurian,- Chinese roots and particles have been carried westward to the Nipal and Milchanang vocabularies. The Gyarung is closely and immediately connected with the latest of the dorainant North Ultraindian families, the Burmanic, As this family “extends from the Singpho and Jili in the north of the Irawadi basin to ~Burvaan in the south, it is probable that it arose from an extension of the G to the south, butit also appears to have had a common basis ‘with the older Ultraindian dialects. The Abor dialects appear to be partly ‘embraced in this system, and it has influenced the Nipal vocabularies. The Nega-Manipurign branch appears to be older than the Burmanic, and ‘bo be specially connected with Gyurung. in its older or less emasculated form, and with Takpa, But as it has archaic affinities with Thochu, Hor- and Bhotian, it is probable that seyeral Tibetan dialects have marched ith the Gangeto-Ultraindian, and, pes long course of time, succes- ively or simultaneously disseminated their vo@ables to the southward. The Mon-Anam affinities of the Naga-Manipurian vocabularies greatly n the inference that their connection with Ultraindia is very archaic. The Himalaic glossary is, in great measure, primitive and homogeneous. The poreieiny ine eet iplee Nn vee Vine ty and a well mark- ed, as to show that the group retained its independence and segregation frora the very commencement of glossarial development, and that distinct dialects were formed during that era, So far as other vocabularies are radically connected with the Himalaic, the connection is mainly to be pusibed th their having been primitively branches of the some stem,— dialects of the sume mother-tongue. The Himalaic branch has remained more homogeneous anid more faithful to the primary phase of the com- mon glossary, because the location of the tribes who have preserved it. has iven them a high degree of exemption from foreign domination and influence. The glossary is less mixed then that of most of the other lin- istic families, while it has radical affinities with all of them. The Cunoamat group occupies a similar sequestered * race and it is radically , related to the Bastia, Semitic, African, Indo-European and Draviro- Australian, in the sume mode as the still more primitive Himalaic is related not only to it and to these, but to the Mon-Anam group and to the Chinese. 24 ETHNOLOGY OF THE INDO-PACIFIC IsLaNDs. The Himoalaic glossary has spread to the southward over all dence much of India, and most of Asonesia, Whether the Ugro-Caucasian an other cognate glossaries spread from the Himalaic prevince, or the Hima- laic were derived, with them, from some other primitive seat, is a question that does not admit of so ready an answer. But from the pronouns and particles, it is probable that the ultimate basis of the Himalaic glossary waa a Chinese dialect, and that the great mass of the substantive vocabulary was introduced from the primary Scythic province, The Draviro-Aus- tralian glossary appears to bave been also formed at a period long pte~ ceding the Ee of the Himalaic glossary in its present form to the southward of the mountains, by the engraftment of a Scythic form and glossary on a Himalaic basis. | NAMES OF ATTRIBUTES. Mr. Brown’s vocabularies contain substautives only, so that the Mani-+ puri dialects are not included in this comparison, The omission is unfor- tunate, as, next to the detinitives and pronouns, attributive words (quali+ tives and assertives) are the most stable. . The relations shown by the distribution of the names for the colours and their various epplications, are, for the most part, archoic. In Tibet various forms and applications appear to have arisen in an early bese of the history of the formation. Both the primary full forms and the s6- condary contracted ones, ure found in the Gangeto-Ultraindian vocabula- ries. | : Fer er nae the most important of the Tibetan roots for black is th liquid. In the existing Tibetan vocabularies it has several forms and ap pleations. The full archaic form was probably nag, nak, lag, lak, rag, rak &c. It retains such a form in the words for black and crow in Bho- tian and Gyarung, and in several of the lrawady and Gangetic vocabula- ries, It takes the dentol or guttural prefix in Gyarung and several of the southern dialects. A form with the labial prefix is also very archaic. It _is found in the word f#r the evow in Bhotian ond Gyarang, and contree- tions of it are current in Bhotian words for Jive and red, in Gyarung and -sonthern words for night, and in Naga words for the crow. These sppli- citions show that the root must have been at one time current with the labial prefix in Tibet, in its primary meaning, black, dark. The Bhotion mo-n, wo-n dive, Murmi mo-n nig/t, Gyarung wo-r might, even render it obable that the form mo-nag, mo-rak &e, hod acquired the contracted rm mon, mo-r, before it cessed te be used with its prisaary meaning, It is not prabable that the same dislect would have both the full and contracted formes current us black, The formation and preservation of distinct varieties of the serae root, and the restriction of eat to a specific use, are mainly effects of the existence of dialects, 'The application of other varieties t) red must be explained in the same way, The Bhotian ma-r, Gyarung ve-r, were probably deriyed from a dialect in whicn the labial pref. was ma- an not mo-. The form of the root in the more cormamion word for red, ngi, ni, shows that it originated in a dialect in which nak, black, had taken the slender form, nyok or nyik, This attenuation of the archaic forms distin- “pee the later from the older Tibetan phonology, It is a Sifan and orpa trait. The contracted form ni, with its application to red, must there- fore be comparatively modern, died must ie been known by other terms or forms in the earlier ages of the formation, Horpa and Tho- ¢bu having the slender form of nak, with its primary mcaning black, the ETHNOLOGY OF THE INDO-PACIFIC\ ISLANDS. 25° source of the secondary word ngi, ni, red, is manifest. As Thochu has. a” distinct vocable, it probably spread from Horpa to Gyarung and Manyak, - It has not been received by Bhotian; and’Gya ung, in adopting it, has re. tained also the older word ee pape ). Theni, ling, ri, nya, of Burman, — Angami, &c. show that it has spread to the south. . Bluch. : nag-po Bh. w., nak—po Bh, s., and Takpa, nya-nya Horpa, nyik Thochu, | ka-nak Gyarung, da-na Manyak, © ) : Obs, The Gyarung form is the same as the Bhotian sp. The vo- ealic and elliptic nya, Horpa, is perhaps the original of the Manyak na. Tho-' chu has the amplified vowel of Horpaand i for’, as in so many Sifan words, In most glossarial groups the toot for dlae# is applied to other dark’ colours, b/ue, green, rod &ce.—to darkness, night, the crow, &c.—as that for white isto light, bright, day, shy, air, sw, moon, sileer &e. The Tibetan vocabularies are too limited to euable us to trace the applicatigns and afi- nities of the root for black. That for b/ve is not given. ‘Vhe foots for qreew are different. The Tihetanand Scythie roots for black are applied to the ervm ;—nyag-vo 'Thochu, ak-po Takpa, ab-lak Bhot, sp. (/a-lak Serpa), ta-b-rok Gyarung, a-lok Lepeha, ka-tha-rak Khoibu, (rok bluck, Milcha- nang); gone ¥¢ (Ginrang (also black ); khere Sokpa, kal Horpa, kali Ma- nyak (kara, b/aok, Turkish, chara Mongol. &c.) The ultimate Seythic root is probably found in the wr. Bhotian khata,Sanwar khad, Newar ko. In Sanskrit the root kara has both applications, as in Seythic. . _ The Tibetan roots for wiyit are diiterent, but I give them here as they are applied to black, blve, green, in some of the southern vocabularies, 1. m-tshan-mo Bh. wr., a-sha Thochu, chen-mo Bh. sp., sen-te Takpa, ti a5 in gok-¢i head, nven-ti day &c.). Comp. achsham Turkish, so, chet ong, sui, sii, shiz Yeniseian. In Chinese the root is black, tso (also, hak) Quang-tang [hi, wa Kwan-hwa}. A 2. Ate Horpa [spa, shpa Pashtu, shab Hind., chshefe Zend, kshapa, ansk. : 3. #to-di Gyarung [tin Turkish, oti, at &c, Ugrian]. 4. tong-rmor Gyarung, | 5. kwiuka’ Manyak, The root tshan, chen, is also used, in combination with another root, for green, h-jang khu Bh. wr., jhan-gu Bh. SPs Horpa, zyang-ku Thochu, ‘chan-gu pa. The second rootis green and blue in Seythic, kho-kho Sokpa, ko-ko, ku-ku Mong., Tungusian, ko-k, ku-k Turkish, The Tibetan tshan, sen, zyang &e. is used in Chinese for green with the same form sang, e common Tibetan root for }/aek is not, in the n-g, I-k form, Chinese. It belongs ty the archaic Seytho-Tibetan glossary. Sevthic vocabularies have distinct roots for black, but nog is applied to blue and green ; nog-gn blue, green Tangusian, nog-o, noe-on, nach-on gre Mongolinn, * The Tibetan root is found with the same meaning in the Naga gr. ta-nak (Gyarung ka-nak), #-nyak, myak, nyak- 9, nak,—B mn nak, net, (Koreng, eror, ng) ls nik-nd,—Garo pe-nek,—Abor yak-dir yaka-dak,—Lepcha a-nok,—Milchanang rok, reg, (also b/ve, rak, rok, and. reen, raz). Itis both black and blue in Joboka nak, Nogaung fu-nak, and engsa nyang blue, nyak diwek, Gurung m-long-ya, Murmi m. Jang-ai. Kinawari Bhotian has nang-no as well as mest bs Khan Naga has the form luk in shim-phu-luk, green. ; 26: ETONOLOGY OF THE INDO-PACIPIC ISLANDS. Mak is merely a variation of nak. In Gyarung: it is applied to green, uir—myak. It is found with the meaning Miaek in ‘Toying Mishmi, mak- we, Limbu hi-mak-I4, vod Kiranti maka-chuk-we. _ Both forms, nak, mak &c., enter into ®ames ior night (shy-blach, air. black &wv.), darkness &e. \ amnsang, darkness, rang-nyak (rang-vo light, i. e. sky-white, a-po white); Muthun rang-nak, darhness, night; Joboka rang- nik durkness; Mulung, darkacss, wyuk, night, vang-muk ; Tablung, dark-~ miss, nyak, night vang=—niak; Lurinan, right, nyin, nya. The Tibetan tshan, sha, chen, sen, night, is both right and black in southern vovabuluries. As night it $3 found in Naga e-sane-de (sen-te Takpo), Limbu /fw-sen, sen-git (Takpa sen+t.); Newar cha, Kapwi zying- pha, Koreng n-chun, Manijiin «-hinges As bluck it is Singphu chany, cham, Bodo ga-cham, Mon chang, /a- tsan, /a=chok, Changlo chang-/o. . . ‘he Naren thu, thus, su, is not Tibetan but Chinese, tso. ® The Tibetan form is also applied to grven, as in Tibetan, and to blye and ved. Blue Khari ching+m? and Namsang ¢-ham. Green, Tengsa and Nau- gaune ¢a-cham, Khari shim-phw-luk, Namsuny athing, Joboka hing, Ki- rantichak-la, It is applied to fed in Kyan.c*tahen, Khyeng sen, Bongya tsin,.Mon chang, Namsang a-chak, Garo pisak, Bodo gasja, Milchanang shing, The Thochn shi-dzt, red, is Shauty the same root, ; — The Magar double chik chi appearsto be a slender form of chak,—as che, sen &c. is of tshan &e. . The Gyarting to-di nzyht, (Seythic oti, tin &c.) is the root for Black in Angami het ti. Itis prolmbly found in ti-zi night in the sonant form zi (= di, Gyar.), ti being sky and ti-so day (shy-white). Tengsa a-saig-di, niylt, The Gyarung mor, night, is not a common form in the Tibetan voeqhn« lary of colours, Lt has the same meaning, night, in Murmi, mon. Thé n of rang-pan, night, Namsing,is the same vocable, and it is also found n Garo Walo, Maram mula, Champung nga-yula, Lubuppa and N. Tang- kul maya (y for |,r), 8. Tan@kal ayan,—torms which accord with the inference deducible from those used for red, that the root is the liquid la, ra, na &c. identical) with na-k &c., and that mon, mar &c. are contractions of which the primary Tibetan form was probiljy mo-nag, ga-rag &e. Coinp ab-lak eran ke.” Bhotian has mon-/o0, s-won, ble, It is red in the formanar Bhotian, ver Goyarang, wol, bala, &c. Gungeto—Ultraindian, | The primary meaning of black, dark, is necessary to explain the various applications: It also explains its use asi name for the crow, walo, waru Naga, ola Lhopas at wth The Lhopa phi-ra night, appears to be a similar vocable. The Manyak kwaku’ nigh? appears to be the same reduplicated cuttural t that is ayplied to five and green in Scythic. It is current for night in Kiranti khakwe: The Deoria Chutin sa-ko-koi and Mikir a-ku-k black ( Mikir ingting kok dark) are the same term: The Lhopa nam=mo, Magar nam-bik, Sonwar na-do, Lepeha, Jili, sa- nip, Singpo sa-na nyt, contain the Tibetan word for sky nam (Khamti nap-sinw dark) : Nam, shy, may itself be identical with the Chinese lam, d/ue, and thus be merely oue of the archaic forms of the Chino-Himalaic and Scythic Liquid root for Llach, ¥ ETHNOLOGY OF THE INDO-PACGIFIC ISLANDS. 27 The Ms bik in nam-bik night, is Seythic, pit, pit-n night Samoiede, piti, pi ied fens black Ostink ( bis blue Veuiveinn}.” ' oe My _ The Lau khua night is not Tibetan, re Bed. 1. s-muk-po Bh. wr. 2, muar-pv Bh. sp., ka-ver ni Gyarung’. 3. ie Horpa, ka-ver mi Gynrung, da-ni Manyak, leu Takpa. 4. shi-dai Thocha. 1. The vid Bhotian muk is not fonnd with the meanimg red in the southern vocabularies. Limbu has ube and it may occur with that application in others. - 2, The sp. Bhotitn mar, Gyarung ver, is # common Seythic, Cancasian and Semito—African root. If the labial be radical, the vecable is rare in the south Himalaic tongues. The Murmi bala, wala, Gurung wol-/ya, resemble it, and they sugwest that la &c, ond bala, mar &c, are ultimately the same rost (ma-ra, ba-la). See Black, 8, The Horpa, Gyarung and Manyak nei, ni and Takpa len are forms of a Chino-Limaliic root common in the Ultraindian tongues, Burman ni, Kumi p-ling, Avn-lein, Toungthu ta-uya, Angami m-ri, ke-me-ri, (comp. Gyarung ver, prob, ve-r), Naga gr. ma-lam, ma-lam-la, td-mu- ram, me-ram, Kambojan, Aa-rhum, Deorm Chutia sa-ru, Abor ya-lung, lu—dak, Sunwar ln-la, Kiranti ha-la-la-wa, Limbu ku«he-la, Lepcha a-he-ur, (Comp. Mong’, ulan ved). The root in its broad consonantal form is b/ve in Chinese, lam Quang- tung, lan Kwan-hwa, wlience the Karen la, tu-la, Limbu leh-la, Chinese has also lu green. The application of lam to red and D/we seems to show that its primary meaning was black, dork &c,, for the same word would hardly be trans ferred from red to blwe, oy vice versa. In the older Himalaic formation of Ultraindia—the Mou+Anam—the root retains the meaning black, dark &e, Siam, Lacs, Ahom dam, Khamti nam, Laos nin (the slender Tibeto-Bur+ man form ior red, niy i, ling &¢.), Simm dam nin, Anam den (night dem), Kunii da-oum, Kiri wetnun, Kasi darkness dum, Nicobar biwel ringus lum-#, Touns-tha patleng, The roct is very ¢ommon in the Indonesian vocabularies in various forms, applied to blick, nivht, dark, fog &e. tam, lam, lom, rang, rom, ri &e. &es } : The slender form is also red in the Lau family deng, heng, len, forma corresponding With the ‘Kumi ling, lein red, Tungthu leng, lLau-Anam nin, den (laek, and indicating a special relutiouship between the Mon Anam ond the oldér Irawady vocabularies; ' The Dravirian and North Ginvetic languages have the same root, The Male mar-90, black, is identical With the Mhotian mar re’, The Kol ara, Telinga erra-pu, and the Hindi lil and Bengali ranga, resemble Nipal forms of the Chino-Hitnaluic root. ' Finally, it appears probable that the Tibeto-Ultraindian nak, nang, lok, rok, long, lang, nyik, na &e. Ke. black, Pive,—the Chino-t ltraindian lam, Jan, In, nan, ram, lunge, ru, nun, num, dum, &e, blue, green, black, red,— andthe Tibeto-U Itraindian ni, nei, mi, nin, ling &e, red, are all variations: of one primary liquid root, which, in the eastern branch o# the primitive ‘ossary, early touk the form la—m, la-n, and in the Tibetan the form na-w, ae &é. The Sifan ni, ugi red (whence the Ultraindian ni, ri) has the slender form proper to the liter Sifan phonology, and the original was 28 ETHNOLOGY OF THE INDO.PACIFIC ISLANDS. probably the common Tibetan root for black, which has undergone varia- tions that approximate it to ni, the current yradations being nak, nyik, nya, na. The m forms for dluck, blue &c. in the Irawady and Mon-Anam vocabularies are evidently eastern or Chinese, and not Tibetan, in their im- mediate affinities. They appear to have been communicated by the Mon- Anam to the Tibeto-Ultraindian vocabularies. . The Tibeto-Ultraindian tshan, sen &c. Dlack, night &e. is, as we have geen, applied to red in Singpho and some of the Yuma dialects, A guttural root is found in Singpho, Khyeng, Jobuka khi, Mon kit, ket, Karen go, gho, Magar gya-eho. © Green. 1. A-jane khu Bh. wr.,* jhan-gu Bh. sp., Horpa, zyang-ku Thocha chan-gu Takpa. 2. pws Gyarune. 3, chu gin do Manyak. (? chu-gin-do a form of 1). i 1. jang, chan, zyang &c, is the Tibeto-Ultraindian root for black, dark, night &c. already examined. It is gpplied to qreen in Naga dialects and in Kiranti, Thatethis was an archaic application is shown by the Chinese tsing, sang. The guttural is the Scythic root for blve and green, also cur- rent in its double Scythie form in Manyak and some of the Gangetic lan- guages for night, bluch, dark. The double form with a slender vowel is green in Sunwar gi-gi. Miri has ge-dak. 2. The Gyarung myak is one of the forms of the Tibeto-Ultraindian nak, nyak black. The southern names for green are derived from the roots for black. In the Naga group we find ¢a-cham, a-hing &c., Kiranti chak-/a. Limbu has leh-/a, Serpa and Lhopa num-mo, nhyam-bo, Gun ur- kya, Milchanang rae, Ti Shik hha dele aaa forms of the liquid root ans used for black &c. Chinese has ln green, ; A labial is common. Angami he-peje, Je. Sa phung phong, Murmi ping-ai, Newar wa won, Magar phi-phi dan-eho, White. = 1. d-kar-po Bh. wr., kar-po sp. The Bhotian kar is probably a cons traction of ka-ru (Comp. ke-ru Takpa). 2. phru phru (pi—rn) Horpa, kke-ru Takpa, a-prom Gyarung (p-rom), da-lu Manyak. . 3. hes Thochu, 7. 1. The Bhotian root, if not a contraction of ka-ru, is archaic Scythic,— . kyr Samoiede, gil-taldi Tungusian, It is a plied to star in Bhotian, s kar-ma, kar-ma, (Abor ta-kar), Horpa re, Manyak kra (Gurman kre). ; its primary meaning kar is only found in the south Bhotian dialecta of Serpa kar-po and Lhopa ka-po (star ka-m). The Murmi tara, Gurung tar—kyq (also sta) are moditications of it, 2. The more prevalent Tibetan ph-ru, da-lu, ha-p-rom, is Mijhu Mish- mi kam-ph-long, Singphu ph-rong, Burman ph-ru, pi-yu, Kumi k-lung, kan-lum, Kami a-lum, Kyau ag—nung, Bongju k-lang, Garo bok-lang Mikir ako-lak, Dohphia pung-lug—pa, Lepeha a-dum, a-dom. A slender form is found in Taying Nisha ishmii, leo-na, Kasia ba-lih, and Deoria Chutia The Mijhu, Singphu and Burman forms, phlong, phrong, phru, are re- ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 29 ferable to the later concreted Horpa and (Gryarung phra, prom. But the more common form in which the root does not take the labial prefix must be of ofder Tibetan derivation: The root is applied to air, light, day, sky, moon, star, Sc. The Tibetan forms for air have both the wv and @ vowel, as well as the slender moditica- tion thak, dak; rhot, lang, ryu, zyu; li. The archaic final consonant is pre- served in some of these forms. Similar forms are found in the southern vocabularies, The variation nung, nong, occw's ia the Manipuri gt.; lom in Lau; and rang, nang, lam, lan &e. in several Naga—Gangetic vorabu- laries. The Tibetan names for the moon have the same root im the forms Ja, da, lik, le’, Ihe’, le,—lik. le? &e. retaining thé guttural final as in Thak, da’ air, The u form is found in the Lau lui} Siam duen, nung, Yuma s-lu, lo. Anam has &-lang, d-lang. For star Gyarung has ¢tsi-ni ; Nam- sanix me-rik ; Monnong, Lau fam., lau, dau, nau. [See 4, Names of inani- mute natural objects]. The Chinese leak, fight in colour, is probably the same root. 3. The Thochu phyok may be a softening of an archaic form of 2. phyok for phrok. But as there is nothing to justify such an opinion, it must be considered as a labial and exceptional root. The Chinese term is the same root. Knang—tung preserves the full form pak, Kwan-hwa has pe. Itis found also in the Tungusian bak—da, wag—da. Fin has wal-gi, Ugrian woi-kan &c., but the prevalent Seythie roots are different. - Inthe south it is only found in the Lau family, and some of the Irawadi- Gangetic vorabulivies that have most affinities with that family. Lan pheuk, Ahom phok, Khamti phuk, Khyeng buk, Bolo gu-phut, Garo bok—lang, Naga a-po, Miri kam-po-dek, Toung-thu and Pwo Kuren “ ba, Sau Karen wa, Magar bo-eho, Sunwar bwi-sye. ( Naga ting-puk shy). rom the vowel u, 0, occurring throughout, all these forms appear to be referable to.a single vocabulary, probably the Lau. The contracted Naga—Karen forms are the parents of the Nipal bo, bwi. The o, u vowel connects the Lau with the Thochu form and not with the Chinese, The Lan fam. has also a distinct root khau, khong, The Naga ma-sang, eae me-sing, heng, che, choh, Angami ka- cha, Dhimal jee-/a, ‘Mon tchu, Nicobar te-so, ¢?-so-ab, Sunwar sye of bwi- sye, Tiberkhad chong, is a root common as applied to light, star, moon &e. ‘The Thoehu chha’, moon, appears to be the same root. It occurs with the same meaning in the Manipurian ka-chang, Milchanang ge-tchang Tiberkhad chang’ white), Manyeak nash-chah day, sun. Jili has ka-tsan, ingpho tsan, Bodo shan, Garo san, ra-san, Naga san, Kol singi, burman a-si, N. Tangkbul a-sun, Tiberkhad zhaug-ma ; tsing-mik saa Luhuppa (day-eye), shi-mit N. Tangkhul. Jaght,—Jii has thwe, Singpho ning-thoi, Tangkhul she, shea, Kasia ba-shai. Star,—In Chinese it 1s ap- plied to star, sing, sing, shan Kwangt.; tsin, tsing, is dighé (in colour) elear &c. Sky tsaug tien, tien Chinese, yi tien day. Oxzs. 1, Both broad and slend rforms of the roois have been anciently eurrentin Tibet and have receivad different dialectic applications. The sama root has ulso been applied differently in different dialects, ‘Thus nak or lak is ilack in one voe., crow in a second, blue ina third, green in a fourth, darkness, night, in afifih. Shan &c. is black in one dialect, might, dark, blue, green, red, ‘in others, Where the root has more than one application in the same dialect the different mennings are sometimes distinguisled by the definitives, as well as by the form of the root, This usg of the def, is generally abiirary, Jt sometimes | 30 ETHNOLOGY oF THE INnoO-PACIPIC IsLANDs. runs through several dialects, indicating a diffasion of the term, but in othef coses the samedef. occurs with a special force in one dialect and without it in others, Bhiotian has nak.no Alsek, ab-lnk crow; Thochu nyag-wo crow, nyik biack; Gyarung has &a-nak black (corresponding with the Serpa form of Bhotian Aa-tak), while for ersw it prefixes ato the Bhot. compound of the root and tha Jabial pref., é2-5-rox ; Gurung uses the labial form with both meanings—m long-ya. In the southern dialects the root oceurs with the labial pref., and with the gute toral or dental, in its primary meaning black, while one of the Manjpuri dialects prefixes the gut. to the dentalin ifs word for the crow. All thia it a covecquence of the present dialects having been formed when the roots were not concreied with the prefixes, and when different def. might be used for the same purpose. 2. Slender forms of thé root occur in Thochu nyik, Burman net, Ko+ reng nget, Bongju nik. Cageees Milch, reg, dlick ; Bhot. sp. chen, night, — Takna, Limbo sen dlack,- -Naga gr. ching (due, hing green,—Yuma sen, tsing Mileh. shing red; Horpa ngi, Gyar. ,Man., Burm. ni, Angami ri, Yuma ling, Lan fam. len, deng &c. re¢,—Lau, An. nin, den, Toung-thu leog deck, —Barms nyin night. The liquid root for whe has only broad forms with that mrao- ingin Tibet, but slender ones ocevrin the names for the moon. Deoria and Kasia have slender forms for wArte. 3. The special East Tibetan gonnection with the southern languages is ' wellshown by the word for red in Gyarung, Manyak and Burman, pis and by the word for white in Horpa, Gyarung, Mijhu Mishmi, Singphu and Burman, M. M. preserving the double pref. of Gyarung, It also iltustrates the spe- cisl Gyarung and north lrawady element in Burman as distinguished from the older cialects ofthe same family, the Yuma forms being Talpa (which again appears to be Bhotian). 4, The Karen thu, thun black is Chinese ; fe Ja, la blue is probably frem the Chinese lam, and some of the Naga and Nipal terms appear to belong to the game relationship; wa, bwa while (taka po Ught) have Naga and Nipa! aff. 5. The Mon-Anam family have a distinet archaic form of the liquid root for black and night. It is retained in the Yuma dia ects and Kasia in a broad form; Lau and Anam havee forms. Slender forms ere also applied to red in the Lau fam., and were probably communicated by it to those dialects of the Tiveto-Irawady family that first spread to the south, as these forms are found in Toung-thu and Kumi. The Lau word for wh:fe—similar to the Thochu—has been communicated to Khyeng, Bovo and Garo; and in Karen and some Naga and Nipal dialects it is retained in a softened form. I add a few more words of this class for the qraroee of illustrating the connection between the Himalaic and Asonesian languages, but without attempting eny exactcomparisons. They are roots ofa class that have many applications, and several of the published vocabularies do not contain em. Large, Tineran. 1, chhen-po Bh. wr., then-2o Takpa, kam-thu Horpa, ka-hti Gyarung> See Long 2. Chinese, long, cheung, chang &c, a er 2. s-bom-bo Bh. wr., bom-bo 8, acd Lhopa, pwi-tha Thocliu, Chinese i gréat. . 3. kah kah Manyak. f Chinese kn; broad kwan kwoh, ; SourHEny. 1. the-ba Gurung, a-ti-m Lepcha, a-chung Muthun, chong Joheka, yong Mulung, yong-nong Tablung, joh Angami, jo-pur Mozome A., yom-ta Jimbu, nga jang Murmi. 2 gu-ba Singpho, ta-pe, té-be Tengsa Naga, ta—pe-tiau ixpari, jo-pur M, RTHNOLOGY OF THR INHG-PACIFICG TsLANDS. 81 Angumi (jo-su long), bote Abor, (fut, ta-bok hes , Nogaung, ta-bit Khari, wa Barman, phum dnech? ales Jat =gu-phung Bods, LWa pal Gyar., round wa’ wa’ Manyak. . 4. ka-tai Mijhu Mishmi, a-ta-dak Miri, u-to-yang Kiranti, tan-go Newar, go-da Garo, go-det Bodo, dham-ka Dhim., (fat g-ta-t Numsang ), This root is Chinese, tai, ta, Anam dai. 5. d-rang Taying Mishmi, lone, lungz, nang, Lau f., ta-lulu Nogaung Naga, yongsnong Tabliné, a-done Namsan, nan Muthun, do Karen, Jen Khyeng, Kumi, leng Kami (also fut), tha-not Mon, (fat, u-do Abory tok, tup Burm.). See Lony 1, Tib. ring, Southern long, lang &e. diound, ka-\nc-lar Gynruny, lo-lo Horpa, z-lam-po Bhot. wr., ri-ti spi, de-rhi Takpa, a-sya-ra Thochn, fa-rany \ogauny, din-din Sinepho, k-lom Siam, Ahoin, #-om Lan, pv-lnun Kami, py-lu Khyeng, ph-in, ph-lon Karey, lun, long Burman, tung-lung Toune-thu, rer-rer-do Lepcha, ril-to Murmi. The Serpa gir-mo, Sunwar kul-kul rownd, Ser-pa gir-bu fle, Burm. kri, Mazar kran, are probably contractions of the common mono- syllabic kind of forma similar to the preveding—A-ri, or /M-ri, gi-ri, ku- lu—hu-lu, #r-raa [=Gvar. ha-lar, Now. fa-ranz, Mon fat kara]. 5 a. kri, kyi Barman, gir-)n Serpa, kran-cho Macur, (rowed, Serpa gire mio, Sunwar kul-kul, fat ka-ra Mon, kal-bo, gal-vo Horpa, cher-wa Thochu). Prob. 5, 5 b. Lau f. yai, yau (also long). Prob. fiom lau, rau, forms of 5 used for long. Long. ‘TIBEran, ¥. ring-yo Bh. wi, rim-%o Bh. sp., ring-to Takpa (2ug-ring tall) 2. ka-chi Horpa, dri-thi Thochu (ur thi Sokpa), sha-sha Manyak, See Large 1. a 3. ha-ari Gvarung (also tal’); dri, Thochu and sri Gyar. may be sri, d-ri(1), Inlike manner the Manyak hra hra, tal/, may be h-ra, SOUTHERN, 1. ka-lone Tayine Mishmi, go-/h-rang Mijhu M., ga-'u Singpho, Boda, thin-k« Dhimal, fa-lan, a-lo, lo, lau, lang-/tv-lo, Naga, ji-lo Garo, lot-cho Mavar; rhin Burm., k-er-k-re Angami, fa-lein Mon, lui Deo ria Th., vau Lau f. [laa Nax.], reng-)a Murmi, rhim-Jo Gurung, a-rhem Lepcha, (See Larve 5, Tall 2. 2. she Burm., jo-su Mozome Angim’, See Large 1. Tilt TIBETAN. i. thom-bo Bh. sp. ‘ . ga-khye Horpa, Chinese ko, kau high, 3. ‘tha. Thochu. 4, ka-sri Gyar., zug-ring Takpa, hra lira Mary. See Long 1, Large 7. Southern. 1. a-tho Lepcha, sune Lanfam, See Derge 1, 4. 4. m-rang, m-yen Burm., lang-la Naya, lhun Khyeng, tha-lon Mon, Bee Longl, Large 6. : 5, tau Mulune, Tablung; Chinese kau; (¢ for k as in tau J, thu 9 &e.) tau-ga Newar lurge, 33 RTHNOLOGY OF THE INDO-PACIFIC ISLANDS, : On The root in ch, th occurs in Tibet both with broad and slender yowels, 1. thu large Horp., thom tal! Bh., tho Lepcha, dri thu Jong Thochu, jo su Moz. Ang., a-tum rvvnd Namsang. To this are related the Naga large chung, chong, yong, Limbu yorn,— Lau tal! sung, which are still closer to the Chinese cheung fong. Kiranti to. “ae . 2. sha long Mony., bra-tha tall Thochu, 3. chhen large Bh. wr, then Takpa, thi Gyar., the Gur., tim Lep.; chi long Horp., she Burm. | The liquid root has similar variations, 1. lum, round, Bh. w., lom Siam, lo Horpa, lun, lu Yuma, Burm., Kar.; long, long, Taying Mishmi, In Sing,, Bodo, Garo, lo, Nag; rung large T. Mishmi, lung, long Lau f., lu, dong, nong Naga, do Karen, @. lar rownd Gyar., ra Thochu, rang Nogaung; ran fat Magar, ra Mon; rang long Mijhu; rang ¢adi Burm., Nag. $. ri round Bh. sp., rhi Takpa, din Singpho, rer Lepcha, ril Murmi; k-ri fat Burm,, gi-r Serpa; ring long Bh. wr., rim sp.,s-ri Gyar., d-ri Thochu, rhin Burm., reng Murmi, rhen Lepcha, rhim Gurung. ‘Allowing for purely local changes, the distribution of Guess feunh shows a special relation between Bhotian and Gyarnng, and between both—bnt éspecially Bhotian—and the Lrawady or Burman group on the one side, and the Nipal on the other. As some of the dental words are variations of the liquid, I will only add the lubial. é ; 1. bom Jorge: Bh., phum ,fat Singpho, phung: Bodo; bote large Abor, bok. fat fener Nag., po-tsu Angami, po-moja Moz., pur large, ke-mer round, Ang. 2, pan fat Gyar., tok pan Kironti; ba large Singph.; pan round Ahom, man fvt Siam, mon rovnd. hamti, 3. “ large Chinese; pwiThochu, pe, be Tengsa, pe, bi Khari, pi fat Lau f. ‘The Lau lom rovad, long; lung, Jorge, show a Bhotian affinity, which Mishmi partekes; yau long is 2 Naga form, lau Tabl., ¢2-lhaun Khari, nau large Muthun; sung tall, high, is also Nara, josu Moz. Ang., chung large Muthun; pan rovnd Ahom, man fat Siam, are Gyarung; tui fat Lan, is Nava, po-tsn Anp. a-syu-m Lepcha, sui-ne large Deor. Ch.; pi Jat is Chinese, pi large, The Mon tha—not farge, is Tobokn fat nut, Magar lot-cho long; ha-lein lonzis the common T. U, term; tha-lon tall is Khyeng &e.; ka-va fat, Mavar /-ran &c.; kha-toung rowndis Anam ton, Toung-thu tung-lung, Abor, Namsany, Deoria tum. The Kambojan tom large is thu &e. of Bhotian &c., but in the Bh. form for tal] thom, Namsang &e. for round, tum; mon ravad is Khan ti (fut.Siom toan;Gyar. pan); ri-sing Jorg has the T. U. ri, ring; ka-pos high, (bote large Abor); tuit small is a form of the C. H. roct for gnvell, short, occuring in the Naga gr. tut for short, Joboka tut, Khari tut-si, Nog. tat-su; Ai-le shortis smell in the Lawf, lek (Nats. a-ring), ° / The Anam dai, large, is Chinese (Kwang-t.) tai, and it has been com- municated to Mijhu Mishmi, like many other Anom words; jei thiak anil siei Jong are probably connected with the Angami se, siand the cognate tr U. words; ton rownd is Mon kha-toung, Toung-tha tung-lung, Nam- > + ETHNOLOGY OF THE INDO-PACIFIC IsLanps, Sa sang -tum, Deor. Ch. tumo-ru &e., Miri a-tum-dak ; kan high is Chinese kau ( Kwan-hwa); nyo sma/l is Chinese lioh ( K=h.), or Lau, Burm. &e.; ee ede Lau, tam,—an archaic form of the Chino-Tibetan twan, thung, tha, oc. The K-t. Chinese tai, /arge, has been received hy Anam and Mijhu; the aff. of pi and ku are archaic; kau high (K-h.), Anam kan, Mulung and Tablung tau, Karen tho, to, Toung-thn a-kho (K-t, kd); the close aff. of the roots for small, little with the T. U. appear to be all archaic; chang, cheung long has archaic sos > ; mall, : TrIneran. 1, chhimg Bh. wr: chun-chung Bh, s., chung-Jo Takpe, ka-chai Gya- rung, bra-tsi-tha Thoechu. Chinese tain, ai, shin, sin, sai, sie. 2. phra Bh. wr., pru Takpa, [bra-tsi-tha Thoehn, bra-tha. ta//]. oe 3. kam-ma THorpa. é a, yu Manyak, SovTnerx, o 1. ka-tsi Singpho, ka-tshi little Pwo Karen, te-sn Tengsa Noga, sui Tablung, (soh short), Moulung, ke-chu Angami (¢/in shya Burm., @-cha Namsane, a-chim Lepcha, yo-Shu Limbu, ! ¢-hi-pia Muthun, hiJoboku. 2. a-ring Namsang Naga, ti-lala Naugone Naga, ning-haji Khari, lek, let Lau f., (thin, kam-rhang Horpa, ka-ri Manyak, rid-po Bhotian, nen+ ma Serpa, ru-cho Magar. }. 3" 3. a-me-dak Miri, po Sgau Karen, pho Pwo, (phu short), a-hi-pia Mu- thun, (tien, ma-bo Bhotian, ta-pa Kumi, pam Khyeng a-po Tenysa, a-po- prr Nogdung, bye kb Lhopa. 2 che-kaTaying Mishmi. 4. a-ngi-do Abor, (gi-dak, thin), nze Burman, noi, on, lek noi, Lau f. nao Khyeng, nyo Anam, lioh Chinese. Short. 1, thong-po Bh, wi, thun dung Bh, s., thong-pe Takpa, ka—chan Green x pee Thochu (man), ga—de Horpa (man). Chinese tun, twan, 2. kalze Horpa. 3. won hi-tha Thochu, ~ 4, dri-dra Manyak. 1. ha-tyoa Taying Mishmi, ga-thi Mijhu, jw-tun Singpho, to Burm., do Kumi, doi Kami, twe Khyeng ¢—ton, ¢a-tsu, tut, so Naga, a-tan Lepcha, tang-ba Limtm, dung-ta Kiranti; tam Lau fom.,thap An, See Smail 1. 3, See Small 3. 4, See Small 2, > Fat, a 1. zo Bh, w., Takpa, so Bh. s., ta-zo Gyarung, a-dz Thochu, nga-jen Manyak (Chin. shik, shi). ne Same Burm. cha, sa, Singpho shan, Naga chau, tyu, cha, sa, sang, ha, chi, Mon tsi, Lepeha zo, tha, Limbu che, Kiranti cho, Murm. chou, Gurung chad, Sunwar jau; Magar chau. - 2. na-ngi Horpa (? ki Chin.), Lan kin cat, drink. f Drink. ' 1. }-thunge Bh. w., thung's., thone Takpa, wa-thi Horpa, a-thi Thochu; ngiieestte Manyak, me. RTS F ae “y ol ETHNOLOGY OF THE INDO.PACIFIC ISLANDS, ehu-ma Taying Mishmi, pong eho Mijliu (-chw assertive post.); Burm. sok, thank, Abor tu-pu, tai-py, Limbu thung-ne, Kiranti jes Murmi thung, Newar ton, Gurung thunu, Sunwartung, chu &c. mater. The root for mater precedes another root in severs! dialects, Namsang jo-k (jo water), Joboka ti-ling (ti mater), Muthun si-ngha, Tablung yang- Ying (riang water), Tengsa tu-num (tu water), Augami zu-krat, i, A dzu-kret (zu, dzu water). 2. ta-mot Gyarune. : Sleep. 1, nyan Bh. w., nye s., nyet Takpa, aman Thochu. Naga ana-27, Murmingung, Lau fam. non, nap, lap. 2. gur-gyun Horpa. —— oe yum. - 3. kor-man Gyarung (Chin. fan, min), Mijhu raui-riu. 3. khai-ya Manyak. ; ® @4. Naga jip, jup, ipi-silo &c., Limbu ip-se, Kir. im-sa, Sunwar ip. . Come. , 1, hong Dh. w., hai Thochu, s-byon Bh. W, Ja-pun, pa-pim, Gyarung. Sunwar pyu, Newar wa. Mijhu hoi-cku, Taying M. hona-na, Len fam. nia, 2. syo Bh, s., Takpa. Singpho sou. ae : 3. kwi-lhen Horpa, le-mo Manyak, (Chinese le, Ini, lam, Sokpe ire). Burm. rok, yauk, la, Naga a-ring, a-ha-lu, Magar rif&-nz, Kirenti ba-na, Lepecha di, Linbu phe-re. = Ge. F : ry : ef song Bh. w., ta-shin, ma-shin Horpa, da-chin, ye-elLin Gyarung th. lin). . ( eat ewa, Naga tsu, tong, Miri sn, Murmi sye go, Newar hon. 2. gro, gyu Bh. w., eyo Bh. s., (k-g7o, gryo, move, walh). Kiranti ka-ra, Sunwar lan. * 3. da-kan Thochu (also more, malh), gai Takpa. : Burm. kwya, Naga kao, Limbu be-ge, Lau fam. ka. (a softening of ya- chin), yu Manyak. 4. ye-yen Gyarung, nh a 5. bo-nd Taying Miskin, phai-clw Mijhu., pai, men Lau f. Names or Jxaxrmare Natvrat Onyects, Tor air, sky, day, sun, light and fire, there are three principal roots in the Tibetan vocabularies, each occurring with all or most of these appliéations. 3 ‘ : Ist li, ni, ne, nyi, nye; lang, dung, rhof, ryu, zyu; Thak, da’, nam, ung, nga. ¥ = ad ey me, mi, mah, meh; mon, meun, mun; wuh, wot, hod, pho, uik, - we, eu. 3d koh, khah. _ The 1st and 2d of these roots are also applied, as we have seen, to. white, moon, star, = RTHNOLOGY OF THE INDO PACIFIC ISLANDS, 35 = Air. Shy. Sun. Day. ~ Taght. ; wr.r-lung-ma namkhah nyi-ma nyin-mo hod 3 Dhotian, i lhak-ya nam — nyi-ma nyi-ma hwe 2 Horpa puryu _ koh nga nye-le sho 8 Thocha mo-ZyU mah-te mun stynk-lo et 4 Gyarung ta-l tu-mon heit=ni, nye ulk teu-meun 5 Manyak mer-da’? mah nyi-ma nash-choh wuh 6 Takpa rhot num-—dung p-lang nyen-ti wot Fire. Moon, ~ » Star. ‘White, 1 sa-meh 2-la-ra. s-kar-ma — d-kar-o 738 cima kar-ma kar-no 2 w- . #-lik-no sore . phrauphrn 3 meh chha’ _ _ ghada pieokt J » 4 #i-mi . tai-le, chi-le ¢si-ni ha-prom 5 me, ma, Itie’ kra’ da-lu « 6 meh le’ kar-ma- khe-ra 1. The liguia root. Air. The Gyarung fa-liis the Tibetan representative of the most common form ia the Burman branch,—Toung-thn éa-li, Klyeng ha-li, Karen /:h—li, Mra ies Bocas le, &c. Itis also Aka do-ri. This slender form is Ugro- Turkish, - : The Manyak me-r-da’ is allied; 1st, to the Bhotian sp. Thak-pa, Murmi lha-ha, Kiranti hak; 2d, tothe Naga ra, rang &c., Mishmi @renga, Gurung nang-mro, Mileh. lan. ' The w variety of Bhotian wr. lung-ra, Serpa lung-ho, is allied to the Takpa rhot, Horpa pu-ryu, Gurung w-ro, Thochu mo-zyu; the Takpa rhot to the Maram w/-lut. A similar form of the slender variety occurs in the Chanele ridi, and’ Fhoibu nong-lit. The -t—, -d, isthe -k of Bhotian, which Horpa pre- serves in lik moon. . The Lau. fim. has lom, lon, Mon b-loei (also k-ya= ho-la). Shy. The Bhotian nam sky appears to be an archaie variety of the root, Jn the uw form it is also archaic Scythic, nom, num, nob Sumoiede, nu- ma, nomen Ugr, (lumen Lat.), and, with other vowels, a widely spread name for sun, fod, prophet, king &c., nim, nem, neb, nab Ke. The Bhotian nam is found in Takpa, nam-dung, in combination witha @® form of lung, nung. tis not found in any other non-Bhotian vocabulary save Kiranti nam-cho, and Kashmiri nab, Applied to day it also ocvurs in the Murmi nam-sin. As sw it is Limba, Kiranti, and, in the contract- ed form na, Sunwar. Maar has nam khan, which is the Bhotian nam kha, shy. As sun the root is Hungarian nap, ai Ost). It is found in some names for night, in which itmust have had the me1n- ing shy (shy-black, shy-dark). Magar nam-bik, Lhopa nan-mv, Subwar - ‘na-do, Lepcha and Jili ga-nap, Singpho sa-na. The Takpa dung is found in the Naga rang-tung; in the original 36 RTINOLOGY OF TUB INDO PACIFIC ISLANDS. form in the Tengsa a-nung, Manipuri nung-thau, in the Anam tung-tien (fen ‘hinese); and, with the slender vowel, in the Abor. fa-ling (comp. Gyorung fa-li air), Kheri a-ning. | In the Manipuri dialects, the Takpa and Naga dune, tung, may he the parentof the ¢ forms, ting-puk, ting- em, tinr-am, ting-wram, ke-zing, Ke-zi-rang, ka-chi-rong, but it is more gbeble that these are variations of the sibilant root. The Turkish and Mon:rol combine a similar form of the root, teng. ten, (immediately re- lated to the Chinesé, tien) with the Uero-Turkish ri ar &c. (teng-ri &e.). The @ form: is still more widely diffused, J-ra Mishmi, ram, rang in the preceding: Manipuri compounds and in others, tang-ban, thang-wan ; rang- tune Naga, no kio-rang Bodo, ta-liang, sa-rag, sa-ranei Nipal, sa-rane Male.* ie Mon-Anam vocabularies have p-leng Chong,, b-loei Anam (air in Mon), F _ The Bhotian variety has a very narrow range; and as the forms dung, mune &e. are found associated with zing, 2i, ling, rang &e, the probability of similar wu, 0, forme having been currentin the Sifan dialects as well as im the Bhotian, with the meaning a?y,-ia increased. . The Naca-Manipnn rang, ram, (with the fa-, sa- prefixes of that group) appears to have been carried along the Gangetic basin to the Bodos, Ni- palese and Rajmahalis. It corresponds with the Bhotian Thak air, la moon, The Tibetan liquid root for white ru, lu appears to be the same root. In the Gyarung rom it preserves an archaic m final, asin pam, ram &e. sky. The Lau lom ar has both the G. vowel and final. Comp. also the southern forme for mhite, lam, dum, lung, Jun, long, rong, nung, lang, Ine, Ink, ih, ti, The Bhotian nam sky must be considered as a variety of rang, ram, none, Ihak &e., air, sky, white &e,, and not of the Chinese Jam dlve (ante yi. 26). The Takpa nam-dung sky, is evidently the same as the Namsang rang-tung, Sun, Day. The same root is svn in Takpa plang, and Horpa has also the a form, nen (Anam neni day). DBhotian and Manyak have the slender form nyi-ma, and it isslso Gyarung kt-ni (Comp, ta-li air), With the meaning day this form is Bhotian, nyi-mo, nyin-me, nyi-m, ni-mo, Horpa nye-le, Gyarung nye, Takyn nye-n-t#, N, Ultraindion, im all the groups,—si-ni, ta-ni, fa-na, ¢i-ni, Rani, ni, ne &e.,—Dhimal and Nipal. In the last it-re- tains the original meaning of sw also, Murmi di-ni, sun, day, Gurung dhe-ni, sum, di-ni, day, This form is referable to the Naga ti-ni. The Ultraindinn sibilant variation current in Singpho, si-ni, may be tho original of the Kambojan ti-ngei (also ta-ngai). Anam has agai ay from whieh it may be inferred thatin the Kambojan, Chong and Ka ¢a-ngai, the root is ngai ; comp: the Horpa nga syn. The Takpa nyen-ti day appears to be related to the Bhot.nyin, Horpa Me-leon one side, and to the Limlm len-dik, Kiranti Jen onthe other. Abor has lorige and Manipuri ka-lhan, lan-la. | Kapwi ri-mik (day eye’. The Kambojan tingei, tangai, (also Ka, Chong), Mon mun tata-ngwe, Koreng ting-nai mik (day’s eye), Luhuppa tsing-mik (id.), Tangkal, Naga, ting-lu, sw, resemble some of the preceding forms for sky and day. * Jn some dialects rang is very much used. Namsang has rang-tung shy, (Muthun rang-han svn), rang-vo light, rang-myak darkness, rang- yi dey, ring-pan_ night, ka-tha-k-rang Yod Grangydling Muthun ), rang- | wok thunder, (Muthun rang-bin air), rang-fom elou ETHNOLOGY OF THE INDO-PACIFIC ILANDS, o7 Light. In the Tibetan vocabularies the root oceurs rarely with the meanin light. In Lhopa nam becomes dam with this meaning, the Chang! dialect sree ngam. Serpa has the slender form -rhip, Magar rap, ' Gurung dh-la, Kiranti u-la-wa cha-m:. The Lepcha aom, ia probably a contraction of ngam. The Aka hang, Sunwar hango, Chepang nughs ap- pears to be a variation of the sillant form, sang, shang &c. Naga has rangai, rang-ro, Burman lang, len (day in Limbu and Kivant:), Garo k-lang, Bodo sh-rang, chu-rang. Lau has leng, Anam den, rang-sany, bo= | jan p-lo (comp. m-ro of the Gurung nang mro sky, Horpa pu-ryu; also jim-p-ro mhite Gond.). Lepeha a-chur light, Limbu thoru, is a form similar to the Bhotian hur wind, and Mongolian a-hur, a-chur, uhr atr ; but it may be a-chu-r, (=chu-rang Bodo), tho-ru. The Chinese word for day may be the same root ngit Cheo-hu; jit, ji, rit, get, yat in other dialects,—swn ngit thioe, jit thau, yat tau, (day’s head). Cryami has re-thou (re for ne, ni) sun, re-yai light, Anam has for sun nhit, ngat, nut Chin. Fire. The Chinese and Tibeto-Burman root for fire is the labial (see 2). But some of the Tibeto-Burman forms of the / root are found in Mon- Anam vocabularies for fire,—pi-lung, p-lung Kambojan, r-lea Chong, lia, lua Anam, ding Kasia. The antiquity of these vocables is attested by the root occurring in the gn Bate other meanings, p-leng sky Chong, b-loei sky Anam, air Mon ;’ light Kambojan; lum, lom air Lau. The Katmbojon form lung is the Bhotian lung air; Takpa dung, Naga nung, tung, Anam tung, shy. Maon, , _ The prevalent Tibetan name is the liquid root,—the vowel broad in Bhot. and slender in the other dialects. See White p.20. The variations are similar to those which the root has with other meanings. Thus the Bhot. z-la-va moon is similar to the a form for shy sa-rag &c., the -k be- ing preserved in Horpa s-lik, moon, Bhot. lhak air, although lost in la, moun. The Bhot, a form is found in Anam, The Lau fam, has the u form, com- mon in white, air, shy, fire &e, In the south the broad Bhot. form iscommon. Aka pa-la, Mru pu-la, Mijhu lai, Maram Jha, Khoibu, Maring tang-la, Burm., Karen, Kami la, Newar mi-la, Sunwar la to si, Chepang la-me, Lepcha la-ve, Limbu la-va, Kiranti la di-ma, Murmi lha-ni, Chango la-ni, Lhopa dau, Gurung lau ngi, Khyeng ki-lau, a The u, o, form is found in Abor po-lo, pa-lo, Dophla po-lo, T. Mishmi ha-lua, tho, Koreng cha-rhu, Toung-thu Ju, Kami lho ; Lau fam. leun, lun Khamti, deun Laos, duen, duen nung Siam, den Ahom. The slender form of Gyarung, Manyak and Takpa is found in Mithun let-nu, Garo rang ret, Tablung le, Khari le-ta, Dhimal ta-li, Changlo la-ni, Ksranti la di-ma, Murmi |ha-ni. Star. The forms for star are similar to those for white. 38 ETHNOLOGY OF TUR INDO PACIFIC ISLANDS. | 2. The Sibilant Meot. Aw, The Thochu mo-zyu air appears to be merely a variation of the common Tibetan liquid root, which has thie similar form ryu in Manyak, the change from the liquid to the sibilant occurring with other roots (see Nemerals 4). The same phonetic change connects many of the Southern words in s wit those in 1, r, d, t. But some of them appear to be connected with the Chinese sibilant root used for star. For air words resembling the Tho- chu occur in Taying M. zhung, the Manipuri and Nipal gronps—M. ma- su, ma-si, ma-r—th (comp. me-r-dah Many.)—N. nam-sn, pha-se, slia—m?, sa-mi-t, sag-ma-t &e. (comp. Ihak, hak). It is found alsoin Milch. hash (Abor asar), The Lepcha sag-ma-t (day sak-ni, swa sa-chak) appears to be also found in Limbu tam-sak-pa shy, an archaic sibilant form allied to the Lambu se- mi-t. The Kiranti hak appears to be referable to the old Bhot. Ihak-p, and not to a sibilant variety. The antiquity of forms like sak is shown lyy the Turkish sok-/a (Yenis-‘Turk.). Comp. also Turk. syod light, or (fire, with Bhotian hod, Limbu ot, Takpa wot, light. Moon. The Thochu chha’ mvon, appears to be an archaic form allied to sak. The Gyar. tsi, chi, of tsi-le, chi—le, is probably a slender variety. Manyak has the broad form in nash-chah day, swn. In the south the Thochu form and application are found in Manipuri, Kapwi tha, Singpho, Jili s¢-ta, Nogeung yi-ta, Khari le-ta, Tengsa lu-ta, Nams. da, Dhim. ta-li, Sak that-to, ” 8. Tang-khul «a-kha, Kamb. ke, Ka kot, Chong kang, Luhi- pa ka-chang, N. and C. Tangkhul ka-cheang, Khoibu and Maring tang- 1. Champhung has a-su-bi (Nicob. ti-so-ab white), Mon ha-tu (cha white), ka-tok; Anam tho bak. The Abor variety with the liquid final, a-sar, is Scythic in form. Comp. a-sar, a-chur, light, (supra.) But it may be a-sa-r (=sa-ra). The same root occurs, in several of its varieties, in the words for sky, sun, day, light and fire. Sun, Day. The Singpho, Jil, Naga, Garo, Deoria Chutia and Bodo tsan, shan, san &c. svn, is similar to sha, sag &e, As day it is Naga, Manipur, Bue- man, Garo, Bodo and Kol, Nipal and Tiberkhad (a-sanga, tsing, tshon, sun &e.). The Naga tsing is reproduced in the Kol sing, Magar sin. For day the Manipuri gr. has rga-sin-lung, nga-sun, a-sun, ma-sung, md-su- tum, tam-lai; Anam song, mang song.; thi, The sibilant and slender form of Naga used for sun, day, current in Kol and Magar as day, is also, with a postfix, applied to, fire in Kol and Gond, singi-], senge-l, sumgu-l (comp. ting-lu sw Naga.) The root occurs in similar forms as white and moon ( p. 19). It%s probable that some of these forms are Dravirian. Comp. white te-Tlz Telugu (na-lla black); light mar-sal, ma-skal Kol, (mar, ma, man min &e, sky), bhok-sha Tuluva; sky sir-ma Kol; fire tu Tuluvs, avha—l Tamil, thi, t-ya Malayalam, chi-4 Uraon, chi-che Malé, singi-/, senge-/ Kol; moon tinga—1, Tamil, Mal., tinga-2« Karn,, Tuluv.; swx singi, sing pay oy ; day sing, sugi Kol; star chukki Karnataka, chukka Telugu, suku Grond. * Anam that tha clear &e. ETHNOLOGY OF THE INDO PACIFIC ISLANDS. aa” The sibilant root is also current for white in several of the Ultraindo- Gangetic dialects. u. The Labial Root. Shy. The labial roo is sky in Thochu mah-to, Manyak mah, and Gyarung (u-mon, tev-meun, This yocable distinguishes the Sifan dialects from Bhotian which has nam, Miri has de-mur, (Gyarung in pref. and root), Singpho, Mru, Murmi ma, Burman, Toungthu mo, Burm. wr. mogh, No- gaung ma-bat, Tengsa phum-ching, Gurung mun. Manipuri has thang- wan, tang-ban, Lungke wan and Lan fa. Light. As light it is Horpa s-pho, Manyak wuh, Bhotian hod, hwe, eu, Thochu uik, Takpa wot, Naga oitite, Luhuppa hor, Limbu ot. The Manipuri wan, ban, ben, war, Karen #-pa, Lau {eWADE; Anam mang sony, Mon/a-ma,do . not resemble these forms but some of the preceding ones for sky. As they are found in Dravirian as light, yal-chom, bela-kw, a-veli, bili &e. (Kol mar-sa, mar-sal)and shy, ban, van, banu &c., these{forms and the similar Gangetic vocable for svn, beli Asam, bela Dhimal, ber Male, and moon no= kha bir Bodo, appear to be archaic Draviro—Ultraindian and not Tibeto- Ultraindian. They are Indonesian,—banua Nias, awan Sasak, Sabimba, wang Madura &e. _ The forms wot, ot are fire in Turkish, air, wind in Ugrian ot, wot, (in Sanskrit at-ia, Armenian ot &e). Sun. As sun it is Thochu mun, (Gyar. mon, Gurung mun sity), Sak sa-mi, (sa-meh jire Manyak), Naga wang-hi, Anam vang hong, and Lau wan, Lan,—the Manipuri forms for liyAt and shy. Air. As air it is Singpho nbong, Naga pong, ma-pung, ma-bung, mong, Newar phai and Sunwar pha-se. Fire, In Chinese the broad form is wind, air, fang, hong c., Gyami sphun, The Chino-Tibetun word for fire is a variety of the same root, Tho- chu, Takpa meh, Manyak sa-meh, Gyarung fi-mi, Horpa w-mah, (mah sky Thochu, Manyak), Bhot. me, Ahor, Burman, Naga, Manipuri, and Nipal groups mi, me, Aka w#-ma, (Horpa). Manipuri, Kumi mai, Lau fai. Bodo wat, Mon ka-miot, ka-met, ta-mot, ta-mat, (Thochu mah-to sky. Chinese ho, fo, fua, hue or hwe, we. The. Mon mot, Bodo wat, have the Turkish form, a3in the Takpa wot light. The root is also an archaic Scythic vocable for sire, abe, apeh, ambe Aino, fi Japan, bi Fin &e. Some of the more archaic forms of the labial in the one applica- tions, as wot, mot, resemble the Tibeto-Ultraindian labial root for white, phok, phuk, phut &c. The Chino-Himalaic labial root is found in both the Chinese and Hi- malaic forms in Dravirian,—day paga-luv Telugu, Karn., paga-/ Tamil,, Malayalam, pagi-/ Tuluva, poga-! Kurgi, pokha-/ Toda; sur paka-d-on Tamil: ; light bokh-sha, Tuluva ; sky magi-lw Karn, anc. The antiquity of” 4h. ETHNOLOGY OF THE INDO-PAGIFIC ISLANDS. this form and its application to the sun,,day &c., are shown by the Yuka- hiri sun bug-on-she, Caucasian sun bok, buk, baak, day bigula, ba ; Malagasy sky ha-baka-baka, Galla sy waka, wak, god wak-wak. - 4. The Guttural Root. The only examples of the root in the Tibetan vocabularies are the Bho-» tian khah, Horpa koh sky, which is Turkish kuk, gok. It may be connect- ed with the Bodo no-kho-rang (no kha—bir moon, Angami kharr moon) Kambojan kor, Mon kya; Chinese light, kong, kuang &c.; Angami Naga ti-khra, ti-khe, Anam khi air, An. ki day, and with the Lau guttural root for white khau, khoung. Affinities of each dialect. 7 1. Bhotian. The form lung, air, has the vowel of Horpa, Thochu and Takpa; and the aoe Se ix Ate an the Takpa dung, Nai nung sky, Kambojan lung ire, p-lo light ke, “a Tie form lhak, air, is allied to the Manyak da’. Similar forms are pre- served applied to the moon lik, le’, da, la &e. The Naga-Gangetic rag, rang, raé&e., aw’, are referable to thisform of the Tibetan root. They appear to have been very early diffused, and indicate a distinct transmis- sion of the no from And one nies Bhotian. Sa Bhotian is found in its proper form in a few of the Nipal languages only. The Bhotian nam, s/y, is also an srohats form. It 4 only found in Tak- pa, and in the south also it has a very narrow range. But the common southern rang‘ is the same reot in another form. : Khah, sky, Horpa koh, has. still narower rrange. Nyi, sun, is also Gyarung and Manyak,—Horpa and Takpa preserving the broad form, nga, lang. It iscommon in all the southern groups, but the prefixes show that thesouthern terms were chiefly derived from Sifan. The forms of the labial root in.its application to fire occur in the other Tibetan vocabularies also. - 2. Horpa. The Horpa pu-ryn, air, is connected with the Thochu and Takpa forms. Koh, sky, has Bhotian Mon-Anam and Chinese affinities. Nea, sun, has also Mon-Anam representatives (ngai). | Nye-le, day, is Gyarung, Takpa and Burman. {Pe light, is remotely connected with the Lau sa-wang, Burman mo sky. Taaie Jire, i3 also Aka. . Like the Thochu and Manyak these vocables show archaic affinities both with the Burma-Gangetic and the Mon-Anam glossaries. 3. Thochu. The Thochu mo-zyu, air, is Manipuri and Spe Muh-io sky has the same affinities as the Manyak mah, that is it is Mon-Anam rather than Burma-Gangetic. Mun, svn, is directly connected. with the Gyarung mon, meun, sky, and, through it, with the similar Ultraindo-Gangetic terms. Styaklo, day is peculiar, Kis Kalombojan and Angami Naga. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 41 Wik, light, is an archaic broad form not found in other dialects., but differing little from the Manyak form, Meh, fire, is the common Tibetan form. This dialect appears from some of these words to be archaie and pecus liar like Manyakwith which it-has some special affinities, and this accords with our previous inferences. 4. Gyarung. The Gyarung ¢a-li, air, distinguished from all the other Tibetan forms, is distinctive of the Burman group in Ultraindia. yaad tu-n.eun, sky, (mun sun, Horpa), is Abor, Burmanic, Murni cavl Gurung. Ki-ni sith, is Yuma (ha-ni) and—with variations of the prefix, such as « sath A ema ian in Ultraindia, some Ultraindian forms being aati JVI DG, Nye, day is Burman, ne. Ti-mi, fire has the common Ultraindo-Gangetic form of the root. The Gyarung forms are connected with the Ultraindo-Gangetic gene- rally; but they have a close and decided agreemeut with those of the Burman branch of Ultraindian, 5. Manyak. Me-r-0a', air, appears to be connected with the Naga forms. The labial prefix isa commou Ni age one, and the compound with r ilso occurs—ma-r-thi air Maring. The root da’ resembles primarily the Bhotian lhak, and secondarily the Naga, Manipuri rang &c., of rang-che phanere, thi-rang, khi-rang &c. Ma, sky (Thochu mah-to) is a link between the Sifan u form and the Yair) Lungke, Manipuri and Dravirian @ forms, ban, wan &e. Taken with the similar forms for light it indicates an archaic connection between the Sifan and Mon-Anam vocabularies, and between the latter and the Dravirian, but no spread of the Manyak form specially. Nyi-ma, sun, is Bhotian. : Nush-ehah, day, appears to be an archaic broad form of the root, as in Jili, Changlo and Sunwar. Wu’, fight, has no close southern affinity. Sa-me’ has the Thochu and Takpa form of the root. The Manyak forms appear, on the whole, to be archaic, and not closely connected with those of any of the Ultraindo-Gangetic dialects, 6. Takpa. The Takpa rhot, air, is found in Maram n-hlat. : The dune of nam-dung, sky, is Naga, rang-tung, and Anam tung-tien (a similar compound). ; A P-lang sun occurs as day in Maram lan-la, Naga rang; as light ia Naga rang Burman lang &c.; and as sky in Chong p-leng. Nyen-ti, day, has no special southern affinity, save with Limbu and Ks ranti, Wot, light, is N and Limbu, Meh’ pi 5 is tho uaceronai Tibeto-Burman form. These words show a close agreement with Naga-Manipuri. Night. : The Tibetan names and their southern forms (niyht, black, blue, green, 42 ETHNOLOGY "OF THR INDO-PACIFIC ISLANDS, ved) have already been given. 1 tshan, sha, shen, sen; sang, sing, ching; zying, hing, cha &c.; 2 spha or s-pha; 3 di, ti, zi; 4 mor, mon, welo, Lee => “ie mo-r, mo-n, wa-lo, mu-la &e.; 5 kwa-ka’, kha-khe, ko-koi, u-k, ko-k. | Other names are also found in the south. : . 6. kamo Abor, kham Siam. i. A ~ 7. nak, mak, nyak, nya, ye, myin &e., black, sep; or as a qualitive with the word for sky, or air. | 8. bik Magar in nam-bik (p. 27). 9. -ma-kung Kami, kung-keng Burm. This is Chinese, kung chung K-h., hung chung K-t. Obs. Two broad forms of the liquid root appear to have existed in the archaic Himalaic vocabulary, distinguished by the final consonant, the first having —k variable to -ng and t, end the second-m. That the m form was not merely a local variation of ng is rendered probable by the | Scythic examples on the one side and the Lau on the other. But the —m like the -k form has produced —ng, —n forms. fe ee 3 A. The -k, (-t) and derivative -ng, —n forms cre . chiefly applied in the Tibetan vocabularies toair and moon. The two meanings appear to have been originally distinguished by the prefixes, but the variations in the form of the root are now sufficient to make it a distinct word in each of its uses, and even in most of the vocabulayies. =. > » The archaic forms appear to have been Ihak, Thuk, lhug, [Dophla haa lug in white]. Their antiquity is shown by their prevalence in the Scytho- Iranian glossaryras;atr, shy, light, day &e, (e.g. a-rak a adil tora licht, light, lux, leukos, logy lug, lok &e, &e.) The u, o, form is still retained im 4outof 7. Tibetan dialectsfor air, (Takpa has it also in sky and it is the most common vowelin white)... if a variation off the liquid to the sibilant takes place in Phoehw it may also be found in the south, where the archaic Tibetan form, for air may be partly represented by the s,z forms. _ The Taying zyung, N. Tangkhul su, Naga yak (=sak), Nipal ro, su, sag, s0, Milch, bash, Abor sar may thus be remnants of the primary Tibetan cur- rent. To it are andosbted ly referable the Mishmi, Abor, Ko Toung- thu, Kumi and Lau forms for moon Iho, rhu, lo, lu, lun, lua, duen &c., and the more common tha, In, Jau. The fina k preserved in Horpa is not found in-any of the southern forms, and, on the other hand, there is now no example of the u, 0, form with this application in Tibet, | Tn Tibet Takpa has the-only example of this form applied to sky, but, it is common in the south} im various forms similar to the Tibeto-Ul- traindian used for air, and generally referable to the same primary current with it. These forms are probably contemporaneous. with the Bhotian The root-does not occur as firein the Tibeto-Ultraindian vocabularic although the labial root is used for light, fire, sky and sun. But some Mon-Anam vocabularies have it with this meaning and in the archaic wu form. The Avam lua fire is identical with the Taying Mishmi lua of kin-lua moon, to which the Siamese luen moon is allied. The Kambojan pi-lung fire, resembles the Lan lun, nung, Abor pa-lo moon, and the ar-: chaie Tibeto-Ultr. pu-ryn, lung, nong &e. air, sky. This is one of these epplications of the common Himalaic roots by which the Mon—Anam_ branch asserts its archaic separation from the Tibeto-Ultraindian. . The slender forms of the liquid root belong chiefly to the laterSifan > ETHNOLOGY OF THE INDOMPACITIC ISLANDS. 43 current to the south. The t-, k~ prefixes show that..one_of its principal sources was Gyarung. These forms are not numerous, and they are chie found in the emasculated Irawady group, the connection. of whieh wi G ng in its late form is distinctly marked by such vocables. _Abor and Aka have similar forms, as in many other instquces.+ dit) B. The archaic -m form of Bhotien, nam sky, which has been commu- nicated to rd is the only Tibetan example of the preservation of this variety of the liquid root for white, p-rom Gysrung, to the sky &c., al+ though the Takpa lang sun, and the sonthern ram, rang &e,.may be variations of it and not of Ihak. The Lau lom, air, has the archaic form, as in the white of Gyarung, Kumi, Kami and Lepeha, The Bhotian form is found in the Nipal dialects as sky and /ight, nam, nap, rap, rhip. 2. The forms nia, (nash) nyi, nyin, nye spplied to the svn and day in most.of the Tibeto-Ultraindian vocabularies are evidently archaic. The root ere to be distinct from that which we have been considering, and’ to be allied to the Chinese. ' Horpa has the only Tibetan example of the broad form, nga sun (unless the Manyak nash-chah be na-schah). The older southern current preserves an example in Anam neai, Kambojan group ta-ngai. The most common Ultraindian form for sun—which has spread to the Nipal group—appears from the prefix to be Gryarung. . ‘ The sibilant root presents difficulties from ita interchange with the liquid. In Tibetit is not current as white, and the only undoubted examples in the present series are the broad form chhah moon Thochu, chah in day Manyak, end the slender tsi, cli Gyarung in moon, star, af In the south the root is still:current with the primary meaning white. Naga cha, the; sang, song, sing, heng; thoh; Tiberkhad chong, Mon chu, Nic. 80; Anam sé, sach, (clear sang, thanh, tot, that tha), Light Singpho thoi, Jili thwe, N. and C. Tangkhul she, shea, — ' The connection between some of these forms and those used in namey for day &c. is unequivocal, e. g. day M. Angami ti-so, (shy-white, night ti-zi sky-black) Khari a-songa; white Nogaung. ta-ma-song. © Anam clear sang, light su sang, su song, jire si sang, sang lang; oe fe owes -rong white), day song, mang song shky-white (also light), sum vang hong (=mang song’). Joboka white che (=se Anam); arrang- che, Mijhu an, day, light, (Anam song), Tayimg sona lights: In sun of Naga rang-han, san, wang+he, wang-hi, su-hih, Garo ra-san, san (also Jan Bodo shyan, Mrang day tsa-lo, (hur-ro night), M. Kumi day: a-hong-nat, sage day sun. In several of these forms the primary qua- pou meaning of white, bright, light; (sky-white or bright &c.) is still yvious. From the rarity of the sibilant element in the Tibetan namés—its ab- sence with the primary meaning whife,—its preservation with that mean-. ing in Dravirian—and the resemblance of the Ultraindo-Gangetic to the Dravirian forms—it seems most probable that the former are Dravirian and not Tibetan, The same difficulty meets us with the labial root, which! is both Tibetan and Dravirian. In both cases.too, archaic Himalaic forms: similar to the Dravirian are preserved by Thochu, which in these, as in\ many other roots, separates itself from the other Tibetan. dialects, andy . > In names for star some slender forms oceur hi-vek Mru, me-rik.. Namsang, le-thi Muthun, le-tsi Joboka &c, (le, le-t-lu, le-ta, moon, comp. tai-le moon Gyar.) . ; bias Peta 4a w= 44 ETHNOLOGY OF THK INDO-PACIFIO ISLANDS. connects itself with the Mon-Anam and Dravirian. Its mah sky (alsu Manyak), styak-lo day, chha’ moon. hada star and phyokh white are iar, and both phyokh and chhw are Mon-Anam * and Dravirian. tis possible that zyu air Thochu is a radical sibilint and nota variation of the liquid of the other dialects. The Thochu sibilant series may be ¢chha’ moon, styak-lo day, zyu air, all based on an archaic sibilant for white similar to the Draviro-Ultraindian; Styak resembles the Nage sak, sag, the independence of which from the Tibetan liquid hak (p- 38) is further supported by the Anam sach, The Anam se, sang, sach, thet, tot, su, thanh, thi &. may be compared with the Dray. te, thi, ti, chi, sha, chha, za, tu &c. In many of the Ultraindo-Gangetic vocabularies both Dravirian and Tibetan ingredients are found in the same or in closely allied names, Thus in the Garo lam-par air, lain is Tibetan and par Drav. Naga has the Ti- betan ngl, nyi-ke, in day, and the Draviro-Anam san, han, hi, sa &&c. in sun. amsang has the Tibetan rang for sky in its compounds, w =k has the Draviro-Ultr. wang, vang &c. 4. The labial root is evidently one of the most archaic of the formation in its use for white, light, fire, sky, sun. The radical vowel appears to have been u, 0, and the final cons. k, as in the Thochu white. The Gyarung variation mon sky (Thochu mun sw) is the form to bh ey. of he See Sk are referable For abi ee ap Gag ae urung &c.,), and the Singpho-Naga pung, mong &c. aer is the same Variety. Lau, Anam dnid Bodo pl a a final tin fire (mot jire An.= mogh sky Burm.). The forms ban, wan, beli, ber, bir, sky, light, sun, moon, appear to be Dravirian. They are chiefly found in the older southern vochs. Lau, Anam, Mon, Manipuri; Bodo, Dhimal, Male, Asam. ‘he Lau and Anam have close affinities with the Manipuri and Yuma vocbs. (sky, light, fire). Tn addition to the forms given above, Pallegois’ Siamese Dict. supphes vela day identical with the Dhimal bela sun, Drav. bela light. The Dray. broad form for sky van-am, man-am, ban, banu &e. (van- main star) pon-e, [also vin, min ; fire ben-ki &c. &e.] is very common in those Ultraindo-Gangetic vocabularies in which Dravirian and archaic Tibetan vocables are found most abundantly. Anam has mang song day, light, vang hong sua, in which the sibilant is white, car, right é&o., and mang, vang, was probably sky or air primarily. The Lou fam. has sa-wang light; ta, fa-fon, sky ; fai fire ;} ban, wan, ta-wan, kang-wan sun ; ban, wari day. In the Manipur’ group we find air phan-ra Champh.; Jire, wai general (Lau fai); Light ban, ben, war; sky tang-ban, Kapwi, thang-wan Khoibu, In Singpho we have jre, wan, Jili ta-wan, ( Law sum in Mijhu Mishmi fire mai ( Manipurian, Lau; in M. M, ai is @ com. : }. In the N oup Mulung and Tablung have generally weng where Namsang, Matkun and Joboka have the Tibetan —sun wang hi, wang he, (sy, white), night vang-mak, vang-niak, (shy, blagk), god kah-wang, lightning wang-lp, thunder w hung, air yang, cloud wang, Garo has wal, ver jire (war light Manipuri gr.), lam-par air (Dhimal bhir-ma, biri~ma the slender Drav. form); Kumi ku-wang light, Khyeng a-wa light, Lungke sky wan, wyn. * The Anam phuck white was omitted inp. 29. It is distinct from the modern bak, of Chinese derivation. + Acom. Scythic form, pai, wai, &c, ‘ ‘ETHNOLOGY OF THE INDO-PACIFIC ILANDS. 45 5. Some of tha guttural yocaljles appear to be the contracted forms of the liquid root with the guttural pretix common as white and star, ka-r, q-re, k-va’ Tibetan. Khyeng has khro maoa, i.e kh-ro, The Angami term may be kha-r, Kambojin ko-r, Mon and Kyau /-ya (for k-ra, comp. k-re, k-ye star). Angami #i-ra air, Gurung has pi-ra star *, (bh-la light, m-ro air, ta-r—-hiya white), Ja The southern guttural roots may be merely variations of the dental, Comp. Anam thi, ki day, khi air. These forms and the allied ting, ti of Ultraindian are Seytho-Chinese and not Bhotian in their affinities. Ch, tin, tien, tsany tien &e. shy; ti ki, hi &c. air. 6. It is probable that the several applications of the same root were ar- chaically distinguished to some extent by descriptive words added. In Tibet there are very few such double words, Bhotian and Takpa have compounds for sty, Thochu and Manyek for day, and Gyaraune for moon, In the south they are very common. In many cases they are merely two synonyms, but in some one of the wordsis descriptive. Day is frequent- ly merely sun; bat sun is eye of the sky or day. Anam has mat ¢-roi sun, (t-roi sky, t-rn day, mat eye); mat nhut sen, (nhut dey); mat t-rany moon, Hie iaet tho bak moon (su song light, song day, sang clear, tot clear, se white; thog white Milch., su Nic, &c.; bak white), The Manipuri group has several examples of names for the sun similar to mat ¢-roi, e. @. ri-mik, tsing-mik. Mijhu Mishint has le~mik. Tayiag M. has ring nging (ning sky, ringa Mijho fwee). In the ta-ngai of the Kambojan group te is perhaps eye and not a mere det. pref. The Chinese jit than Xe. is head of day (day head). 7. The Himaluic vocables that have been communicated to the Vin-~ dyan branch of Dravirian are the Male sa-range sky (Nipal sa-rangi), | and jim-—pro, or o -roa white, aud tha Kol hoiyo, hoyo aw (Anam hoi, hoi gio, unless the Anam name be of Kol derivation). External Relations, ‘1. The Himalnie liqnid root used for white, air, sly, moon, star, ia also an important one in this class of names in the Seythic, Indo-Huropean and Caucasian families. Itis also found in Dravirianin Scythic forms, Some full and broad forms are preserved more largely in Himalaie-and Indo- European than in Scythic, owing probably to the assimilative and slender phonology having made less progress in some of the dialects of the former than in most. of the latter. 2. The Himalaie labial root has several distinct forms, 1 phyok, phiuok, mogh, wot, mot &c,,—2 mun, mon, mur &c¢.,—3 me, mi &e. Nach of these has a wide and archaic range of external affinities, Chinese, Seythic, Cau- easian, Indo-European, African. All are found inthe Seythie glossary in forms similar to the Himalaic, e. 2. fire bok Yeniseian, + wot Turkish, a-ba Aino, bi Ugrion, Japan; sen bug-on-she Yukahiri; mraing an-bok, on-bok-sy Yenisein, sit-bag Torkish ; air, wind wot Ugr.; white bag-da-rin, wag-da-riz Tungusian. Caneasian has bok, buk sun, but, bot-so, moot-z, mo-z &c. moon, mu-ch sky, air ft. These forms from their distribution and rarity must belong to the most archaic eraof the Scythie glossaries, The prevalent forms of the labial are similar to 2 and 3 of the ‘Himalaie. * But it may be Dray. pira (—i-pil an : + Enrop. fok, fogo, fo, a-fo &e. ch Lat. foc- ws). } Vapour Turk. bugu,bug, Armen, bug, Tanil bug-ei. ~ 46 ErYVvoLosy OF THR INDO-PACIFIC TanaYns. and they are rare in comparison with the dental, sibilant and guttural roots. © The pure lebial for fre, mo, me, mi, be, bi &c. is not @ late variation of mnok, bok &¢., but an archnically distinct and very widely distributed variety. The Himalaic forins phuok, mogh &e, are immedi- ately connected with the archaic aod nearly obsolete Seythic buy, bak; and the n of man &c. shows that it is probably alocalsottening of forms like muk, and not a modern derivative from the prevalent Scythic forins of the labial which have a,—bar, wal, maré&e. The Sokpa wun-dur day com- bines a Tibetan with a Mongolian name, The more archaic labio-yut- tural formsare now found, notin, bataround, the central Scythie provia marking the older migrations. inthe north they are found in Yukahiri and Yeniseian, in the south in the Himalaic glossary, in the west in Cauca- sian, Indo-European and African. 3. The Draviro-Ultraindian forms of the labial root, ban, van, ‘wan, mar, bar, val, pal, ber? pel, vel, vil, kc. &e. are connected with the pre- valent Scythic and Caucasian forms, The Arian and North Indian yvoca- bularies have similar forms, and the directly western and non-Tibetan rela- tionship of the Draviro-U\traindian and Draviro-Australian group of forms and vt is us fully evidenced as any archaic glossarial induction can be. Caucusian, moon bar-s, ber-s, ba-z, mi-s, ms-se, Georgian m-t-ware, ¢- wa-i, Ossetic mi-i; sue -bar-ch, bar-ke, mal-ch, marra, beri, Georgian mitli, mixh-si, mik. The application of the labio-liquid root to %Roon is one of those numerous vlossarial links between Dravirian ‘and Caucasian (and even Caneaso-Semitic and African &+.) which, with its non-Tibetan ideologic and phonetic traits, indicate that its course from the Scythic rovince to India was across the fndus and not across the Himala virian has ir-ei, za-hilli, bil-ye moon, identical with the native rodt for white, light, bil, vel, pel, bal, bol, and different from all the Chinese, Tibeto-Ultraindian and Scythic names for moor, The western affinities of the labial name are also illustrated by the Indo-European mona, man-k, mond, maan, moon &c., African vola-na, berra &e. Scythic: air, wind, Ugr. war-ma, bar-s-ka, mar-d-em, mer-d-ish, mer-z, wire, Sam. bil, mer—ga, mar—k, mer-se, bar-shi, Korean par-an, Yeuis. pai, poi, boi, pei (Ostiak woi; Fin, day poi- wa, pei-wu &e. sun pol-wa, pei-we &c.); sua Fin wal-ky; fre Fin wal-kéa, Koriak my], mil he-rnil, mil-gan, mil-gu-pil &c.; white Fin wal-gi, bal-gie, wel-, ; &c., Ugr. wai-gam, woi-kan &e. Here we find the root with its origina Meaning white preserved in bin and Ugrian as in Drayinian, although losg. in most of the other families, and the special archaic connection between ian, Caucasian and Dravirian illustrated. The Indo-Buropean weiss, wit, white &, is probably a variety of the same form, r passing into 3, t,, in the Scytho-Iranian phonology. The Scytho-Caucasic bar-s wind Ugr., moon, sun Cauc., is light in Armenian bar-z( Sanskrit bha-s), as in Dravi- rian, mar-sal, bera-chi &e. Indo-European has other forms and applica= tions, e.g. air wad Ossetic, vata Sansk. wat-em Zend, bat-as Bengali bad Hindi &c, &c.; vent-us, wind, wan, win &e. jire fuir, vuur, fire &c. Semitic air a-war, a-wiru &c. 4, The sibilant root is the most common in the Seythic glossaries ; and in the Caucasian it is equally important with the labio-liqnid. Scythie : white a-sho, zag-a, chaga, sak-rin, sai-ran, sa, siri, sir, shora, sor-ny &c. &c.; light achik, sir-dik &c.; fire tog, tug-ut, tat, tol, tul, tuli, tuz, tut, ETHVOLASY OF THE INDO PACIFIC ISLANDS. 47 San, tm, tui) shu,’ siu, &e.; sem shi, chat, cham, hai, shun, shup—dy, sinna, siig, chotal, tir-/?-tir; day-shi, chaya, tel, e—lur, doh, tua, chel, shan-da &.5 moon Yenia. tai, shui, ehe—p, cha-ip, Muar. sara, chara, sira-2 5 shy Yenis. eis, osh, es do. air sal-ki, chil, sel, tol, tyl, &e. Cwneasian : jire za, 20, zi, tue, m-za, m-xe, (Ossetic sin, din); sun m-se, b-sha, Georgian tuts; day dini,deui, den; shy m-teo, chi, mi-chi, p-chu, p-sha; ster za, Zon, su-ri, su-ta, te-ru &e. In the Indo-Karopran family Scytho-Cancasian forms are common sunsur-ya, heli,sol, sonne, zon, sun; day dina, den, dan, dies, dit, det, daz, tu, day dec. They ave also common in the Semite-African glossaries, e. x. sun shom, shenys, tani So. | The exeeptiowsl Thosha ebha’ mova ( Manyak nash-chah dy) ia Seythie, cha-ra Mong., cha-/p Yenis, | ‘The Deaviro-Anam forms appear to be partly western from their Cau- caso-Seythic and indo-Kuropean aifiaities, and partly archaic Himalayo- Secythic. Some ofthe Anam torms. e.g..sachor sak, that, tot, are more Scythic than Dravirian. The Draviriun and the similar Gangeto-Ul- truindiaa, aguin, have a close resemblance to Caucasian and [ndo-Eu- ropean forma. Sone of the Anam and other similar Gangeto-Ultraindian forma closely resemble Chinese as well as Seythic forms. Probably in this as in other cases, the early Himelaic current—Mon-Anam—brought » archaic Scytho-Chinese forms, and bleajled them in the mixed southern glossaries with the archaic Scytho-Caucasian and Scytho-lranian of the native Dravirian family. Water. River. Blood. 1 Bhotian WT. * chhu ee g-tsang—po ae kh-rag, kh-rang is Rds chha .. chang-po .. thak 2 Horpa gee Bea. oy hra’ ae sye se’ 8 Thochu os ee cha-bra’.. sa” 4 Shed 3 «. iechi .. ti-chi .. fta-shi 5 Manyak mer TEE VIA | ap dya’ ie shi’ 6 Takpa e. chhi oe ehhi ea * kh-ra 1. The sibilant root. ) Water. Blood. Chhu; tsang, pg chah, cha, sha’, sa’; chhi, chi, shi, sye, se’. The Sifan vocabie for blood is Chinese; and itis also the prevalent Ul- traindy-Gangetic one. Chin. chiue, hint, hue, he’, Gyami sye; Horpa syé, se’, (ryarung ¢é-shi, Thochu sa’, Manyak sha’, he forms curreat for water are also Chinese and Scythie. Chin. chui, shui, sui &e., Scythie su, shui, tui &v.: Gryaruny ¢-chi, Takpa chhi, Thoehu chah. Que of the most prevalent Ultraindo-Gangetic forms for blond adheres to the Chinese form tor water, shui, chui, sui, tuip—Tiberk. shui, Deoria Ch, chui, Bodo thoi, Burm., Yuma swe, thwe, Karen thwi, N.'Tanghal a- su, Nogaung q-zu. The Garo kan-chni, Singpho sai, Jili ¢a-shai, Koreng ta-zai, Yuma sai, chai, 'Tengsa ai, are moditications of the same form, As water the full Chinese form is preserved in Kumiand Kyan tui, toi, Khyeng tui, tue, Bongja, Kuki tne, Maram a-thni, Koreng tu-dui, Song- pu dui, Bodo do, Khatha, Maring yui. Nogaung has tsu, Khari a-tsu, ‘ " 48 ETONOLOGY OF THE INDO-PACIFIC ISLANDS. ‘Angami dzu, zu, Tengsa tu, Namsang jo, 8. Tangkhul tu, C. T. tun-da, N. T. ai-chu water; Khari a-tau, Nog. tsu-la-tsu, Tengsa tu-la, Joboka swo-kha, Muthun shoa, Namsang joan river. The Changlo and Lepcha vi, Chepang wi are perhaps referable to the full form shui &e. ‘ ‘ The Bhotian form for water, chhu (Chin. chui), resembles the Naga and Manipuri tsu, zu, tu, chu, &e, Abn Le The slender Ultraindo-Gangetic forms ma-chi T. Mishmi, ?-si Aka, a- ‘si, a-he Abor, thi Karen, ti Lungke, Muthun, Joboka, Chep., Milch., si Mulung, tei Mrung,, x-tsin Sinepho, m-chin Jili, i-siny Manipuri, resem- ble the Gyarung chi weter, shi blood, Takpa chhi water. ‘ This form is also common as applied to blood, hi Khoibu, Marine, Ne- war, hi-ki Bodo, he Namsane, if Mulung, Tablung, i Abor, the Sak, thi Kapwi, Kumi, Kyan, ¢-ji Muthun, Joboka, a-zvi Maram, a-zi Chu mphune, a-shi Luhuppa, wa-si C. Tangkhul, wn-hi Moz. Angami, a-thi S. Tanc- khul, Kami, Kumi, Lungke, Khyeng, ka-thi Khyeny, w-si anwar, *dbd Garo, ma-khi Limbu. ; ~ The Kambojan chi-em, Mon chi-m blood is the Sifan-Ult:aindian form with the final labial sornetimes found in Mon-Anam yovables, when it is absent in Tibetan (comp. bird chi-m, shi-m, chi-va). It is probably con- nected with the Singpho chin, tsin. he root occurs with the same final or postfix in Aino ki-m, and Tungusian sho-ma [See also River], — The Ultraindo-Gangetic shui, tui &e. preservesthe full Chinese form now lost in Tibet, and probably marks the oldest Tibetan current. The tu, su forms are allied to the Bhotian, and may indivate a second current. _ The slender Gyarong, Manyak and Takpa form appears to be that ia which the lust of the great Tibetan currents diffused the root to the southward. It is possible that all the principal yartatious may haye been independent- ly formed in each of the Tibeto-Ultraindian provinces, and that the slender forms many be of equal antiquity with the broad. But the broad forms are common to Chinese, Seythic and Tibeto-Burman, while the slender have a more confined range, similar to that of many other Sifan roots and varieties. Itis probable therefore that all the Tibetan forms were original- ly broad like the Chinese; that the Sifan chi was a contraction of the archaic Chino-Tibetan chui, or an imported Seythic form; and that the Ultraindo-Gangetic chi, khi, thi, ti, &c. in general belong to the Sifan-Ul- traindian current, or to the same phonetic era. The Irawadi-sutlej chi, shui, tui &e. were probably derived from the pure Chislese forms once pre- valent in the Sifin province. River. The first series for river has the common Chinese, Scythic and Tibetan root for weer in a Scythic form. The Scythic forms are shur, shor Ugrian, su, sug, |o-suk, sea, Kemach.], w-sun &c. Turkish, w-sun, chun Mongo), The-same form is found in Anam saung, sung, song, som, Kamb. tun-li. Bhotian and Thochu have @, Bhot. g-tsang-po, chang-po, san-pu, Thochu oha-bra’, A West Bhot. yoo. bas muk-sung, _ The Murmi syong, Serp. hyung, Lhop, chhu kyone, Lepeh. ong—kyong, (ong watvr), Limbu wo-hong, Kiranti hong-ku, Gurung kliwong, Luhuppa, Fungkhul and hhoibu kong, Koreng shing-gu resemble the Anam sung, song,—but trom the Manipuric form gu and the Nipal kyong, khwong, ku, it is probable that the k form is a distinct root; Anam has kang. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS« Av ‘Thesoureeis the Chinese kong K-t., kiang K-h., which would thus appear’ to have spread from N. Ultraindia to Nipal. ; The Pashtu sean, sin, sin-t (whence Sin-d, Hin-d, Ind-us, India) may be. Tibetan and not-directly Scythic. ? | The STAD sun, sug, sung is a common Indonesian word, sungui, &c. * , The ‘sibilant, in the common forms for water, is also river in Gyarung and Takpa. Garo has chi (ti-chi Gyarung, chhi Takpa), Abor a-sie, Newar shu-si. _The Chinese and Yuma full form for mvwter toi, tui &e., is current as river in Bodo doi, Manipurian dui-dai, tui-koak, tu-thau, tu-lil, tu-rel, tu, Limbu chua &e. 2. The Liquid Root. hra’ Horpa, dya’ Manyak, water, river ; kh-rag, ih-rang, th-ak (for: th-yak) Bhotian, kherea Takpa blood. ; » \ The root of the old Bhotian hp Takpa kh-ra, Sokpa kho-ro-gyve (? kho-rog-gmwe), blood, is found in the Mileh. pu-lach, Nieobar ka-nak, forms which lave been derived from a very archaic Tibetan éarrent. The tag rak-ta, Sindhi rat, may be referable to it, The root occurs in the Ugrian wug-rak. The softened current form th-ak, Lhopa th-yak, Sarpa th-ak, have made no progress. The Anam tiet blood resembles the Kambojan slender form for water, tik, (tak &c.). The Anam form may bea purely local variation of Chinese. As water the root is Mon-Anam dak, dat, tak, tag, tik, nuk, in Nankowry rak (Nie. ka-nak blood), Bengali u-dak, Newar lau, la, Jtiver tu-la, tsu-la-tsu Naga, ha-loung Khyeng. The form in rk, dk,nk &e, is river in Ka dak-tani, Kambojan p-rek, Burman m-rik, m-riet, m-yit, Khyenglik,Mon water dhik. This slen- der form is distinguished from the broad rak, dak &c. water, blood, and is similar to the Rakhoing re, ri, Changlo ri, Tablung ri-ang, Magar di, water. The same slender form without the final / is also found in words for river, tun-li Kambojan, tu-lil Manipuri (two roots), li ku Sunwar &c, Burman has alsoa broad form g-rach, + In the Thochu cha-bra’ the last element is probably a similar vocable b-ra’,?m-ra’; comp. bra’ Horpa. he Sambawa b-rang is the Burman m-rach softened. A similar form of the root occurs in Mikir lang water, lang—pi river. The root occurs with the labial final in the Lau fam. nam water, river, oil, Kasia s-nam blood. he same root also occurs in Tibeto-Burman names for o#/, in some cases by itself and in others with the name of the plant from which the oil is obtained. Bhotian has /-bru-mar in which bru or b-ru appears to be an archaic form of the liquid root for water. Horpa has mar-nak, in. which the liquid geot (hra’ water=rhak) has the Nicobar form. has kya-mar, in. which kya appears to be au example of the guttural, root, * Sung-ai water’s-father may have been an archaic Tibetan form, . ai father, Lhopa, Thochu, Ugrian, + This form is found in the name of the river Barak (Ja-rak), In Marco Polo’s time the Irawady appears to have been known by the slen- der form of the same yocable, Brius (d-rius =m-riet, m-rik). e , 59 - ETHNOLOGY OF THE INDO-PACIFIC IsLANts. The Thochu ching-yu aKuag to combine a Tibetan name for water with ” the Chinese name for oi/, yu. (Gyarung has chin-swi (chi water ), Ma- nyak i-chi-ra, %-ti-ra. | The sp. Bhotian num oi! appears to preserve the Lau form for water. The Lepcha nam oil is identical with the Lau form, : ; 3. The Guttural Root. Most of the guttural forms appear to be variations of those in ch. The Chinese kiang. kong, khoi &c. may be radically the same as the Bhotian - chang, tsang, hut it is more probably connected with the Scythie guttural root. A few of the southern words appear to be directly. connected with ee eattenn & i Maring, kyong, khiong, hot ce iver, kong Luhuppa, kong-pvi Maring, kyonz, khwong, hong, Nip: River, kyi Toun rth kha Biomebne tui-koak Kapiti eta cis ta-gha M. Kumi, khe Ahom, khye nam Khamti, unwar liku, Magar kho-li, Newar khu-si, Kiranti hong-ku. Water, kwi Murioi, kyu Gurang, pan-khu Sunwar; chi-ka Garo. Blood, ma-khi Limbu, ka Murmi, koh Gurung. hii Newar, hyu Magar, hau Kiranti. Qil, ma-khu Lhopa, chi-gu Murmi, chi-kang Newar, chu-gu Gurung, eyo Sunwar. | 5. The Labial Root. Mon has the labial only, bie, Ps river, and it is also current for water in Changlo vi, Chepang wi and Lepcha vi. In Lepcha yi is also dloed. Kiranti bas awa ot/, For sea Mon has i, bui, Anam bien. The N. Dravirian vocabularies have a peculiar form of the labial root, mater am Male, um Uraon. Kasia has also um and for river wah. These forms are distinct from the Southern, but the Scythic original both ofthe Toda pa and Northern am, um, is preserved in the Pashtu aba, abu. t A labial root occurs in several names for river, Tt appears in some to be the root for father or mother. The Lau fam. has me nam, nam ine, mo- _ ther of watcr (me mother); Mikir lang pi (lang water, mi mother, Kasia, — pi, pe female Mikir), Sak pi-si. The broad forma, generally gether, is found In Kaori ka-va, Kuri ka-wu, yang-pang (=rang-pang), Mru a-u, Lung- ke fi-wa, Mrung tei-ba. Kyau ti-poe, Karen has ti-mo (water's mother) Miri a-bunge, Aka su-bang, Sunwar.pan-kha, In some of these forms also the lalial only its retained. * Other roots for mother are also used. Mulung and Tablung have yang- ° nu (riang water, nu mother), Nams, jo-an (jo water, anu mother Muth. 'The name for the sea is sometimes formed in a similar mode, Burman has peng-le ( father of water), Khyeng pen-le, pan-lei, Tungthu pin-lai, Obs. 1. The ui form of the aspirate root—the mostimportant of the. Ultraindo-Gangetic and found as water and blood in the older Ultraindian up, Yuma &c.—is probably en archaic Tibetan form, as it is preserved 1 in Chinese and Scythic,—water Sam, tui, sea Aino a-tui, a-dui, a-zui, '* "This idiom ia Scythic, e.g. Turkish aga-n-su, yei-su, ir-mak (water su Turk., e-mak Chukchi, father aga Turkish, ai Ugrian (Perm), mother ila Koriuk), . > ETHNOLOGY OF THE INDO-PACIFIC TELANDS) \-* 5 The aivariation is also Scythic, but in the Irawadi tongues it appears to _ be a local variation ofui. The short forms tu, tsu, su, appear to be also local, although similar varieties occur in Seythie vocabularies. -A singlo. . form imported in one Tibetan vocabulary may have originated all these southern variations of the broad form, j eal 2, The slender southern forms are mostly connected with each other, and derived probably from the Sifan (Gyarung-Takpa) similar form. Its alence in the northern Irawadi-Brahmaputran vocabularies, Singpho, ishmi, Abor, is strongly in favour of its immediate derivation from a’ southern Sifan didlect; and the extent to which it has been diffused amongst the Irawadi-Yuma vocabularies—especially in its Gyarung appli- cation to blood—render it probable that it was the vocable of the lutest great Sifan migration, the Gya:ung character of which has so often-been noted. 3, The forms in n have a much more limited range. But their appli- cation to river, their close Scythie affinities, und their occurrence in Bhotian and Anam, show the form to have been arehaically distinct from chui. The Dravirian roots for water, river, blood, are distinct, and their line of . connection with Scythic roots is also distinct. 1. water, niru, nir, nillu, tiru, Brahui dir; blood se-nnir (red-watir, se-yya &c. red), u-dir-amy, ke-nniru (ke-mpu red, ke=se); Comp. also cho-ga-naa, cho-ndad red, cho-ra, cho-re blood, —and the Uraonand Male khenso, keso red, khens, kesu blood. 2. water, pun-al, vell-am ; river varu-pun-al, aru, eru, yer, polo, hole, pa, puzha (=pula). 3, water tanni; river tude, don-da ; b ne-tturu, ne-ttar, na-ttur; the root tur, tud &c. is probably a variety of tir, dir. The Kol and Uraon name for river adds a 4th root, garra Kol, khar Uraon. Itis found in Angami, karr, kerr, and Chepang, ghor. The Tibeto-Ultraindian liquid root is found in Kol dah, dha water. A slender form of the aspirate che-ii water has been received inte Uraon, in addition to the nativelabial, um, ‘Foreign Relations, The Sibilant Root. This, as we have seen, is Chinese, water, and Scythic, water, river. The archaic broad form appears to be Chino-Himalaic. A similar form is found in the more primitive Scythic vocbs. Most of the formg for river are Scythic in form, the older Himalai¢é vocb. of Ultraindia preserving, the ure ie variety sung Kc, (Mongolian, Turkish), The sibilant and enta) root is the most important in the Seythic vocabularies. In the older it has the labial prefixed. Vin we-si, Ugrian wi-ty, wi-t, vi-s, Japan mi-zu, Aino river be-zu, be-z, water wa-z-ka, wa-cha. The root is also found in Ugrian forzasin the Caueasian tzun, shin, che, p-si, p-seh &c. water ; chi, p-si, chor &c. river (Georgian wi-tz); zi, tsha, cha &c. blood (also bi)» Indo-European has wa-sse-r, wa-t-er, wa-da &c. The Liquid Root. The primary forms are rak, rat, lak, dak, nak, nuk;—nam. The root * Insec. 11 of ch, v. the Ultraindian tara is compared with some of the Dravirian forms. But it is now clear, from the collation of the Ti beto-Ultraindian names, that the root is ru (ta-ru), 52 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. is rare as mater. In the great Scythic alliance the only analogous forms in Klaproth’s vocabularies are the Yeniseian dok, Fin ta-tze, za-tse, which however appear to be referable to the sibilant root. The Lau form is pre- served in the Tungusian word for sea nam, lamu, Ueriag has. lei, a-ner, a-nyer river (Wolza); Mongolian nuhr, nor sea; Turkish nehr; and with, the prefix ta-,da—both roots are used for sea te-nger Hungarian, to-nyar Tungus., ta-lai Samoiede, da-lai Mong., cplt dake, de-nyis Turkish. A common Seythic name for sea and river also contains the liquid root, ' mnu-ren, mu-ran, mu-ra, mo-ri, wu-di, mu-ny, mu-l, The Seythic roots also occur in the vocables for blood, milk &e. The root occurs applied to bload in ut, leut of the Lan fam., in the Aryan rak-ta, rat, ru-dira. It occurs in similar forms in African and Asonesian vocabularies (water, river, blood), Possibly the Indo-Luropean lac milk may be the same root, but it is at preset too sporadic to take its place within this circle of affinities. | This root appears to be one of the primary ones of the Tibetan family, and, from its rarity and the nature of its distribation in the Aso-African glossaries, to have been one of the earliest dispersed. It appears to belong to the western or Ugro-Iranian side of the Tibetan affinities and not to the Chinese and eastern Scythic. The Labial Root. This root is common in Scythie, Ree ae Caucasian, Dravirian, and Semito-Atrican, as water, river, sea, blood, &c, But it is doubtful whether inthe full forms, in which it is followed by the sibilant, dental, liquid or gut- tural, it is to be considered as the root oras a prefix. = forms mu, W bi &c, are comparatively rare, and, in general, evidently contractions o those which have a second element. Thus both bi and zi blood of Caucasian are referable to the Ugrian vi-s &c. water, wi-r &c. blood, be- cause, from the prevalent forms for water and river, the vocabulary is clears ly Ugrian (e. g. water p-si Cauc., wi-tz Georgian). In Ugrian the same contractions take place, e.g. water Ugr. wyut, ute, uit, wu, wa; Samoide bi, it, i, bu &e. The Caucasian vocabularies appear in this, as in other cases, to be immediately counected with the archaic Ugrian and not with the Tibeto-Burmun. The eon Toots. * aimee toa ~~, range of mutations in every considerable group of vogubularies, as in ey East Scythic, Tibeto-Burman, Asonesian, Onnealian and Semito- African. In the Ugrian branch the labio-liguid is the common vocable for blood Fin wuo-rak (the full Tibetan form of the liquid), wa-r,-we-ri, ma-le, (lei« pe); Ugr. wa-r, wi-r, ille, &c., Koriak mu-lu—mu-l. | Sanskrit has Va-rl, Wa-rl water. The Guttural Root. The guttural is Chinese, It is also a distinct Scythie root, kai, Ad Bede kus; gol, kul, kol, gun; Semite-Libyan khar, khor, kol &esj. Earth. Mountaw. Stone. arth. Tibetan, 1, sa Bh., k-cha Horpa, se’ Gyar., sa’ Takpa; zi-p Thochu, ETHNOLOGY OF THE INDO-PACITFIC i3LANDS. 53 . 2. ma-li, m-li Manyak. : Southern. . 1, sah Lhopa, sa Murmi, Gurung, cha Newar, jha Magar, n-tha Ma- ram, ha Namsang, Joboka, Bodo, Garo, ha-wan Muthua, ya Deoria Ch., han Toung-thu. [Mountain 3). 2. th-li, ta-ri Taying, a-li Tenrsa, Nogaung, Khari, kan-li Songpu, ka-di Koreng, ni Kyau, le-kong M. Kumi, w-le Lungke, m-re Burm. wr., m-ye sp, nhe Girung. ; ‘ ta-lai Kapwi, tha-lai Khoiba, nga-lai Champhung, Luhuppa, C. Tang- khul, &-lai Miriaz, ma-lai N. Tang-khul, lai-pak nipurl, 4@-lai-hong Kami |nai Chinese, Kt]. : noi Anam, nyai Mijhu, bho-noi Dhimal, (nui mountain, noi, loi, doi hilt Lau fam.) [nai Chinese K-t]. lang—nin Khamfi. a-lu 3. Tangkhul, Ru-long K. Kumi, (ka-lung stone M. Kumi), 2-roung Mru, ruong Anam. dong, (also mountain) Anam, dung country Mon. 3. a-mong Abor, muang Lau, Siam, phuong, huong Anam, phat Lep- cha, lai-pak Manipuri, ba-kha Kiranti, bho-noi Dhimal, ha-wan Muthua, Khamti, phen-din Siam, mien Anam, 4. ka-tok Mulung, Tablung; tho Anam [Chinese tho].’ 5. ki-ju Ang. (the-ju iron), su Anam, ki-je M. A, (je won). [See Mountain 5). . 6. ta—ka Jili, nyga Singpho, ba-kha Kiranti, kham Limbu, kho-pi Sunwar, ka Sak, ho kho Sea Karen, gon kho Pwo K., khon Anam. khe-khel Uroan, ke-kal Male. . 7. dia, dia phan Anam, deiye Kambojan, (Ch. ti, dei, land ti, tien]. 8. dat Anom, det Kkyeng, te Mon. o-te, o-t, wa-the Kol. 9. din Siam, Ahom, lang-nin Khamti, phen-din Siam. This is pro- bably a variation of li (2). Mountain. Tibetan. 1. ti Bhot., T: ri—rap Horpa, | Harth 2, Stone 2). Chinese lia, ling Tungusian a-lin, u-ro, w-ra; Mong’. w-la, a-la, Fin war- 2. s-pyd Thochu, ta-vet Gyar., m-bi Many. ° . Southern. ; 1. [Kurth 2, Stone2). m-lo-di Dophla, no-di Aka, ma-long Koreng, (ta~ lo earth), ka-long Maram, kh-lung Maring (also earth), kho-lon Pwo Karen, rong Lepeha, dung—kang Tiberkhad, dong Anam, rok Lhopa. k-lang, slang Lung-khe, A-lang Kyau, mu-ra Doing-nak, mina-ram Nogaung, ra Dhima!, - Jai Mrung, loi Laos, doi Ahom, noi Khamti, nui Anam, hill, non, thé non, thi don Sm, non, non sanh, ngan Anam, pi-nom, ohare Sgt ka-phung N. and ©. Tangkhul, 2. bom, bum Singpho, ka—-phung N, and C. Tangkhul, Champhung Luhuppa, mue K. Kumi, a-pih Khari, min—a-ram jer moi Trae pa-awon hell Mon. . . Ba. [Barth 1, sa, ha &c.]. ha-ho Namsang (ha earth), ha-hoa Joboka # 540? ETHNVLOGY OF THE TDu-PACIFIC’ ISLANDS!” ” cha-ju M. Anb, ha-jo Bodo, lia-chur Garo, a-su Deor. Ch, Ja teu Sgau Karen, tok—song Limbu, sa-kha Ang. {Chinese shan]. ; ma-sin Tengsa. fi shan Chinese}. 30. tok—song Limbu [/e-tok earth Mulung, Tablunz, tu Mon, to-he Male, dong-ar Gond]; son Anam, sa-tong J i Chit Burman wr., taung’ sp., shung, tung Mru. 3c. m-—lo-di Dophila, no-di Aka, a-di Abor, ram-thing S. Tangkhul, Khoibu, ching Kapwi, Manipuri, cheing Songpu. dan-da Magar, Sunwar. 4. to-kune M. Kumi, (le-kong earth), ta-ko Bak, kon L. Khyeng, koung Toung-thu, gun Newar, kwon Gurung, gang Murmi, se-kha An- gemi, kha Mon, do-kang Milch., dung—kang Tiberkhad, go, giong, kon all, Anam, (kon, ku lau island ), phu-khan Siam, kong jai ( jai Aig) ). Dravirian kono-m, konda, kunnu, gudda. [Zurth 6]. Stone. Tibetan, 1. #-do Bh. wr., do sp., doh Serpa, Lhopa. 2, gho-lo-pi Thochu, go-r Takpa, 3. r-ga-me Horpa, ra—cu Gyar. » 4. wo-bi Manyak. Southern. 1, #-tau Maram, Songpu, da Anam, thach Anam, ——, 2. [Mountain 3, Barth 2). mp-hla Taying M., laung Mijhu, fv—-laung Mrung, i-lung, i-ling Abor, a-long Dophla, »-lung Singpho, ta-long Jili, long Namsang, Muthun, Tobcka, a-long Khari, lung-zuk Nog., lung-mango Tengsa, youg Tabl., yangi-yong Mulung, k-yok, k-yauk Burm., lung Kapwi, C. and 8. Tangkhul, Kyau, nung Manipuri, xga—lung Champhung, Luhuppa, thw-lunz Khoibu, lung-gau N. Tabgkhul, kh- lung Maring, ¢e-lo Koreng, ka-lung Kumi, long-tchong K. Kumi, lun, lum Khyeng, Lun r-khe, ke-lun Kami, ta-lon Sak, lung loang Toungthu, lu Sgau Karen, lon Garo, Pwo Karen, long Lepeha, lohong Newar, lung Limbu, lung-ta Kiranti, lhung Ma- gar, phung—lu Sunwar, yuma Gurong, yum—ba Murmi. 3. ti-mo, ta—-mo Kambojan, ta-moe Ka, ta-mok, ka-mok Chong, ka- mok, kv—mouk Mon, man Kasia, ta-wha Mru, pakh—nagUraon. 4. .ka-tse Ang., ha—che M. A. (ah, shik Ch. |. 5. wi-thur Dhim., on-thai Bodo, ; 6. hin Siam, Khamti, Laos. Prob. from hil Asam, and hil from the Bengali shila, Hind, sil. bd 7. tongi Gond (mountain dong-ar). 8. dirv, diri Kol. _ The names for earth, stone, mountain, land or country, and island, often involve the same root. _ Inthe Tibeto—Burman voch, the sibilant root is earth and mountain ; the liquid is earth, mountain and stone ; the labial is mountain, earth and stone. 1. The Silrlant Root. Eurth. | The sibilant root is used for earth in all the Tibetan dialects save Many- ETUNGLOGY OF THE INDG-PACIFIC ISLANDS, 55 ak,—sa, si, cha, sé, zi. In the South itis rare, occurring in the Naga- Bodo and Nipal groups in forms referable to the Bhotian and Worpa, Mountain. In the Naga group it also enters into names for mountain. The similar root-generally combined with it has u, 0,—ho, hoa, ju, jo, chur, song,—in Deoria Ch. and Sgau Karen this last form is itself the name au, tsu. Itisthe qualitive great, high &e. In the only full vocabulary of any dialect of the Naga group which we possess—Mr. Hodyson’s Bodo—ha, with qualitives Gdaaposets forms numerous other numes. 2. Lhe diguid Rovt. Lurth. la. The only Tibetan example of this application is the slender Manyak ga-li, m-li. Burman has the same form moditied to re, and the same prefix. It occurs with the dental and guttural pref. (Ceyarung) in Taying Mishmi and some of the Mavipurian dialects: The prefix in the Naga a-li may be either from ma— or tiny 1b. A broad form lu, long, dong, ding, roung, ruong, has this ap- plication in 8. Tangkhul, K. Kumi, Mru, Anam aad inn oomne le. Thea form is rare—Yuma, Nogaung, Dhimal. 2. The form lai of the Manipuri group appears to be connected with — the Anam noi afd Chinese nai; and with the noi, loi, doi Ai// of the Lau fam; The Mijhu nyai preserves the Chinese vowel better than the Anam noi; while it isan example of the Anam elemeut found in Mijhu. Mountain. la, Bhotian and Takpa have the slender form which Manyak ap+ plies to earth, li. 1. In the south the slender form is not used. But the o, u form occurs in Dophla, Lhopa, Lepeha and Tiberkhad on the one side, and in the Manipuri group, Pwo Karen and Anam on the other. That thisis an archaie Tibetan variety appears from the forms and. their distribution,— rong, rok Lhopa, Lepcha ; dong mountain, earth Anam, dung mountain Tiberkhad, eauntry Mon, tone mountain Jili, Burman. The dental forms, however, may be referable to the qualitive sibilant root. Stone, The 1, r, roof occurs with this application in Tibet in the Thochn glio-lo~ * pi, Pakpa go-r,. : ‘ LieT rit The broad form is more common in the south with this application, than any other. It has all the variations that occur in the names for mountain and earth—long, lo, lung, lun, lu, yok [=lok, rok], laung. It is found in the upper Brahmaputra-Irawady band—Dophila to Sing- pho—in the Nagi Manipuri, Yuma and Karen groups, _ Abor has both the broad and slender vowels Inne, ling; Manipuri has an nform, nung; the Burman k-yauk, k-yok is referable to a form k-rauk, k-vok similar to the Lepcha rok garth. . 3. The Dental Root. Earth. . ETHNOLOGY OF THE. INDO-PACIFIC, ISLANDS. &6 _. Kshs tin ti K-t., lands tin to K-t,, is allied tothe Anam tho and dia, and eraboian, delve. . igh cesealt at ait Mane The Anam dat, Khyeng det, Mon te, and the Kol te, the; are of doubtful affinity. Similar forms of the sibilo-dental root for large &c. occur, e. g. det Bodo. In the Anam? dat dai, dai is the ya ge great, but dat may have been an older form.. The Aryan desa, ma} ibly be the ori- ginal of det,dat.. , » ih Oye Bi Rite age ! ; ~ ren Mountain... |... ; 5 rere dental and sibilant root appears in general to be the qualitive, large Stone. | ie, , wy ‘ _ The Bhotian do is probably a form similarto the southern dong, and » its.etymology mustshare in the doubts that attach to it.. The Avam da, thach, Maram n-tau, are also similar forms of the dento-sibilant root dor large, ta, tai Chinese, dai Anam, tai M. Mishmi, tau Newar, da Gare. 4. The Labial Root. n Worth. | sc) - The labial root is not applied to earth in Tibetan, \ . __ In the south it occurs in the Mon-Anam family, Muthun, Dhimal, Abor, Kiranti—generally in conjunction with another root. Hig: Mountain. Nain tl The Gyarung ta-vet, Thochu s-pya’, indicate that the fyll form had s final k, t, and that the Manyak bi isa contraction. == “ In the south the root is rare and the forms are different from the Ti- betan, Sinepho having final —m, Manipurian and Nogauug' -2. Stone. ) In Tibet, Manyak alone has the labial, wo. | The Motj=-Kambojan group has itina more archaic form, mok. In dis- * tinct forms it is also found in Mru, Kasia and Uraon. From the rarity and the mode of distribution of the labial, the little connection between several of the forms, and the identity of all with eur- rent forms for great, round, fat &e., it is evidently the same root ; and it - was probably originally used asa qualitive in all cases, asit still is in several of the names. Thus inthe Muthun ha-wan it follows ha earth comp. great wa Burm., ba Singpho, rownd pan Ahom, fat kwi-pan ryarung); lai-pak Manipuri has the common Manipurie root tor earth lai, followed by a form of the root for /arge which may be Manipuric (no qualitives beingrincluded in the vocb.), and is similar to the Lepcha phat | earth, Gyarung ta—-vet, Thochu s-pya’ mountain, and to the form of the qualitivein Abor bhote, large, Tengsn ta-bok, Khari ta-bit (=ta-vet Gyar.) fut, Manyak wa’ wa’ round &c., and tthe Mon-Kembojan ta-mok, - tu-mo, stone (=th—bok ee Khari), Manyak wo-bi. ‘The byes bom ~ bum, mowntazn, is the Bhotian large bom. Singpho itself has phum fat, in Bodo phung, which is the Manipurie form for mountain, phung. e Kumi muse, moi mountain, Dhimal bho-noi earth (noi earth), Kiranti ba- Ukha, Abor mong are also similar to current forms of the qualitive—ba large Singpho, mon round Khinti, pwi large Thochu &e. ) WTHUNOLOGY OF THE INDO-PACIFIC I:LANDS. 57 The Khari @-pih motintain has the slender form applied to lerge in Khari itself as in ‘Tengsa, Thochu and Chinese, and to,fa¢ inthe Lau fam . 5. The Guttural Root, Earth, Mountain, The guttural root is not Tibetan. It is found in that range of the south- ern vocabularies which has the bsg erooneaee af archaic Himalaic anil Indian’ vocables— Anam, Mon, Yuma, N. Gangetic &c. The most common forms khon, kon, kung, kang &c. best adhere to the primary form—kunnu &c, Drav. Stone, The Horpa ga, Thochu gho, Takpa go, Gyarung gu, is not found in the gouth. Te ie probibly a comparatively late Scythic acquisition, The Dra- virian kan, kol, kolla, kallu has distinct Scythic affinities, The qualitives for great, round, fat, high, long—into many of the names for which, the same root enters—appear to be used in the present class of words to Some extent. The inperdeetion of the vocabularies is a bar to our discriminating them from the substantive roots, with some of which they are phonetically identical. The words given for earth vary in their meanings from the mere soil, to land generally, to the earth as a whole, to @ particular country &c., and it is quite possible that, in the wider applica- tions, the qualitive ib may have formed part of the name, That it should occur in words for mountain is more obvious. The same root oc- curs in names for earth and stone, and the epithet large might readily be applied to rocks as well as mountains in the Tibeto-Himalayan region. n the names ha-ho, ha-jo, cha-ju, ha-chur, tok-song mountain (5), in which ha, cha, is the sibilo-aspirate root used for earth, land, the second element appears to be, without doubt, a Himalaic qualitive for great, high, tall &c. su, thu, tho, sung, ie &e. (ps 80 to 82); and in the Deor. Ch. a-tsu, Sgau Karen /a-tsu, the same qualitive appears to be used as the root The Manipuri ching, cheing, thing, are much closer to other forms of the sibilant root for darge &c. than to any of the undoubted variations of the liquid root for mountain, earth. The adjacent Naga has the broad form chung, chong large. The Jili, Burman and Mru tong, tung, tung, ‘mountain, may also be referable to this root. Tong again is so close to dong &c. that it unsettles its relationship also. 3 In lo-di (8) the slender form of the same qualitive follows the liquid root (large thi Gyarung, ti Lepcha &c.) The labiul is clearly the root for large &c. ‘The most common liquid root for earth, mountain and stone is phoneti- eally identical, in most of its variations, with the liquid root tor great &c. But as this is necessarily the case whenever the same sound forms differ- ent roots, each with several applications, it does not seem possible, with the present small vocabularies, to say if the identity be more than phonetic. External Relations. 1 and 3, Earth. The sibilo-aspirate sa, cha, ha, tha, zi earth is probably connected archaically with the Ghinese ti, tien, tin, tho, 58°" RTHNOLOGY OF TUB INDO PACIFIC ISLANTS!” In the Scythie alliance this root is not common in the sibilant form; but ~ the guttural, which is very prevalent, appears, from the gradations in seve- ral of the groups, to be in general a variation of the sibilant. Itis not Ugro- Fin in the sibilant form. But Samoiede has the Tibetan form, ja, decha » &v., Aino fui, toi, Japan tsu-tai, tsi, zi, Korea ta+ti. The Japanese and Korean are probably Chinese. Tarkish has sir, ser, zir, yir, &.j' Aino siti-. kata, Korea chala, chli, hilyi, Tuazusian turu, tor, Turkish tor-pach &c., Mongolian has cha-dsar, ga-syr &v.; Votiak has mu-sem. Caucasian has sach, seoh (J rriaa Aill, Chinese eurti) Ossetics di-cha, mi-tza (Fin) Georg. ; tzau-la, tu-la, chu-llah, (Korex chu-lu. &c.) cheb, chy, mu-sa, mi-sa, bi-su ne mointain), Canc, proper, apres Seythie forms are fouad, as naual, in [ndo-Furopean. The pure sibilant occurs;in Zend sa, sao. The sibilo-liquid sir, ser, tor, turu &c. is repre- sented by the Celtic tir, dor, duar, Latin terra; and the labio-sibilant. of Scythic and Caucasian by the Aryan mei-ti, mi-ti, ma-ts &c. The Scla- vonic family has sem-lya, sem-ya, sem &c., Zend semo, Sanskrit sima, Persian zam-ia (symun-it, shim-ta &e. Kamsch.) re Stone. The sibilant root is also applied to stonein Chinese shi, shik, chio, sa’; in Seythie—Fin tschi-ni, zi-wn (ki-wi), Ugrian is (Permian),,Japan,- jai, ishi (tsi earth), Aino shio-ma, Tungusian za, hy-sha, Yeniseian shish, Mongolian tscholo, tschila-chon, Turkish tschol, tash &e.; in Lranian sil-ex, shi-la, sila &¢.; staina, stain, steen, stone &e.; in Caucasian i-zo, che-zo, hin-zo, she-ru, Ossetic dor, durr (Tungusian tor earth). Semito- African ha-dsar, ha-sar, a-sar, i-sha. Mountain. The same_root is applied to Mountain in Fin mo-tschi, mo-zi ( Aill me-to, ma-tas ); Ugrian, is Perm (also stune), tachoi Perm (sysi, mu- syesn earth); hill Wolga u-sach, ech-sait : Samoiede sye, seda, soti &c.; Ye niseian dschii, chai: Turkish syrr, kirr,* tu, tach, tag &e.; Ugrian (Wowul) ill sal, Mongolian ald shili. Semitic has also the dento-liquid of Scythic, tiru, turo, toira; serri Hata-~ gi, tulu Galla. Armenian has zori hill. . From these examples it appears that the sibilant root.is one of the primary Asiatic ones, and that the Tibetan torm associates itself with the guttural Chinese, Samoiede, Caucasian aod Zend, shi, se’, dacha, sach, sa &e. &e, and not with the liquid Scytho-[ranian, Georgian and Semitic tzula, - tur, airi, sila dc. &e. 2. The most important root is the liquid, of which the more archaic broad forms, lung, Jong, rok &c, are preserved in the South,—Tibetay hav- ing lo as an element in Thochu gho-lo-pi stone only, the form in earth and mountain being slender. If the root be not the native liquid root for * The Turkish syrr, kirr suggests a connection between the sibilo- liquid and the gutturo-liquid root—the latter b-ing common to Scythic, Indo-European, Caucasian, Semitic and Dravirian. Scythic,—gora Fin, kur-uk,-ker-as Uyrian, kar, kono-ny Yeniseian (which has also kai, chai, dschii); Indo-Eur.,—gora, kore, gura Sclavonic, gor Celtic, giri Sanskrit ; Caucasian kur-sh ; Semitic gar, gor &e,; Pashtu gar; Drav. Gand &e. ETHNOLOGY OF TUB INDO-PACIFIC ISLANDS, _ 59 et great &ex (pp. 31, 32)—the forms. being, toa greatextent, identical—it _ has few external affiaities. . This circumstance. is strongly in favour of the - . € sibilant beiag the substaative Himalaic root, and the liquid being merely” the qnalitive large &e. —- Prt . In, the Seythic alliance the liquid root Per to occur only in two dia- lects.of Ugrian, loch Ostiak mountain, rek earth Wolga. a 4 i . ; 4. "The labial is a Scvtho-Tranian and Caucasian root for Eurth, Stone, Mountain, but it does not appear to have any conaectioa witu the Tiveto- Barman forms. Mountain. One of the most widely spread varieties for mountain is the dabio-liquid. Seythic—ware, wuorl, Fin panda, wan-da, ponda, uwa~ run,pel, pala, Aili mar, moli-ma Ugrian, mari, mony-mon-si, borr,” bre ore iede, buri, nv-buri Aino, urra, uro Tanzusian, ula, ola (Aild bol-duk, bori) Mong., mur-on Turkish, a-mar ill Yeniseian, Caucasian mehr, meer, bil, pil, du—bura, bir-d. Indo-European par-va, par—bat; berg, bare, biarg &c.; mons. Deavirian varei, male, buru, par-ta; Australian an-birik, wari-at, mur-do; Indonesian bulu, palu. Aclabi ttural is found in Circassian buch, buko-du (earth bak Les- ciam bach Yeniseian, ma, mag, myg Ugrian, tor-pach Turkish) and Mala-_ gasy vohi-ts, buhi-tra &e. : Earth. Seythic.—ma, mu, mag, my, myg Ugrian, pan, pany, biny, bach Yeniz. Drav. puda-vi, pulo-va. Nilotie—barr Tumali, baro, ‘bara Dan= kali &e. Stone, pai, pui phi, pi, Samoiede, poi-nah Aino; vatu, batu Malag sy} i eet era? wal- i, bura, maramio, ke. Australian, &u, ewe Names or Parrs or tHe Dopy. : ; Head. — Hair. Lye. Ear, 1 Bhotian wr. m-go Wr. my Ne Lag e208 sp. go kra, ta mik am-cho. 2 Horpa - gho s-pu mo nyo ie e . hom-pa ss kan’ nukh 3 Thochu ka-pat . Hy | oe We u tai-myek tir-ne © 9-8)" 4 Gyarung _ ta-ko ; 7: } emai, SS : omni na-pi_ 5 Manyak wali. ; 3 re ) ie ter ie agg 9d the-long ne-bla 6 Takpa gokti ff heey peome t ot Ee ’ 60 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. ‘Mouth. Tooth. Hand. Foot. Bone. Shin. 1 kha 80 lag-pa r-kang-pa rus-pa pag-spa kha £0 , lan kango = ru-ko - 2 ya syo a ko re-ra i 3 drukh swé jipa ja-ko —rri~pat ra~pi 4 ti-kke ti-swe ta-yak = ta-mi sya-rhu ti-dni S -ye-ba phwi’ lap-ché lip-c#é ru-khu — gera’ 6 khe wil In le-mi fos-pa phyekh Head. 1. The guttural root connects Bhoticn, Horpa, Gy2rung and Takpa. It is Ugrian og, oike, ug-ol, ug-om, Yeniseian koi-go kc. The kent form is prevalent in Abor, and the Nega-Yuma group. Mishmi has 2- kau, kou. The Takps gok is found in Manipuri kok (Limbu tha-gek). It has been communicated—by the Naga fam. apparently—to Male and Uraon, ku-ye, ku-k; kho Namseang, ta-ko Tengsa, (ko Aair), ta-ko-lak Nogaung, m-ku-re Mishmi, kho-ro Bodo, kha Karen, kau Champhung, kui Luhuppa, a-kao N. Tangkhul, o-kao 8. T., lu-gu Khyeng, a-kbu Sak, khang Muthun (kho fair), khang-ra Joboka (kho hair), {kra hair Bhotian &c.], khong Burm. wr., ghaung sp., nggum Jili. | It is also applied to the hair ko Nogoung, ku engsa, kho Mithun, kws Khan, kin Deoria Chutia. 2. The Thochu kapat is one of the peculiar vocables of this dislect, It appears to be Scytho-Iranian, and to have no Ultraindo-Gangetic repre- sentatives. ‘ 3. The Manyak wu-lli appears to be an example of the Scythie liquid root. The Dhimal pu-ring is derived from it, and the Khari ¢e-lim hes the same slender form. The broad one is found in the Manipuri gr., a@-lu, lu, Yuma lu, lhu, a-lu, Ahom ru, and Magar mi ¢a-lu, (but the last may be talu, Drav.). The broad form appears tobe that of the Lat family, in which the common form, ho, hua, seems to bea softening of the liquid root-pre- served in‘Ahom. It is also applied to Aair in that fam. See Hair. This root for Head is found in the forms long, ron, rung, ru, lom, attach- ed to the names for the eye and ear in some dialects. For eye Takpa has me-long (eye of head, me is the root for eye), Taying Mishmi mo-lom, Garo mok-ron. For ear Abor has nya-rung, no-rung, Mishmi ma-/-ru. 4. he dental and sibilant root is common in the South for head (Sce Hair 3.). Hair. : ee are two prevalent terms, pu Aair in general, and kra that of the ead, _1._ The labial is found in all the vocabularies. The form pu connects Bhotian, Horpa, Gyerung and Takpa, The ‘Thochu hom and Manyak Mul are exceptional, * ermyouocy OF THE tYDO PACIFIC I3LANDS. 61 The root is Chinese, mo Kwangtung, fah, *mau Kwan-hwa, bo &e. in other dialects. It is also Scythic. = / The teria is rare prseatalaes cn Read haGe rie form fut We he Tablung Naga min, Sak.kn. mi (ku head),.Garo ka-man, Bodo ie-mon, are probably, of tiacmot Ohthess origin) The Naga mii msembles the Chinese word for fuee min, and the Kwandiwa piu (Awangtung pan) the hair on the temples. Kiranti has moa and Gurung moi. “ar Applied:to the head it oceursin Singpho bong, in the compound Bur- man chhan-bang, s’ha-ben hair, in Songpu pi, Maram a-ju, hKoreng cha- pi, in Sunwar pi-ya and Magar mi talu, The Kol bu, buho, bohu head (ub, up hair) may be Tibeten.. ; 2. kra occurs in all the vocabularies save Manyak. The Thochu form, ng, is exceptional. The root, if the guttural, may be Scythic, kar-y Vin, 5 det Aino. The guttural is an element in words for the beard in Yeniseian and Tungusian. van This root also isnot common in the south. It occurs in Singpho kara, and with the meaning #icad in Gurung kra, Mishmi m-kura, Bodo khoro, and Munipuri ta-kolak. ‘ Fyrom the disposition of Bhotign to merge the prefix in the root, kra is probably a contraction of ka-ra, In the Thochu g-rong the root has a form ‘similar to the Anam long; ph-rum Ahom, ph-om Lau, Siam; lu Khyeng, ta-lu Tung-thu. In some of the Southern forms the tirst element is not. a def. pref., but the euttural root, and the Tibetan k--2 may also be radical, e. ¢. fa—ko-lak Manip,=s—k-ra Dhot,, m—Ru-ra Mishmi &c. 3. The Bhotian ta, Manyak tsi and Thochu /a-chu, is alsé Scythie,— ata, at &c. Fin—and Chinese, Kwongtune, tsa hair of the upper lip, (Manyak tsi), su beard (ke-chu Thochu). The Chinese than, sau, head? may be also connected with the Tibeto-Ultraindian dental and sibilant roota for hair and head, In the south it is the most common term; but the source of Some of the forms appears to have been a broad archaic vocable early introduced by the Mon-Anam family, Kambojan suk, Mon tok, sok, thwot, suet, Karen thu -Rasia shuin. This form is allied to the Thochu chu. Mikir has chu, ani Namsang Naga ka-cho (=Thoehu ka-chu), Mulung su, Dhimal. tu (mui _ ta), Taying Mishmi thong, Lepcha a-chom, and Newar sony. e Bhotian ta appears to be related to the common Irawadi form. Mi- ~ jhu eham, Angami tha, Burman chhan, s’ha; lu-sam Khyeng (lu head), _ a-sham Kami, shamgkumi, Mru; Sunwer chang, Magar chham, Limbu tha-gi. It is also one of the most common southern names for the head. Anam mang song, thu, Naga sang, Angami a-tsu, Abor mi-tu-b, mi-tu-k, tu- ku, #a-tau, Toung-thu fa-tu, tu; gu-toh Lhopa, a-thi-ak Lepcha, tha-gek * Limbu, tang Kiranti, tho-bo0 Murmi, chhon Newar. The Chinese thau, shau, may have been the original of both a and u forms for head, hair, Auam has the derivative dau head. Lye. 4 1. The Jabial root is found in all the dialects save Thochu. The slen- der form vonriects Bhotian with Gyarung and Manyak. While these slen- der forms preserve the final guttural, the Horpa mo preserves the vowel of the Chinese form,—muk Kwang-tung, mo Kwan-hwa, mok, ma,. ba in other dialects. The labial root igcommon in the Scythic vocabularies as applied to the . , . 62 ETHNOLOGY OF THE INDU.PACIFIC ISLANDS. Head, Hair, Beard, Faces” Mouth, Lip, Teeth, Nose, and Ear, but it is other in the pure labial form, or with a liquid or sibilant final, r, 1, n, s. The guttural occurs in the Ostiak wai-mik fate. The Japanese me, ma- nako eye, are related to the Chino-—Tibetan root ; face is o-mote, 2. The Thochukan is the common Chinesengan, Kwang-tung &c. (yen in Kwan-hwa). , The Brahui and Dravirian khan, kan, appears to have been derived from the Thochu form of the Chinese root. The Turkish kar appears to be also connected with the Thochu and Drav. form. ' The Manyak mni and Takpa me appear to be referable to the slender form mik, myek ; mui is explainable'as min from ming, mik. The form min is found in Mrv, mi in Mijhu Mishmi, Singpho, Dhimal, Angami, and.several Yuma dialects, mi, me, @-mi, Karen me, Khyeng mi-u-i ; min appears to be an archaic form of the root, as it isapplied to face in Chi- nese mien, min, Abor ming-mo (eye a-ming), and probably in other Tibeto-Ultraindian dialects. The word face is not spied in the short vocabularies, : The form in t, d, occurs in Lhopa mi-to, mi-do (also mig), Miri a-mi-da, N. Tangkhul a-mi-cha, As -to, os is a Lhopa postfix (gu-toh head, gong- do egg &c.) the root may here have the contracted form, But mito may be mit vocalised. Kyau has me-et, me-to. The Bhotian form mik, mig is the most common U]traindo-Gangetic. The broad Mon-Anam mot, ka-mot, pa-mot, mat, Kuki mut, appears to bea distinct imporfftion by that family. It is directly referable to the Chinese mok, muk. The form mak is also found im Champh. a-mak, Garo mak-ar, Bodo mag-on, and Kiranti mak. Naga has ¢e-nok. The Deoria Chutia muku-ti appears to have the full Chinese muk (¢i is the Takpa form of the dental posttix). | ‘The Gyarung myek is found in Burman, myek-cht, myet-st. The Abor nyek, Naga ¢e-nyk, te-nik appears to be a modification of a similar form. It is found in Lanpa-ned, and Kambojan pe-ne, pa-nek. The form met, med, has been received by the Kol dialects. The common ta of the Lau fam. appears to be from mi-to Lhopa, mi-da Miri, mat-ta Shan, Face is na ta, in Namsang than. The Takpa long (/#ead 3) of me-long, is found in Taying Mishmi mo- lom, ma-lam, Garo mok-ron, mak-ar, (See also Zar). ‘ Luar. — ; 1. na Bh, wr., and Manyak, nyo Horpa, ne Gyarung and Takpa, are connected. The Thochu nukh or nu-2h is probably a full archaic variety of the same root. If so,it is neither Chinese nor Scythic with the meaning ear. Chinese has rh, ro, ngi, li, i; Kamschatkan illa, yel-uth, all-od, il-yud ; Can- casian en, i, lai &¢.; Indo-European ohr, ur, aur—is, or-eil ; African ilai, iroi, ulu-k, uil-ge, mia—ru, noa, no-f. In some languages the ear derives its name from its resemblance to a leaf. In the Takpa ne-d-lap ear, b-lap is leqf. As the Tibetan na does not closely resemble the Scythic and Chinese roots for ear, it appears to be itself an older application of the same root for leaf. Bh. lo-ma, Hor- pa da-ld. In Manyak it has n (as in the word for ear) mone Sokpa hag nai. In the south ne, na, lai occur, as well as other | forms. In the Manipuri up na is the common form, and in some of the dialects the forms for paf and ear correspond, e. g. Songpu nhui leas, a-nhy-kon ear; Cham- ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 63 phung sing-nu leaf, khu-nu ear; Kapwi na leaf, ka-na ear; Angami po-nye leaf, a-nye ear. ang The Chinese yp, ip, ye leaf’ is perhaps a softened form of an archaic root similar to the Tibetan nip, . The Burman rwak leaf (yuet sp.) and the Kambojan si-I6k are ar- chaie forms with final k for p, as in the Thochu nukh éar The Tibetan root is very common in the south, mostly in the a form— Singpho, Naga gr., Manipuri gr., Yuma gr., Karen, Burman, Nipal gr. The o, u, form of Horpa nyo and Thochu nukh, is found in Kumi hu- “no, Khyeng hno, ka-nhau, Lepeha a-nyo-r, Sunwar no-pha; while the common Nagu form is na, Khari has ti-nhaun, Nogaung ¢e-naung and Tengsa te-lanu. a The slender form of Gyarung and Takpa ne, is found in Angami a-nye, Limbu ne-kho, The Mijha Mishmi ing, Maram in-kon and Mikir in, an, are peculiar variations of the slender form. In the Dhimal nha-tong, tong appears to be the Tibeto-Ultrainda den- tal postfix asin Tib. me-teg flower, men-to Takpa, Lhopa me-da eye, Abor lam-te road &c. It ovcurs with the same form in the Dhimal si- tong tooth, and in the Garo lha-tong tooth, ho-tong mouth. The root {or ewr,is combined with the liquid root (for head probably), in some dialects. nio-rong Dophla, nya-rung, no-rung Abor, Lepeha a-nyo-r. Taying Mi hi has na-kru and r-kru-na (in-kura ), Karen na-ku (ku head), Garo na-chil, Songpu a—nhu-kon, Tangkhia-kha-na, é-kha- na, na—-ko-r, Khoibu kha-na, Limbu ne-kho, Magar na-kyep. Maring has na-mil and Aka na-bar. In the Nipal gr. a labial postfix is frequent, but it appears to be the definitive in some instances,—na-ba Kiranti, na—pe Murmi, nai-pong Newar, na-be Gurung, ne—-pha Sunwar (na-vo Lhopa). . . Tf na &e. were originally leaf, these compounds were probably “leaf of the head”. The Bodo kho-ma, Gara ma-chor, Kasia s—kor appear to have only the word for head, the root for ear bing dropped. (In the Mon-Anam comp. voc., kor &e, is treated as the root for ear, improperly as 1 now think). 2. am-cho Bh. sp. This term appears to be exceptional in Tibet. Serpa has a fuller form am-chuk. This appears to be also an ancient Tibetan word for leaf. It is found in some of the Naga dialects, pan- chak, hu-chak, phum-yak, (pan &e. is tree), Deoria Chutia chia, The ba ie und Nogaung am is probably a contraction of the Tibetan am-cha, Robinson gives nam-cho as the Bhotian term, from which it would appear that nam is a form of na, nap, lap &e. The Lau family has hu, Lung-khe hua, which appéar to batmodifications of the 1, r,n, root for ear, leaf, the liquid sometimes becoming aspirate in the Lau fum. In the sameway the Gurung lau legj‘is hau in Newar ; and the sp. Bhotian hyo-ma appears to be a softening oflo-ma. Mouth. * 1. The Bhotian kha, Takpa kha and Gyarung ke are connected. The Horpa yais probably a softening of kha, and the Manyak ye is referable to it, The root is Chinese and also Seythie,—kau Kwan-hwa, hau Kwang-tung. It is not common in the South. Songpu aka, Kumi uk-kha, Taying Mishmi ta-khu, ku-kwen, Aka gam [Garo wa gam tooth], Bodo khou-ga, Garo ho tong (pha tong tooth). C4 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 2. The Thochu dzukh is perhaps the same root—as the sibilant and den- tal found in several of the southernedialects, Naga gr. tun, chu-sim, Ma- nipuri chil, Kuki taung, Garo ko-sak, Kasia shin-tur, Maram ma-thu, Murmi, Gurung sung, Sunwar so, Kirantidoh, Chepang mo-thong, Shan thsot, Szau Karen tha kho (kho dead). With the labial tinal it is found in Ahom and Khamti sup, sop. The root is Seythic and Chinese. Fin su, sun, sum, Ugrian shob, shus &e.; Chinese sui, choi &c. See Tooth, 3. The inost common root in the South is the labial. Abor na pang, na~ png, Naga gr. te-pang, ta—bang, tu-pin, amu( Moz. Ang.). Kumi la-baung, yeng a-hmanng, Monipuri gr, ma-mun, cha—moun, kha-mar, kha-mor, 8. Tangkhul, Khoibu, Maring mur #@Lepcha a-bong, Limbu mura. This root appears to be of Mon-Anam origin,—pak Lau gr., meng, mieng Anam, pan, kda—mon-pan Mon; Ka boar, Kambojan mat, Nicobar minoe. The labial root is Seythic—o-m, ha-mun, a-ma, a-man, Mongol., whence probably the Mon—Anam pan &e. Singpho has ningup, negop, Jilinong, N. Tanekhul ania, §. T. oni, Burman nhup, nhok, Pwo Karen no’ Sak ang-s?, Mru naur Mijhu Mish- mi njyut, Meyar nger; Chong ra—neng. The root is Scythic nyen &e, Samoiede. Tooth. 1. The Bhoti@ so and Horpa syo are related. The Thochu and Gya- rung is probably the archaic form. The Manyak phwi and Takpa wa’ nay either be referable to it or to the ladial root for mouth. This vocable is the Chino—Scythie root for mouth. Chin. sui, ch’hui, choi. Fin su, sun, shum &c. Ugrian shus, tos, shob &c. [Selavonic also has sub J. It is tooth in Turkish tish, tusch &e., Samoiede tipe, tip, Ostiak tiwu. It is not very common in the South, unless it vavies to the labial. Mijha Mishmi tsi, Anam si, Burman swa, thwa, Pwo Karen thwa, Murmi swa, Gurung sak, Magar syak, Changlo shia, Lhop soh. Several of the forms strongly rescmmble those of the sibilant ond dental root for mouth, head, haw (Hair 3), and the root is probably ultizately the same. Comp. the Mon—Anam sok hair; Abor mi-tuk, Lepeha a- thiak head ; Garo ko-suk, Sunwar so mouth ; Gurung sak, Magar syak, Bhotian ao, tooth. seh _ 2. Labials are more common, Singpho wa (Takpa wi), Naga va, a, ta—phu, ta—bu, pha, ta-pha, soya rr. a—va, a—ha, o-ha, ha, a—hu, ui; Garo pha fong, Aka phi, Daphla fig, (Manyak phwi), Abor i- ng, Siamese fan, Kambojan tim—bang; Seeu Karen me, Yuma a—pha, na, d-ho, hoso; Lepqia a—-pho, Limbu he-bo, News wa, Milchanang bung. The root ‘Is identical with the labial one for mevth, In Seythie lan- guages also it is used for tooth,—Ugrian pane, pin, pon-A, pan-ket. 3. Mijhu Mishmi lan, la. 8. Tangkhul ala-ra, Anam rang, This exceptional term is probably the liquid root for Aead, which also occurs attached to roots for ear, eye, &e. 4. The Lau khiau, khiu, Jilirkong, Kiranti kang, are probably related to the guttural roots for mouth or head. The Sunwar kryu may be con- nected with the k-r root for head, hair. , 5. Kanwi nga, Tungthu ta-nga, Maram a-ha, Manipuri ya, Song— pu nai, Mon nveok, ngeat, nget. This vocable is Chinese, nga Kwang- tung, ya Aw: n-h va, ETHNOLOGY @F THE INDO-PACITIC I:LANDS. 65 Hand. 1. The final Nagra connects the Bhotian lag, lan¢o and the softened Gyarung yak. The “ag 9 Tha and Takpa la are probably contractions of the Bhotian form. Tha Manyak lay mga to be a distinct archaic form, as a similar form is applied to the ‘foot ip. . This form is the same that is used for /eaf, and it is also found in southern languages with both meanings. Gurung lap-ta Aand, Murmi, Newar lap-te leaf. aa) Ie The -k form is an archaic variety of the same root, as it ja also current for leaf. “ Tics oot has both meanings in Scythic also, The Bhotian form resem- bles lag-ol Ugrian, i-lik Turkish, hand. 2. The Thochu jipa is peculiar, 1. The common Tibetan root is prevalent in the South Lhopa la-pa, Abor e-lag, Aka lak, Naga dak, chak, yak, Garo jak, chak-reng, Mrung be phu leta, Burman lak, let, Mru rut, Lepcha ka-liok. The Manyak -p form occurs in Mijhu Mishmi yop, and Gurung lap-te. 3. Tengsa ta-khat, Nogaung ¢a-kha, Khari ta-khet, Mampurni khut, kut, a-khui, kuit, hut, Lungke kut, Kyau ket, Kumi kok, a-ku, ka, Kami a-ku, Sak fa-ku, Khyeng kuth, Bodo a-khai, Dhimal khur, Ahom kha, Chepang kut-pa, Mayar hut piak, Limbu"puk, ¢a-phe, Kiran- ti chuku-phema, Changlo godang, Milchanang god, got. 4. Jili ta-phan, Songpu banyhoreng eha-ben, Maram van, Champhung a-pan, Luhuppa pang; Angami a-bi (foot a-phi, u-phi ), Sunwar g-wi; Lau fam, mu, mo, m1. Chinese. words for hand and foot are found in some of the Ultraindian vocabularies, that for hand being in some ee to the foot, and thut for foot to the hand. Hand shau Kwan-hwa, Kwang-tung; Finger shau, chi, ib. ; Foot tsu Kwan-hwa, tsuk Kwang-tung. Taying Mishmi thyoa, a-tua, Anam tay, Ka dei, Mon tway, tai, Kasia k-ti, Karen tshu, Toung-thu su, tsu, Deora Chutia otun. Sunwar ta-b-le. (See Foot.) < oot. ih tbe Bhotian kang, kango, Horpa ko, is the Chinese root keuk, oh, kha. 2. The Manyak lip-che and Takpa le-mi, have the same liquid root that is also sepia to leaf, ear, hand. 3. TheThochu jako.appears to be Turkish, ajak. 4. The Gyarung ¢a-mi, and Takpa mi of le-m1, may be"connected with _ the Chinese po, a footstep. 1. oF hong, Pingpho la-gong, Jili tak-khyai, N. Tangkhul -kho, C. T. o-kho, 8. T. a-ke, Kapwi ki, Maring ho, Mrnng ya-kong, con ren kho, khang, Toone ep khan, Anam kon, Rang, Kami a-kho, Kumi a-kok, a-kauk, Mru khouk, Khyeng ka-ko, Dhimal kho koi. Anam chen, kon, kang, Rembgian chine ae chang. 2. Garo cha-p-lap (cha—-kreng keay. Lhopa kang—lep, ' The Naga cha, tchya, da, ta-ching, ta chang ta-tsung, Khyeng ashi, Doing-nak teng, Sak a-tar, Lau fam. tin, ten, Mon thsihn, Vepcha ‘ian - liok, have the root used for hand. The Nogaung ta-tsung foot is faithful to the Chinese tsu, tsuk foot. The Lau and Mon forms appear to be of Naga derivation, > 5 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Mulung and Tablung yah-len (yak- lan, finger, ak hand). Garo ja-chok, chap-lap (chak-reng hand, jecohak leaf). Taying Mishmi m-grung, m-groh, Burman khre, khye, Kiranti w-khuro. Mijhu Mishmi m-p-la, Abor a-le, Singpho la- gong. Murmi Ja-le, Newar pa-li, Gurung bha-le, Magar mi-hil, Sunwar The Tibetan lage Ak diang-li - e Tibetan lag hand is found as foot in Aka laga, Le liang-liok Limbu lang—daphe, Murmi ba-le. # aah: rd 4. Angami a-phi, v-phi, Koreng cha-pi. Maram, Songpu, Luhuppa phai, Champhung a-phai, Bodo ya-pha, Khoibu wang, Kyau pat. — Bone. The old Bhotian rus-pe is found in Takpa ros-pa end without the suffix in Magar, Sunwar end Chepang. The current ru-ko is found in Manyak ru-khu and Monipuri a-ru-kau. The Gyarung sya-rhu is also Manipuri sa-ru. The Lau duk, nuk is probably referable to the Manyak form, Names or Famrniy ann Socran Reiatrons. Prey small vocabularies only contain the names for Man, Father and other. . Man is mi in Bhotian, mi’ in Takpa, and tir-mi in Gyarung. Horpa has v-zih, Thochn na, and Manyak chhoh. Itis probable that the Horpa zih is masculine and not generic, as it occurs with a masc. power in the Manyak nea-zi and Thochu zyah bud/. Mi is common in the Ultraindo-Gangetic vocabular.es. Ast, generically, as in Bhotian and Gyarung, varying in form to bi, wi,— e.g. Newar mi-jang man, mi-sa woman, Burman sa-ani girl, Singpho si-wi girl, Garo mi-chek wife, Miri mi-yeng wife, mi-mo woman, Bodo bi-hi wife, bi-ma mother. | 2d, with a feminine application. Kasia mi mother, Khamti me girl (Dhi- mal be—jan boy), Mishm! mia woman, Siam tua—mi fem. of animals (tua- po males), Mishmi k-mai women, Anam mai, fem. of birds, Mikir a—pe fem. of animals &c. The sibilant has a masculine application in the Bodo bi-shai husband ; Siamese chai, Khamti sau man; Kumi tehau man, Pwo Karen p-sha man. The N, Tangkhul pa-sa and Kasia pen-so man is probably also masc. and not generic. The Manyak chhoh is a cognate root, (comp. cho Ostiak). The Thochu nd is Chinese nan, nen, man (vir), lang husband. The rootjoccurs with the mase. power in the Bhotian pa-la father and it is very common in the Ultraindo-Gangetic vocabularies as a masculine root and servile, both for man and the inferior animals, The word for futher is pha in Bhotian wr., pa-la sp., a—pa in Ho Manyak and Takpa, fa-pe Gyarung and ai in Thochu. That for mother is a—ma in Bhotian, Horpa, Manyak and Takpa, to-mo in Gyarung and ou in Thochu. The Bhotian masc. and fem. roots, postfixes and prefixes in b, p, vand in m, are the same words as those used for father and mo- ther. In the southern vocabularies they are almost universal with similar meanings and functions, and with various changes of form. In several of the dialects they are now detinitives absolute as in Bhotian. The mase. application of ba, pa, fa, va, bi, be, bo, bu &c. and the fem. of ma, mo, mu, mi, ine, &c, is common to most formations in the world, and must ETHNOLOGY OF THE ING PAGERS ISLANDS, 67 have been coeval with the beginning of human speech. [See App. Father, Mother). In some families, however, m is masc., and b &c. fem. [See Dravirian Comp, Voc., App. to chap V}. The Tibetan mi maa is probably connected with the Scythic masculine root of the same form, generic words for the species being generally. cog- nate with masc. and not with fem. roots. Ugrian has mis, mes, maz, man &c. for husband, man. . The peculiar Thochu words ai Father, ou Mother, are Scythic,—ai ’a- ther Ugrian, aua Mother Turkish. The Chinese nu, neu, na, woman, female, does not appear to occur in the Tibetan vocabularies, but it has been received into several of the Ultrain- dion. Gerard however gives ane Bhotian. The Chinese fem. tsi, tsz (confined to Birds in Kwan-hwa) is found in Horpa s’—ae girl, yf is common in the south, Mascutine Names, 1. The Labial Root mau male Chin. K-h, tb. (cattle), K-t. fu Jather, husband W-t., K-h. phu, pu, yy Anam phua man (vir), husband, Siam. ho, po male Bhot. sp. a ” bP | hd pha, pa, father, male Bhot. wr. pa-la (futher Bh. sp. a-pa ,, Bhot. wr., Worpa, Manyak, Takpa, Lhopa, Murmi, Kapwi, Maram, N. Tangkhul, Muthun, Joboka; Mon. éa-pe Gyarung. » pho father Lau fam. eae f » Kambojan. a—po » Mozome Angami. a-u » Songpu, Koreng, Tengsa, Angami. ain— » Kum. i- / ” Champhung. bo » AKhbyeng. ba-bu » Abor. a8 if Leathe G “- , urung. thong-po male Changlo. sing—pho man Sing—pho. pong male (an) Namsang, wa-pong pat. uncle, Mijhu. 4—pho brother og poh man Kumi, ¢ua-phu male (an.) Siam wa father Singpho. va » ili, pera Dhimal. pa » Sgau Karen, S.Tangkhul, Koibu, Mra, Kasia, Chepang, . 1-pa » Nogaung. o-pia bd C. Tangkhul, Tablung. o-pah 4 Mulung. s . > « 4 BR ETHNOLOGY @F THE INDO-PACIFIC ISLANDS. Je ages Povo Raven, Toung-thu. pee ha raid i M. Kum. a . pha-e ,, Burman. ‘ ‘cits - . : Bae a et a-pai mile xe M. (ow. ape father ” -pha ,, Garo. . Pe, ha - Mrung. a-va » Luhuppa. na—ba 9» Jay. ak e a—ba » pat Toung-thu, Serps. = ss ° Kiran: ; Kyau. ved E Sunwar. 6g nes syed (an.) Garo. Miri ha birds) Burman. wa-jon & imal. man Kuki. n-me » Taying. 2. The Liquid Root. nan male Chin. K-h., K-t. . . nen 93 43 Shang-hai rin "man Gyami. lang husband Chin. nam male Anam. © na man Thochu, eS Sather oa ‘ ya son aneph ~ sya daughter ree husband N ay ‘ ee. male (an.) ‘Bao, ta-la (fowl ) Shi . Chango (pro dae kin . (an. a Oh oak” ma-ran-ma = M-yan-ma, M-ya-ma, Ba-r-ma. p-Ta man Karen, ph-lai male Spepharis) Siam. t-rai ” nam, k-lang Man id ka-ren ” [= Kh-yeng]. mi-jang : vir, Newar (mi-sa woman). : an Toung-thu, Ta-lein = Mon. leng husband Burm. len-ja © male Marar. “foera man Mijhu. ” Mon., male Taying. - Mon. #2 * ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. . k-loun man Karen, mi-lo husband Miri. nio-lo > Daphla. k-loe man Ka lu 3 Burman, Sak. m-ru 9 ru (= N-ru), mura ” Stinwar. ma-ro 5 ruot (father Anam loh male (small an.) Kumi, lu-hi " (birds) ah trong PE a neuoi = =6oman Anam on ve ” h-po A a chagha @rh = ong &-tchony kung kang hong hiung yiong ) 3} B) N hy n Hb] ather usband _ Kumi. Bhot. (Gerard). Rob.), Milchanang. Tiberkhad’ 3. The Sikilant Root. Horpa, Manyek, Anam Mijhu Siam Khamti Anam husband Bode. bi] male Hh a) ?} 3 TEL ” father i Garo, (many an.) Burm, + ee | (an.) Khamti. ») Anam, (large av.) Kumi. N, Tangkhul, Kasie. Sifone. Anam i? "4. The Guttural Root. male Py | 7 7 ” Chinese K-h., id. (an) K-t. (inanimate), ,, (birds) K-h. 5 eR st. » Shang-bai, 70 ETHNOLOGY OF THE INDO- PACIFIC TSLAN DS. keen | masc. principle in nature Ch. khon =s*vvir Siam male, *, fa ther, A a (Gerard). T ja-ko husband a pes cha-g r / pawiy Garo (mi-chek aif chek=jik generic). u paternal uncle Bhot. (Gerard), rio 'Tiberkh., kea father” Tiberkh. s-keo male Milch. Feaiuntne Nawes. : 1. The Labial Root. pan jJemale Chinese K-t. pin 4 ey K-h. ‘ 39 (an) fu jin woman, Chin. ma, mo, m fem. part. and pref. Bhot. a-ma—_ mother nh ee Many., Takp., Dhim., Gero, Gurung. ta—mo mo-b-jye wife Bhot. a da-mo Jem. Changlo (Gyar.). moi-bo wife “a e ” vo Pa Anm mi-mo woman Miri. ma Jem. Pe Burm. bi-ma ws an.) Garo, a » mother Bodo. ma—pani fem. (an.) Dhim. — 5 phang , aepaen) Siam. fi wife iam . * don bu woman Anam a-mi mother Burm, mi » Kasia. me » Laufam. Anam mia nife Siam : tua-Inia em. (an.) Siam. “ mia woman Taying M ‘ be-jan «girl =Dhim. a—pe Jem. (an.) Mikir, ai—pi-80 ier na-bi sister Taying (na-fo elder brother). si-wi girl Bingpho, mau jem Anam wih k-mai ém., woman, Mijha, — mai Osea, Anam. ‘ETHNOLOGY ‘OF THE INDO-PACIFIC ISLANDS, na-mu mother Abor. ie he » ‘Taying - 2. The ‘Liguid Raot. nil fem. Chin. K-t. K-h. neu ” ” ad na ? . K-h. nu ¢ woman. Anam e-ne mother Bhot. (Gerard). a-ne » Aka. e=nu A Naga, Manpuri gr. o-nu ‘s 3 wm 44 nu-nu Ps Mijhu noa = pel num-sy& ing pho. mig Py on ys ” (an.) a) Kumi. wet roman Siam ma-na mother Min _ne-ka = fem. (an.) 5, Ry it wife, girl ” ing Jem. Siam. ing-yong mother Namsang. n-yong Se em, (an), ing~ » | 3 3. The Sibilant Root. 2 wife’ Chin. K-t. tsai K-h. taz fem. (birds) K-t., K-h, e’~me gil Horpa. us—res woman Gyarung am—cho woman Bhot. (iho? man Many,). a-zhim sister elder sing—mo cas je fem. (an. ) Dhim. jong mother Mikir. hi-n-jo woman Bodo. mi-chek ad Garo (jik-se husb.). the a Anam. #a-si fem. (an.) Taying. t fon girl sSingphe (! st-wi, comp. sing-pha. “hi-njo Race F a wughter ingpho. int girl © Burman (s’-mé Horpa). mi-sa woman Newar. 4. The Guttural Root. Kk Chin. K-t. (fem. principle in nature. ka éfem. def., Coa : ip rai moman Aare tle Me ETHNOLOGY OF THE INOO-PACIFIC ISLANDS. : kai | Sem. (a0-)'y . — 5. Vosalic. ' yang Chin. K-t. fem. principle in catures _ ying Jem. an ~ miryeng , wife Mui, -m-ang. vs arr _ og-yong mot ‘AMSANE. oe Jem. Siem. aus aia Names or Domzstic anp or some Wiip Antmats. , . Thenames of the more familiar quadrupeds, domestic and wild, are ~ more or less connected in all groups of Janguage that vetain a priraitive and homogeneous character. e same root. coisin applied to several of _ these quadrupeds as a generic name, the species being denoted by 2 detini- tive, or a inatitive, or by the addition of a second substantive name, In the progress of dialectic separation and change, the samic pure roet, or the seme '. compound, has come to be applied to different animals in the various dis- lects ; vocables originally identical have acquired distinct forms and ap- plications by phonetic changes in one or tore of their clements, ec or accessory, or by throwing off the latter; and, on the other hand, vocablee _ originally dissimilar have acquired a close resemblance. Roots primarily - denoting only the sex or age have acquired a substantive character ond _ become restricted in particular forms, to certain animals. Thus words that first signified man, weman, child, were applied to mark the sex and age of the jower animals ; and in some cases, by the loss of the substentiva names which accompanied them and by the acquisition of peculiar phone- tic forms, they eventually superseded these names, so that the same root aay, in the same dialect, mean not only man and the male of a lower spe- Gies, hut the species itself, male and female. By the loss of the specitic name and the permanence of the descriptive, the seme animal may ac- ire distinct names not only for male and fernale, and for the young of ditlerent stages, but for other varieties in breeds or individuals,—as those of eize, form and colour. Thus while a single root sometimes continues to be spplied—joined with descriptive words or slightly varied plionetically—to animals of different species, the varieties of the same species, and even diversities in the same breed, may be known by distinct roots. Secondary forms and applications have been communicated by one dialect to others, 60 that while, on the whole, the general glossary of each group has, by the Japse of time, gained in richness and individuality, while preserving the primitive stock of roots, each single vocabulary has become less homoges neous and systematic in its nomenclature, and hag even lost some of the erchaic roots or compounds. T'o ascertain the radica! stock of names and the primitive method of applying them we must consider the glossary in the agorerate. To ascertain the relations of particular dialects we must consider each departure from the archaie system as a substantive histori- eal fact. {See chop. V. sec. lL, Names of Parts of the Body, p. 208, Names of Donestwated Animals p. 240). ‘ _ Dialectic relations can only be fully understood by comparing vords in groups, comprising:all those thatare etymologically related. But to . form such groups with accuracy demands the perfection of 2 science which can hardly be said to have yet taken a definite shape. We must be.in EFHNOLOGY OF THO INDO-PACIFIO ISLANDS. 73 complete possession of all the dialects of the family, and we must have compared all their vocables not only with each other but with those of the cognate families,—that is with all other human languages. The history of every spoken tonfue ascends to the very beginning of speech, or to the origin of mankind. It contains roots that have come down through nu- merous channels and with various changes of form and meaning. Each root has also, through all ethnic time, flowed in hundreds of con- temporaneous currents, pens Am by self division, diverging fir apart, approaching, touching or coalescing, and again divaricating. The genea- logy of every languaye is hence exceedingly complicated, und will remain a subject of research for centuries to come, We must make a begioning - with imperfect vocabularies, and such partial groups aa they enable us to determine. The results which the first labourers in the ficld may arrive * at will appear insiynificant as the science advances ; but they have this en- couragement that every well considered comparison, however narrow, leads “to a positive historical result. What is learned is a substantial and stable gain, It will afterwards take its place as only one among many eviden- ces of the same ethnic movement or influence, or internal Tinguistio change; and connections that now appear isolated or partial will be explain- ed_as the results of ethnic alliances and events that were not at first sus- _ pected, but which have left other records in the vocabulary. Tho earlier ‘generalisations will be corrected when they have been too narrow or too wide, but. the substantial affinities brought to light will) always remain -among the facts on which the science, in all its successive developments, will be based. With the small samples which we possess of most of the Himalaic languages, we must be satisfied with the examination of a few groups of -words, and each of these exceedingly defective. Not to complicate the enquiry, it will be confined to ascertaining Ist, the ramifications of each rout in all the vocabularies; 2d, the vocables by which each object is at resent known in the different vocabularies, and the connections thereby ‘Indicated ; 3d, the affinities of each vocabulary singly. The relations indi- ‘eated under the first head are to a great extent archaic: they must have been formed during a great lapse of time; and many of them must belong to the earliest phase of human speech. The history indicated by these aflinities is complex and must embrace many and great ethnic changes and movements. The affinities examined under the 2d head will throw some lieht on the later ethic movements ; and those brought together under the 3d will help to show to what extent each dialect was affected by those movements, and what its modern and its later pre-historical melations to the other dialects have been. As the Chinese is, on the whole, more faithful to the primary system of nomenclature thin other languages, and the Himalaic family takes the next place in the order of glossarial disorganisation and concre- tion, it will be useful to take «a few illustrations from Chinese of the use of generic names. The root ngau—dialectically varied to gu on the guttural side and to niu on the liquid—is applied, with specific qualitives, to the Cow, wong ngau (yellow ngau), Byffuloe ui ngau (water ngau), Yak man ngau, Zebu fung ngau, and Rhinoceros sai ngau. ‘The root yeung (yang &c.) isapplied to the Sheep _ min yeung, Goat shan yeung (mountain yeung) or tso (tsau &c.) yeung, Chamois ling yeung, and Antilope gutturosa wong yeung (yellow yeuug). 74 ETHNOLOGY OF THE INDO.PACIFIC ISLANDS, ‘The root shu is applied to different kinds of Rats lo shu, chuk shu, tsong shu, ku shu &c., and Mice shik shu, tso shu &c., tothe Squirrel sung shu or wong shu, tothe Weasel yau shu, to the Mole tin shu or an shu, to the Marmot to put shu, and to the Bat fi shu. The following appear to be the roots now eurrent in the Tibetan voca- bulariesin the names for the Cat, Dog, Hog, Goat, Monkey, Cow, Butfi- loe, Elephant, Horse, Tiger and Monkey, the names of other quadrupeds not, being contained in the short, Sifan lists. I. (a.). The labial with a slende# vowel, mi, bi or byi, pi, me’, is an element in the names for the, Caé in Bhotian Coes and T: )s for the Cow in Horpa and Manyak, and for the Bujfalve in Manyak. The Ho me’ suggests that it may, in an older form, have had a guttural final. The Thochu Vi Hog may be a slender Sitan variation, of the a phag ze &e, like ri road for the Bhotian lam, brisnake for the Bhotian brul : (see Sec. 2), so that it cannot be considered asa fourth application of the archaic slender root, ; (b.’) ‘Theaspirate Jabial with the @ vowel is applied tothe Hog. It prperes a guttural final.in Bhotian, but has lost it in Horpa, Manyak and Takpa,—phag, phak, Pha, vah, wah. | (c.) nother broad form ixapplied to the Cow—ba, pha, wo—in Bho- tian, Takpa and Manyak ; and to the Horse—bo—in Gyarung and Manyak. (a.) Cut. byi-la Bhotian wr., siemi Bhotian sp., Sok-pa, syi-m-bu Takpa (-bu, the Bhotian masc. postfix as in the Bhot. pre-bu monkey). Comp. bi-thi rat Bhot., pi-chru-ba Changlo. Cow. ngau-me Horpa, wo-mi Manvak. Hog, pi Thoehu. Bufl le. ding-mi Manyak. The word is not-given in the other Sifan vocabularies. i ~ vd Laer le-phe Manyak (Chinese lo-fu &e.).] “ b.) slot phag Bhot. wr., phak Bh. sp.; Horpa vah, Manyak wah, akpa pha. ° ec.) Caw. ha Bh. wr., ba’ Takpa. pha-chuk Bh. sp., wo-mi Manyak. Lerse. bo-rvo’ Gyarung, Manyak, b-ro’ Manyak. Chinese has mi stag, ti colt, ma horse. II. The liquid root occurs in names for the Cat, Goat, Cow, Bujffaloe, Elephant and Horse, wt. byi-la Bh. wr., chu-la Horpa, lo-chi Thochu ¢a-rhu Gyarung. Goat. ra Bh., Takpa. . ' he Com. sa-lo Sok-pa, lang dang Bh. (Rob.), lang-gu-bull (Pitti), pa- lang com (ih), a Sir ; ; Buffaloe. ding-mi Manyak. Elephant. g-lang-chen bh. wr., lam-bo-che Bh. sp., Tha-bo-che Sok pa, " la-mo-che Horpa, lang-chhen Gyarung, Takpa. The second element in the compound is Chinese. The native term lang is obviously that used for the Cow and Buffuloe, the Chinese name being added us the specific one or qualitive, or conversely. Horse. whi, ryi Horpa, ma-ri Sokpa, ro’ Thochu, bo-ro’ Gyarung, Ma- nyak, b-ro’ Manyak. [Chinese has lu, lo ass, lau hu, lo fu éeger, lau shu, loshu rut, luk, lu deer, lok, loh to came?). : > ETHNOLOGY OF THE INDO-PACIFIC ISLANDS: 75 - Ill. The sibilant and dental rootis applied to the Cat si, syi, chi, chu, cheu, Gout chhe, tsah, so, cha, chang, Cow chuk, Bull zyah, zi, Dog sha, tu’, Horse ta’, Tiger tak, te. A. The sibilant. >. ‘Cat. si-mi Bh. sp., (syi-m-bu Takpa), chu-la’ Horpa, lo-chi Thochu, ma—cheu Manyak. " yp Goat. chang-ra Bhot.,-chhe Horpa, tsah Thochu and Manyak,. ku-so Gyarung, cha-pu, chya-pu (the goat of “the northern region of the sub- Hitnalayas ” Hodgson, J. B.AyS.XVL, 1020), |, € Cow. pha chuk Bh. sp. rend wry Buil. zyah Thochu, nga-zi Manyak. » oF ae - Dog.*k-sha’ Manyak, ‘This appears to be a variation of the dental found in Horpa ka-ta’. ST ste" 5 B. Thedental.. ~~ , Horse. r-ta Bh. wr., ta sp., te’ Takpa, Dog. hata’ Horpa, k-sha’ Manyak. | Tiger. s-tag Bh, wr., tak sp., s-tak Horpa, tée Takpa. | - id IV. The guttural and nasal roots are applied to the Dog, Hog, Tiger and Cow, and appear to be all Chinese in their inimediate aihuitien a Dog. khyi Bh, wr., uyo sp.; khwa Thochn; khi Gyarung, Takpa, (Chi- nese kinen, hun, kau, keo, Kin koi-ra &c.,, Mongol nhoskhwe, na-koi &e.. Hog. ki Gyarung {? chi, ti, tio, chu, tu,du Chinese]. |, Tiger. khé Thochu, Kong Gyarng, [kim Gyamai, hu Chinese]. Com. gwa Thoehu [? Chinese ngau, guj], wll Cow. ngau-mé Hora, nye-nye Gyarung, Chinese, K-t. ngau, K-h. niu Hok-kien gu; cow hwang niu, wong ngau, vong ngiu, (hwang, wong, vong, yellow) &e.; “dul mau niu; nia ku, ngan ku, ngauw kung: &e, (mau, ku, kung, male); buffuloe (water-cow) shui niu, shui ngau, The occurrence of the same root as an element, in different names, and its change of position trom initial to tinal, is, in several cases, expatined by its possessing, or having primarily possessed, a sexual power. ‘he Inbial retains its sexual power in Tibetan, The sibilant is masculine in its ap- plication to the Budi in Thochu and Manyak The liquid does not appear to be current as a masculine root in Tibet, but it is preserved in Bhotian pa-la dather, and in the southern languages of the tamily it is common in the Tibetan forms lang, ra, ro, ri &c., as» masc. substantive or servile, On comparing the Tibetan names of animals in which it occurs with the southern ones, it is clear thatit must ‘originally have been a mase. root in Tibetan. Hence byi-la, chu-/d, Wechi cat and chunp-ra goat, are radically mase.; while si-mt, wa-cheu cat, wo-mi cow, are radically fem. In Tibetan the sex qualitive may either precede or follow the substantive word. In the course of that glossurial metamorphosis to which language is subject, the sex name has, in several fasthndos Mikoiies a substantive one, Ba, wo and lang wre now Cow, ra Gout, rhu Cat, phag Heg, and ding Buffaloe. With the light thus thrown on the Tibetan names we can pro- ceed with more certuinty to examine their history and relations. I. The labial is one of the primary zoological roots. It is also prima- 76 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. in Seythic, and with a similar range of application. Cat, Turkish mi- shik, pi-shik, ma-chi &c., Ostiak mi-sak, Mongol mi-choi, mi-i. ‘hat this was a very archaic application—perhaps the earliest, unless the morse was the first of the house quadrupeds (pipi, mush, mus, pisse kc, &c.)—is shown by the prevalence of the same root, and of the same combinations, in other families, including Semito-African bi-s, mus &c., Caucasian and shtu pish-ik, and Dravirian pi-shi &c. (See chap. V. sec. 11, Car). Similar vocables for the mouse are as widely spread; and those for the dog, goat, sheep and cow are the same (e. g. cow Ugrian mis-ye, mes mus, wys &c.). It is much more probable that the root was extende from the smaller to the larger animals than the reverse, The order was probably from the mouse and rat to the cat, and then to the ddg, goat, sheep, hog, cow, and buffaloe, as they were domesticated. The mouse and rat would be the first quadrupeds to become inmates of human dwellings, and they would be the baits that first attracted the cat and the dog from their coverts and reconciled them to man’s companionship.” The form and the free position of the Tibetan mi &e. in the different compounds in which it occurs, show that it is not a derivative from Scythic. 1t must be equally archaic in both branches of the Tibeto—Scythic stem. The ultimate source, or primary meaning, of the root appears to have been man, male or female. It was afterwards applied to the males or females of the lower ani- mals. In the Bhotian si-mi Cat, agape wo-mt Com, ding-mi Buffaloe, mi has probably its feminine function. e Bhotian and Gyarung mi Man is the same form of the labial, It is also Ugrian mi, ma, mis, mes, mias, mas, muz, min, man, mar, mur, &c., and in that ay may also be the source of the similar names of domestic animals. .The Bhotian pha, pa, ba &c. Sather, and ama, (ma, mo &c.) mother, have not only been applied to ani- mals, to designate the sex, and thus originated substantive names, but have come to be used as detinitives with inanimate substantives. The slender form is not current as a definitive in Bhotian, Horpa or Gyarung, but it is found in 'Thochu -mi, -pi, Manyak -mi, -pi, -in, Lhopa -be, and in Gau- getic dialects. The Bhotian byi of byi-la Cat although primarily identical with the servile sexual bi, mi &c. has evidently had a distinct history. It presents itself as a root used substantively for the Cat, and that this application was very archaic appears from its being found both in the Chinese glossary and in that of the Scythic, and most of the other Asiatic formations. La appears to be the masc. liquid root used postfiually as in pa-la jather, chang-ra goat, chu-la cat. Ba, Cow, is the same root as the pha in pha chuk, Serpa has ma chu. Ba or pha and ma are identical with the Bhotian sexual labial definitives and postfixes, pa, ba, po, bo &c, mase., ma, mo &c, fem, In the Lhopa dialect of Bhotian bha is the bull, lang the cow; the compound baclkagr pa-lang is used in some dialects for the com. In lang-bo- che, elephant, lang takes the masc. postf. The woof Manyak is also ra- dically the same mase. def, It has the same form in Thochu as a postfix, mar-wo Bird, nyag-wo Crow. Bhotian has bo, vo, pho, po. * Ihave found a somewhat similar remark in Admiral Schischkoff’s Vergleihendes Worterbuch ii., 224.—Referring to the identity of some widely prevalent names for the Se ete and Mouse, he explains it by saying that they must have been the first domestic animals. ° ETHNOLOGY OF THE INDO!PACIFICG IsLANDs. 77° . Both ba, pha, or wo and lang must have been originally applied as quali- | tives to the bull, with or without another substantive root conjoined, It is probable that chuk preceded them as the substantive and that chuk, ba, chuk wo, chuk lang,—or ba chuk, wo chuk, lang chuk,—were current like’ pha chak and ma chuk. | The broad form of the labial root for Afan, ba, pa, wa, wo, bu isa very common one in the zoological vocabulary, with different applications, and - with or without a final consonant, s, 1, k &c. It has frequently a maseu- line application, both when used for the bull and ox, and for the males of other quadrupeds. Itis applied to the Bull in Seythic, buga, buka, Indo- Enropean buka, bugu, buk; wol, wul, bull: bus, bos, bu, be, wo ( ic mus, mis, wo-gol &e, &c.), and to the Ow in Circassian, wwe, be, b’by. The Tibetan phag, pha, wa &c. Hog, is distinct in form from mi and byi, and is evidently a very archaic variety of pa, ba &c. The labial does not a to retain a similar form with the same application in the glossaries of the other t families, but it is still current for the goat, sheep, ox, deer &c., in Seythic, Indo-European, Semito—African &e. and it is trequent- ly applied to the mule. It is not Chinese in any of these spplications and it therefore belongs to the Scythic side of the basis glossary, but without be- ing a derivative from Scythic. It is one of the distinctive archaic vocables © of Tibetan. Its relations to the examples of the same root found in other families will be considered when the Ultraindian forms and appications have been given. II. The liquid root is one of the primary or most archaic of the Tibetan, and hence enters into many animal names. It appears in the form lang, la to have become one of the proper native words for the Cow and to have been afterwards applied to the hant; in the form ra it has become a substantive name for the Goat; and in the form rhu for the Cat. A different form of the same root, or a primarily distinct liquid root, appears to be the native term for the horse ro (probably a softening of rok), rhi, ryi. This root has also a masculine application. Its primary meaning is man, male, and it oceurs extensively in the Chino-Himalaic vocabularies in masculine terms,—man, husband, father &c.,—and as a masculine deti- nitive with the names of animals, in various forms, nan, lang, leng, la, lu, lo, long, log, ru, ling, ri, ren &c. The Bhotian word for father, pa- la adds it to the labial root of old Bhotian. In ehu-la’, lo-chi Cat and chang-ra Goat it has probably the same function. It is a wily Serced root, for many—Chinese, Turkish, Mongolian, Semitic, African and Drayiro- Australian. The more immediate affinities of the root in its application to the Cow are found in the Ugro-Semitic band. Fin Cow loh, loch, leh, or, er, la,— loh-ma, loch-ma, Jeh-mu ; Caucasian Or, her-ga, hor-g, or-j, or-2, er- dse, hyer-ko, Ugrian yzh-la, (comp, ish, ysh-kuzh &c.), ok-or, er uk-ys &e.), ‘Tungusian or-gol (comp, gol Turk.); Indo-European aur-ochs ochs is Seythic), ur—-vs; Caucasian ol, al (Lesgian); Semito-Nilotic lo-¢r ara, lahi-then Mahrah, lahe-mi, la-me Tigre, la-m Amharic, Harragi, Gafat, lu, lo-wa, he-lo-a le-wu Agan, la-mo-wi Gafat, la Danakil, loh Saumali, lo-ni Galla, la-n Tumali, ai-ra Dalla ; dull onra Danakil, Amha- ric, uhur Arkiko. The Semitic form, as in so many other instances, must be directly connected with the Caucasian ; and from the Lesgian ol, al, and the Mabrah and Tigre lahi, lahe, it is evident that this 8. E. group a ° 78 ETHNOLOGY OF rie INDO-PACIFIC: ISLANDGs is more immediately connected with the Fin than with the Tibetan. Tn the Chinese, Scythic, Indo-European, Caucasian, Semitic and African, families the prevalent names for the Com contain different roots. The Tibetan name is therefore independent. The Fin loch, loh is a variety of an archaic form applied to the Horse in Ugro-Tibetan, and to the Deerin Chinese, and probably connected with Tibetan forms for the Com only through the derivation of both from the same archaic Asiatic masculine, Ra goat ( ra-ma_f., ra-ba m.) appears to be the same root. The word i8 only found in Bhotian. It is not Chinese. It seems clear therefore that Ya was originally derived trom chan-ra, a masc. form of the root chang. The liquid root is applied to the goat in Semito-Nilotic dangruagees ha- run Mahrah, @-ron Gara, ho-rar, ri, re Galla, ‘illa Danakil, arre Sawali, araha Bulanda, [eru, eri, ere lsoama &c. is probably a contraction of the, com. African e-wure, wuli, e-puri &c.); arre Saumali. But these names do not appear to have any direct connection with the Tibetan. i Ro Horse is Ugrian, and the guttural is preserved in Ostiak log, loch, low, (in other Ugrian langnages lo, lu, lyn; wol, wyl, wal; lowu. The Ostiak guttural form corresponds not only with the Sifan rd, but with the Chinese luk, lu, 16 deer. Ih E. Tibet and Siling it is also used jas a generic vocable for sheep, two kinds of which are known as ha-lnk and pe-luk (Hodgson J. B. A. 8. xvi, 1008). The root may also'be contained in the Samoiede bo-ra, Koria mol, mar, Tungusian mo-ro-n, mu-ri-n, mu- ri-l, Mongolian mu-ri-n, mo-ri—this liquid form corresponds with the Horpa rhi. But. in this prevalent Tartar term the liquid is either one of the common finals taken by monosyllabic roots in the Seythic family or it is a sex postfix, the root being mo, mu, cognate with the Chinese ma. In the Ugrian and Turkish ala-sha, Turkish lo-sha and Caucasian uloh ( Misjejian ), it may be a substantive root, as in the Ugrian lo &e. From the occurrence of the liquid as a prefix or postfix in the human sex names and in several names of animals in Seythic Word itis Lege servile and masculine, or was so originally. Man F'n al-maz, wl-muz, »U- grian /o-man, iri-golos; Husband Vin ol-ma, ul-ma, we-lo-man, Mongolian ere, Turkish ire, eri, er, ir, er-kek, er-in, ar-mi; Cat ir-my-shak, e-gek~ myi, Turkish (so ata-p-shik, ata being father); Dog koi-ra, koi-re, koi-ru Huy, al-tschip, al-ship, il-tschap (also tschip ke.) Yeniseian; Ox or-gol ngusian. If we consider the labial’ as the substantive root in the Scythic mo-ro-n, mu-ri-n, and ro, ri as masc, serviles, which in the Ugrian og &c. have become substantives, the same view must be taken of the Tibetan bo-ré, rd, which are thus placed in the same class with pa-la, pha; chang-ra, ra; chu-ld, lé-chi, rhu; and perhaps r-ta, if the Bhotian,pre- fixual r-, I-, s-, z-, b-, p-, v- be, as is probable, contractions of the origi- nally masculine definitives la, ba, si &c. As the form ro is no longer cur- rent in the southern Scythic languages, it must belong to the archaic U- grian basis of Tibetan, like a large proportion of the other vocables. In the Ultraindian vocabularies the ine liquid root retains the guttu- ral final in several dialects. Indeed all the Scythic forms are found,—zi, rin, ron, log, lok &c. Til. The sibilant in its a plication to the cat appears to be archaic, and native. The root is found in Scythic vocabularies for the Mouse (e.g... ETHNOLOGY OF THE INDO-PACIFIC ISLANDs.. 12 ‘Purkish shish, zis, shi &c., in shish-han, shi-kan, shyshi Ke.) and a simi- lar root.is combined with the labial root in the Ugrian me-tschik, ma-tska, mi-sak, and Mengal pi chet (also mii) eat. : ' , In the Tibetan vocabularies (Horpa as well as Sifan) the sibilant root is the prevalent one for goat, chang, ehhe, tsah, so. It is a common Seythie root, occurring in names for the cow, horse, dag, hag, mouse and sheep. It is doubtless applied to the goat also, but most of Klaproth’s Scythic yoca- bularies omit the word. other highly Seythie glossaries it is applied tosthe goat. Itis the prevalent Caucasian root—zé, tzia, etcha, chan, gu-ku, tn-ka, ze-ki, ka-za, HE. Caucasian; gra-se Misjejian, b-zhen, zhi-ma,; Jireassian; tcha Georgian, sa-ga, s- zan, zan-ek Ossetic; Indo-Kuro- pean zie-ge, chha-gal, chha-g, a-ja, chhe-lo, tsa-wul; Semito-African ne- ' ae, bi-se, e-ge-so, t~mi-shu, fi-zo [fis, fus, sheep Ossetic], de-sha, sidsy sikh, é-su, mbo-zi, si, si-na &e. » The same root has as great a range in its application to the cow. The Tibetan zyah, zi, chuk, are Seythiec in their immediate affinities. Chuk ia Tungusian chyu-kun, hu-kur, ku-kur, Yeniseian thu-ga, tu-k &e. The root has the same application m the Ugrian ish, ogh &c. oa [Indo—Eur. oths, ox &c.], Mongolian shar, zar bu//; Caucasian is, 0s, ots, stu, ust, n-itz &c.; Indo-European ochs, ox, oss, osse ke. All the applications of the sibilant appear to be Seythic in their affini- ties. Chinese does not use this root for the cow, goat or cat. It is applied to the Mure shie, she, Hog chu, chi kc., Musk deer she, Mouse shik, shu (as in Turkish &e.), Rat shu, chuk; and inthe same form to the squirrel and weasel with qualitive roots preposed. The Chinese chi, chat, stallion, appears to be the same root in its masculine Tibetan, Scythic, Caucasian and Indo—European application. The dental root is, in many cases, the same as the sibilant, and hasa similar range. Asa name for the Horse the Bhotian r—ta, ta is cognate not only with the Turkish at, ut, but with the Chinese shie, she; Seythie sha of ala—sha (Ugro-Turk.), Indo-Europ ad ly tzi, Cauca sian shu, shi, che, chak &c., Semitic sus, has-on, his-an, African eis, es, sy, si, su, so, dsu, e-si, e-dsi, a-shi. The Bhotian ta appears to be an archaic form. It is found (reduplicated like s of the Hebrew sus) in the Dravirian and N. Indian tata, tatu. Exactly similar terms in t amd s or sh are widely current names for the dag, hog, and ow, he Horpa ka-td, Manyak k-shd, dog correspond with the Turkish eda, it, ot, Korivk a-tan, a-tar &e., Kamschatkan ke-tan, ko-sha &c., Aino stah-pu, Yeniseian tzi, i/-tscha, tachip, ip &c. The Bhotian and Horpa stag, tak, Jiger, is 1 consonantal and probably more archaic form of the same root. From this.form, the abrupt accent of ta and’ sha, and the application of the root to the dog, horse and tiger, itis probable that the root was one of the primary ones of the Tibetan glossary. ‘Like the labial and liquid roots its primary application was also to Ma and most commonly in the sense of father, Progenitor &e. It occurs Me many families in reduplicated forms tata, pet he &e, In the Seythie vo- cabularies it is equally common with the labialtoot as the word for Father. Usrian isi, ese, atte, ata, tato, tatei &c., Turkish ata, ate, asio, M i etschi, ruchige &c., Japan titi, tsitsi, &c. &e. The reduplicated Seythie. form is also Indo- a er and Zimbian. In the Himalaic family it does + Bot appear to be one of the primary and prolific roots, but it occurs in 80 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, , Horpa v-zih man and Manyak chho’ man, which show the same vitriation from the palatal and broad to the purely sibilant and slender form that is seen in chu, cheu, chi, &c, in the names of the cat, in chang, tsah, so, ehhe in those for the goat, and in zum, tyu, siin those for the monkey. The broad form of Manyak is Ostink cho, choi, but the normal Scythic form of cho is the guttural ku. — | IV. The guttural and nasal roots do not appear to jhave been primary and prolific, unless khi dag and ki hog be both native, and the former only primitively eonnected with the Chinese. From this see survey of the Tibetan names of the domestic animals, we infer that labial roots now having the forms bi, mi &c.; phag, pha &c., and ba, wo &c.,—liquid, now having the forms lang, la, lo, ra, rhu, ro luk,—sibilant and dental, now in the forms si, chi, chu, cheu, chhe, cha, chang, tsah, so,—and dental, in the forms tag or tak, ta, ta,—were among the primary roots of the family. Of these the earliest to acquire a specific substantive meaning appear to have been the labial byi in its application * to the Cat, the labial phag inits application to the Hog, the labial bo in ite application to the Horse, the sibilo-palatal in its application to the Gout and fom, the aspirate-cuttural in its application to the Dog, and the dental and © aspirate in its application to the Dog, Horse and Tiger. The labial in its other applications, and the liquid, appear to have remained servile to a comparatively late period, and they probably still retain their sex function in most of the names in which they occur. The primary roots connect the Tibetan or Himalaic family with the. Scythie as dialects of one proto-Scythic monosyllabic glossary, distinct from the Chinese, but having also affinities with it. The separation between the Chinese and Scytho—Tibetan vocabularies must have taken place ata much more remote period than that of the separation of Tibetan from other proto-Scythic vocabularies. At the latter period several forms of the common roots had acquired specitic applications, which they have re- tained in Tibetan and in several of the widely diffused Seythic and Sey- thoid vocabularies of the Old World. Others again are si to Tibetan, and indicate the great antiquity of the separation. This is also proved by several of the common forms being best preserved by languages now widely removed from Tibet—as the Ostiak. In speaking of the period of separa- tion it is not intended to limit the connection to one age, There may have been successive contacts between Scythic and Tibetan vocabularies in archaic as in recent ages. The only name that may indicate an archaic connection with the Chinese nomenclature is the guttural root in its application tothe Dog. The other - radical Chinese names are different from the Tibetan. The names for the Cow, Horse, Sheep, Cat, Hog, Tiger and Monkey are quite distinct. A Chi- nese root for the Deer is the same as the Tibetan for the Skeep, but this is one of those primordial affinities that niay rank with those of the pronouns. ‘The other Chinese names found in the Tibetan vocabularies are evyident- ly intrusive and compratively modern. Some have the forms of the an- cient Chinese phonology, and some the emasculated Kwan-hwa. Like many other Chinese words in these vocabularies they prove that the Chinese race is that with which the Tibetan tribes have been longest and most intimate- ly connected in the latest era of their ethnic histery. Broad Chinese names — or the Cov are found in all the vocabularies along with native ones, saye © ETHNOLOGY. OF THE INDO-PACIFIC ISLANDS» al in Gyarung which has the modern or Kwan-hwa name only, The Chinese name of the Llephant appears to be annexed to a native root in all the: dialects. The Chinese name for the 7iger is found in the wutturalised Gyami form of Kwan-hwa in Thochu and Gyarung, and in the old Chinese form in Manyak. The Gyarung name for the Hogand the Manyak name for: the Monkey appear to be corruptions of the Chinese. *The 2d step isto examine the nomenclature of each animal, with a view to.ascertain the extent of the dialectic divergency. : The Cat has five names, 1. byi-/a Bhot. wr.; 2. si-mi Bhot. sp., Sokpa and Takpa; 3. chu-/a’ Horpa, chi-le Thochu; 4. ma-cheu Manyak ; and, 5. ¢a-ra Gyarungs In these names the sibilant, substaiitive and the liquid servile are the most prevalent roots, and they connect all the dia- lects. Special connections exist between Horpa.and Thochu, both pos- sessing the substantive and qualitive roots combined in the same order, though differing in form; between Bhotian and Thochu in the slender form of the substantive; between Horpa and Manyak in its broad form; between Bhotian and Horpa in thea, and between Thochu and Gyarung in the o, u, of the servile, Old Bhotian in its use of the slender labial as the substantive, is peculiar, the Sokpa and Takpa being obviously deriva- tives from it. The Dog has 2 or 3 names, 1, khyi Bhot wr., khi Gyar., Takpa, khwa’ hochu; 2. uyo Bhot. sy.; 3. ka-ta’ Hor., k-sha’ Manyak, Here also Horpa and Manyak, at the two extremities of the province, agree. Pos- sibly ta’, sha’, is the primary Tibetan name, and khi &c, a later intrusive one of Chinese origin, The Hog has 2 names, 1. phag Bh. wr., phak Bh. sp., pha Takpa, vah Horpa, wah Manyak, pi Thochu; 2. ki Gyarung ; in which the connection between Hey a and Manyak is again illustrated. The Goa 2 names, 1. ra(the sex qualitive, for the substantive) Bhot. Takpa ; 2. chang-ra Bhot., tsah Thochu, Manyak, chhe Horpa, ku- eo Gyarung. The normal vowel is preserved by Bhotian, Thochu and Ma- n There are other instances in the vocabulary of Horpa affecting ¢ and Gyarung o (and e). The Cow is known by 6 names, 1. ba Bh. wr. (phain 3), ba Takpa, » 1 a. wo-mt Manyak ; 2. lang, ba-lang Bh.; 3. yhu chuk Bh. sp.; 4. ngau- me Horpa, gwd Thochu, nya-zi, dull Manyak; 5. nye-nye G arung ; 6. zya, budl, Thochu (uza-ze bul! Manyak), For this important domestic animal 4 native and 2 Chinese names are current. ‘lhe southern Chinese ngau, gu preserves the archiic broad form, to which the Horpa, Manyak and Thochu nau, nga, gwa are referable. The softened. hwan-hwa niuis the original of the Gyarung nye, through the Gyamineu,nyeu. The Chinese pame is found in the Lhopa dialect of Bhotian, ngo, as the generic, | term, ba being confined to the male and lang to the female, from which it may be concluded that the Chinese name was at one time received into all the Tibetan dialects, . The Elephant is known by the same Tibeto-Chinese compound in all the dialects. The Horse has 3 names, 1. r-ta, ta Bhot., té'Takpa; 2. bo-ré Gyarung Manyak, b-ro Manyak; 3. ro Thochu, rhi, ryi Horpa. The remarkable fact here is that the Bhotian name should be exceptional. Roe ETHNOLOGY OF THE INDU. PACIFIC isbanDSy” @ «The Tiger has 3 names, 1, s-taz Bh. wr. tak Bh. sp., s-tak Worpas téé Takpa; 2. kho Thocha, kong Gyrarang, 3. leephe Manyak. Of these the Bhotian, Horpa and Tukpa words are native. The Thocha and Gya- rung are from the (;yami form khu of the aspirated Kwan-hwa hu, and the Manyak is a mative slender fornt of the original Chinese lotu &e. The Monkey has 3 names. 1. s-pre-’u Bh. wr., she-pri Gyarung, pra’ Takpa; 2. tyu Bh. sp.3 24. zum-de Horpa, 2 0. ti eekennigs wiai-si Thochu,—the Gyarung having the Bhotian form. . The roots possessed by each dialect, and the relation of each to the others will be best shown ina table. Ll have added the names for ish, Snuke, Bird, Crow, Ant aud Mosquito, (See next page). From this table it appears that in the names fof animals comprised in it, there is—when we exclude those of Chinese derivation—a close radical agreement in ali the yocabulariea, the variations being chiefly phonetic. The dialectic relations indicated are :—- f ‘Ast, a very intimate one between Bhotian and Takpa, the latter adher-— ing to Bhotian wheu the other dialects depart from it; and the difference being, it almost every case, merely aslight plionetic one. In its greater vocntc tendency Takpa partakes of the Sifan phonology. 2d, a connection between Bhotian and Gyarung, in the form of the roots for Dog and Fish, in the roots for Monkey, Bird, Crom, and Aat, and in the pretix in the words for Money, Crow (G. preserving the full fornt td, Bhi. has a-), and Ant, The connection is chiefly with the old or written, Khotian, the words for Dog, Moniey, Bird, and Ant preserving the old © * Lhotian roots or forms while the spoken Bhotian has lost them, 3d, a very slight connection between Manyak and old Bhotian, The Main Ss b-ru snuke like the Takpa m-rui preserves the yowel of the Bh. é-rul. 4th, an archaic separation between Bhotian and the other dialects save Takpa, as shown in the forms of several of the roots and pretixes. The special connection indicated under the preceding heads, if archaic, would be is in- consistent withthe early divergence nidicated under thishead, Itis attrilm- ~ table to the dialect of the Bhotians having acquired more or less curreney in the provinces of the otber tribes, during the period when the Bhotians w re ~ predominant, and this must have been while the old phonology still prevail- ed. As illustrations of the archaie separation of the dialects, we may — point to the different roots, or combinations of roots, for Cat, Dug, Horse, Monkey, and Fish, and to the difference of the prefixes in the Bhoto- Gyarung g-rog, ku-rok, Manyak de-rad Ant, and in s-b-rul Bh., kha-b-ri ~ Gyarung Snake. Oth. A special connection between Horpa and Manyak,— Cat, Dog, Hog, Com, and Crow,—and the comparatively slight trace of such a connection between Horpa and Thochu (s—kh-ro, tv-kh-ra Ant being the onlyéxample), — and between Horpa and Gyarung, As this special relation of Horpa to Manyak extends to some other substantive words, but not to the Sara and the mass of the abstract and qualitive vocables, and as the Horpa are kuown to be adventurous alid nomadic, being even now scattered over southern Tibet, it is probable that a Horpa horde at one period mixed with ~ the Manyaks, aud communicated to them a portion of their vocabulary. The intercourse of the Manyaks with the Horpa, however caused, appears + to have been more intimate than with any other ofthe Tibetan tribes, | «< ay NOSF NDO-PACIFIC SLA THE 1 ETUNOLOGY OF | oymbeoyy gt quy 8 MOLD) FT PANT EL ayeug at YW Le Aoxyuo]y OL Iasty, 6 es10H g | gueydorsy 2 aoregng 9 _ MON ¢ quor) > SoH ¢ So & wD I “ “J 5 G - — J ~ 3 - | sou LinYyo-t _ -oyd ~ o-Tq, of dn-ay Bs-na—7| vu-suop-ofs| — mg—uus . + ote cake wI-74) You-oy) — WIYAN? O.ltLY-S DuL-OY}} —- Pll-BO = Yeey) —-you-q-ny| oa—auku 4 > 54-4 By-BYY vad ' uty aXd-atd) evant 0 BATE) nAq THAUL nq L944 al -L1q wal nap} qn.tq- 3 ‘3u nA; ofdu-nyo DYAl ey wou wal wal ney-t esitspiil 8-9} ap-uinz ndy ares yeq- edit 0 ; = 993 eek | ony yry-s aby Sus oa +t J Ot-04 9-04 ot) 14a tye i Bd) uott-Sa wayt-cuLy toqya-du| “* juaTpP-OU-BE | wat[o-09-UTe]) UaYI—FuLy ~ Tue-S UTD} * S32 & aren) quar (1ynq) z-wou (ning unk zt -4 ww-ow “a fu-atu BMS] gu-nest ID ¥q qu} os-n ye a Ds-DUBYO “Wt wl uy yn : rd} ge ua wy Bed my pys-y) yy | PANY Be =n oAn ay be nay wut nyi-vg Ty-o7} — - py-nya is. wu-Is;| p-4Ng 3 “IM edyey, | yudueyy | Sommiy | -nqooyy| — edaoxy usyoyq "| usnoyg — 24 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 7) Fer Aer ae feild eet @ wer ~ ‘6th. The connection between the proper Sifan dialects is not very close ; ‘and they must have had distinct histories from a very archaic period. Each has well marked specialities. The agreement consists in a common, but not identical, soitening of the Bhotian phonology, and in some com- ~ mon non-Bhotian reots aud forms, as in the word for Horse. Thoclu has a slight special agreement with Manyak,—Goat, Bull, Mogguite. In conclusion it should be remarked that, in so far as each of the voca- bularies has received yocables from Chinese or from a sister Tibetan dia- lect, during recent eras, the archaic glossarial relations amongst the dif- ferent Tibetan dialects, have been disturbed and obscured. 7 All the Tibetan roots are found in the Southern vocabularies, They have the same forms, but variations are also prevalent,—some of southern ori- in, and others archaic. The roots have not only the Tibetan applications ut others, which are also, in several cases, archaic, The connection with the Tibetan vocabularies not only embraces all those phonetic and plogsa- rial phases which the existing ‘Tibetan data have enahled us to discrimi- nate, but others which are not now distinctly marked in Tibet, and which indieate thearchuic existence of conditions of the Tibetan language and dialectic peculiarities which are now obliterated. The labial root is applied to the Cow and Hog asin Tibet; and also to the Cat (Kambojan), and Mog (Lau). It has consonantal guttural and dental forms not only in names for the Hog as in Tibetan ; but in names for the Cow, bik, bit, Elephant puok, mag, and Horse puk, mok, The liquid is ajphied to the Cat, Cow, £lephant, Goat and Horse as in Tibetan ; and ulso to the Dog (Mon), /iog (Mishmi) Luyffalve, Tiger and Monkey. thus not only the Tibetan forma la, lang, ra, ro, lo, rhi; but many others, longs rong, ron, rung, rok . e. the full torm of ro), ruk, rat, rak, lut, lok, luak, lak, dak, nuk, roi, loi, ling, li, let, le, ren, re &c. The guttural is applied to the Dog as in Tibetan; and also to the Goat, iger and Monkey. he"sibilant, as;irate and palatal root is applied to the Cat, Goat, Ele- phant and Monkey as in Tibetan ; and also to the Cow, Bujfaloe, Horse and Tiger, It is not applied to the Dog asa primary root, but the gut- tural in this application varies to the dental, sibilant, palatal and aspirate. The dental is, in general, a variation of the more prevalent aspirates sibilant, palatal, aspirate-guttural). The Bhotian dental form for the ‘ger occurs only in two vocabularies, and the same form is applied to the Buffaioe in some dialects. ‘lhe pure dental is not used for the Dog and the /lorse. The aspirates are common roots tor the Tiger, Deg, Horse, and are not distinguishable from those forthe Cut, Monkey, Gout, Cow, » Bujfatoe and Elephant. ‘Lhe nasal, passing into the guttural (ng, ny, ¢), is applied to the Cow, Bujffilve nod Ggat, but it has iu nearly every case a direct Chinese origin, ‘he primaril¥ sexual meaning of several of the roots, and their reten- tion of u merely qualitive function in many of the current names, is placed beyond doubt by the Ultraindian languages. We have seen, in consider- ing the words of family relationship, that the roots applied to males are the labial under the forms ba, pa, wa, va, pang, po, pho, bu, pai &c.; the liquid under the forms lang, la, lung, lu, lo, ru-éc.; the sibilant under the forms shai, sau, chiau, tho, thong, thuk &e. ; while those applied to te- males are the lubial under the forms ma, mo, mu, mi, me, Wal, Moi, mia, . it. rejets the liquid. and appeara as.a simple root or , ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Bor (sometimes. bi, pi, pe. &c.); the sibilant. under the forms si, hi, sa, sya,» twha, chek &e.; and the nasal under the forms nu, num, na, ne, nyong, yong, jong, ing &c, ' Of these the two forms of the labial and the liquid are the common.sex words; and»they oceur most frequently as such, or as substantive words, ° in the names of animals. The sibilant is rare asa sex qualitive. It.is a ve- ry common element in names of animals, but from its rarity as an indubi- table sex term, from its form, and from tho sex words usually joined with. it, we must consider it as an independent root in the existing Himalaic animal vocabulary, whatever it may have been orginally. ; ‘In many cases it is difficult, and in some impossible, to ascertain which of two conjoined roots, both primarily sexual, is substantive, and which ualitive. It also happens, from the cumulative habit of the formation, that a name sometimes contains. three sex, roots,—the one that originally became substantive ; another first joined with itas am. or f. qualitive,, and afterwards losing its.sex meaning and becoming definitive or concre-— ted; and a third superadded to mark the sex Ani Thus. the mase. root lo. applied. to the Elephant took the mase. labial pref., and on thia,con- ereting with the root, p-lo, a,sex postfix was added p-lo-, The root tse» applied to the. Com took the masc. qualitive ha-ru, and this concreting into, a postfix, the fem, form became ma-tsa-k—-ru, equivalent to “ female Bull”, Ifmacte, first concreted, ma-tsa-k-ru must originaily have been applied to. the Bull (“male Cow”), In several instances the same compound of two sex, roots changes the fungtions,of the roots with the dialect, or with the ap- plication. ‘Thus in such a word as la-mi or mi-la, the labial must be con« sidered as substantive in one application, because it is so in dialects where ith distinct serviles, while in a different.application the liquid.is obviously the substantive. In _ marking the qualitiye, roots in the compounds I have been guided by a comparison of vocabularies and by general probabilities in each case, but I am far from.confident that a larger acquaintance with the glossary of. the formation will establish the,correctness of my analysis throughout. The following appear to be examples of the qualitive use of the sex roots. Whether in a particular dialect, they retain the original sexual meaning or have sunk into definitives absolute, or those marking a class of animals, can only be ascertained when the existing habits of the dialect are better known. When the form agrees with that of the current sex words, as it - does in some of the dialects for which we have grammatical detuils, it pro- . bably retains its masculine or feminine function even when it has become a prefix or postfix. 1 give a few names in which the sexual or definitive use of the qualitive appears to be preserved. For the Cat we find la-mi; ja-mt, me-sa, mo-chi, nono: in whith the two Tibetan roots are conjoined with a fem. def.; and newai-pai, pa—kwai, ha-ngau-bj in which a Chinese root has masc, definitives. For the Dog we find choi-ma; for the Hed Jans m., ti-li prob, f. (ti=si), cha-ruk f-; for the Goat pu-run m.; mi-k-re, me-te-le m.; chheng-«r, cho-/e, tso-be, sha-bam m.; mt-cha, md-dze f.; for the Cow chu-ma, man-chu, ma- su, mi-thu, shau-me f., cho—rong, cha-ra, si-ra m., mwoi-tom m., ma-tom_f. ma-tso-i-ru m.; for the Buffutoc WOLLO, pat-nil, pa-na, pu-ren M.; a ag ETHNOLOGY OF THE IN0O0-PACIFIC ISLANDS) +, ; * Othe ‘ for the Elephant mag—wi, woi-pong, p-lo-bi m.; for the Tiger mia, ma- sa, ma-cha, sah-nv, cha-nu_f; khu-bui, khu-bi m.; for the Monkey si-mai, mai-nak, me-nak, mo-kha-ra, si-be f., le-be m. »The nasal fem. root occurs rarely,—~lok-niu Elephant Tablung (neu Chi- nese), sa Tiger Namsang, sah-nw Mulung, Tablung, cha-nw Joboka, chia; mv Muthun (nu Chinese, Kumi). Inthe Angami te-nu, M. Angami ta- nu Goat, Nogaung ta-nu, Angami and M. A. nu-no Cat, it appears to have become a substantive name, ta, te &c. being the most common pre- fix in these dialects. The sibilant is so common as a root that it is difficult to distinguish in what, cases it is used as a sex qualitive, and the difficulty is increas some of the masc. and fem, forms closely resembling each other, The following appear to be examples of substantive applications of the Bex Toots. ‘The mase. labial is applied—in the forms pai, bai, woi—to the Goat in Mijha Mishmi /az-pai, Mon kha-bai, Toungthu bai, Bongju woi; to the Com in Kumi kha-boi; to the Elephant woi, mwi ;—in the forms mi, bi, “me, bhe to the Goat ; to the Cow bi, mih, pi, bit &e.; to the Monkey be, we, pi; and to the Cat mi, bi, be; in the forms wo, po, mo, bo, woa to the Cow; in the forms me, moh, pang to the ey teed ; in the forms vu, phu, pong, mu, mun (phang fem. in Lau) to the hant; man, mang, beng to the Horse ; wun, myu, mang to the Monkey; in the form muk to the Cow ; mag, puok to the Lleyhant ; mok, puk to the Horse ; muh to the onkey. ‘Pie “ndee. liquid is applied to the Dog in Mon ka-la, to the Tiger in Mon and several other dialects k-la, si-ra, sa-rong, det to the Goat’ in several vocabularies k-lung, b-lang, ke-l, [from mi-k-re]; to the Cow in Karen /-lo and Mon /a-rau; to the wots in pont dialects Jang, long, loi, lui, roi, la, le, reh &e., to the Llephant p-lo, lok, luak ; to the Horse rang; to the Monkey lan, lak, nak, ra, rhu, ling, ri, re. I tabulate some identical forms showing variations from qualitive to substantive applications, mim-dve Cat, Kumi, kha-dvt Con, - woi Goat, Bongju. wr-tom Con, Songpu. woi-pong Llephant Maram. khu-dui Tiger » ham-—pai at Mijhu, pai-noh aloe Kumi, a-pang Buffaloe Khari, ~ pony lephant Manipuri gr. phang ,, Llephant fem. Siam. phsak IIeg com, we aw Ao gs ; io W ; q 7 Mes 4 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. » wok ~ bok puok aa—puk ka-phuk mok moh | te Hog i Elephant Horse . y9 Buffaloe Hog Buffatoe Namsang. Tangkhul. _ Lungkhe. Mamesang, Burman, Limbu. Mon. . Mon. Kasia. Bhotian. Tiberkhad. Limbu. Anam. Namsang. Kasia. Kasia. Sak. —— hamp. * Miayake o Serpa. Bhotian. Maring. Mikir. Toung-thu, Anam. Singpho. cee epang. Milchanang. ” 7 Daphla. Daphla, - °°? =f ‘ wats / > 88: ETHNOLOGY OF THES INDO-PACIFIC ISLANDS, mai nak Monkey Muthu. pr k—lak ; Silong. rte lok-niu Elephant —‘Tablung. z, Cat. ‘* y; (a) The old Bhotian byi-la is only found in the Bh. dialect of Lhopa, i-li, but it occurs in a contracted form in the Serpe ee Sunwar be-r—mo. Murmi has ta-wa-r and Gurung na-wa-r. Serpa and Sunwar form is also Male ber-ge and Uraon bir-kha. Similar names are preva- lent in Telugu, Gond, Kol, and in the Sanskritoid age of Northern India, bir—al Bengali, bil-al Gond, pilli Telugu, bulau Maldivian, billi, bil- lao Hind., bilai, billee Sindhi, bra-ir, bra-ur, Kashmiri. A _similar-word - ~ for the Ziger in Dravirian, pili Tuluva, piri Toda, puli in the other ialects. The exceptional Deoria Chutia midige is probably midi-ge from mi-li- ge, biri-ge (comp. Male ber-ge, Toda piri, Tuluv. pili, Hind. billi). ~ (>.) The Bhotian form is also found with the masculine particle pre~ fixed in the Luhuppa /a—mi, N.Tangkhul /a—me, in which the root has the same form as in the Bhot. byi-la. The pian: Tablung and. Mrung a-mi, Kyau mi are probably contractions of a similar, term. : (e.) e common Yuma min &c. of min-cho, mim-—dei Kumi, meng, mi Kyau, ta-myin Mru, min Khyeng, min-yo, tha-mi-yo, sa—min-yo Karen, (? heing Sak), found also in Mikir meng (Kyau) and Ahom men, resemble the Bhotian mi, and do not appear to be vaniatious of the Chinese vocable. But the Kumi cho of mia—cho and the Karen yo ap = omer that it was originally the fem. qualitive in this group also. [iSee Il. The Gyarung form ta—rhu appears to be the immediate parent of the broad Burman /-roung, k-young,—the Burman group having strong special affinities with Gyarung. Ill. The saber root in the prevalent slender form LChinete, Lhopa, Dravirian, N. indian) and with the r ofthe Nipalo-Vindyan forms, is found in the eastern sub-Himalayan band. disjoined from the labial, or with a distinct root or detinitive interposed, Taying Mishmi, ma~a-—i, na-dza-ri, Abor-Miri men-da-ri, ka-da-ri, men-ku-ri, Changlo dai-ni [comp. Garo ja-rang, dai—rang all}. The ma-ja, ka-da,&e, Mishmi Abor terms may have been double prefixes, in accordance with the Ty habit of heaping particles, which is well preserved in some of the Abor directives (ante p. 16), and if so they. probably served to distinguish the names of the Cat from those of other animals having the some‘rovt, and one of the definitives. Thus Horse is ku-ri in Tengsa Naga ; be is sa-loi, nga-loi &c. in some Manipuri dialects; and in Mishmi the root, with one of the prefixes, occurs in leh Mog (wild), dali ib. (domestic), ta—loi Byffaloe, The Lepcha a-len is probably a contraction of a Bhotian or Mishmi-Abor form, the former probably, as the adjacent dialects have wa-r, be-r. The Dhimal men-khou is the Abor men-ku-ri with the liquid elided, and the Newar bhou a to be the labial prefix in a broad form (comp. Maldivian bu-lau). The aie na—ni Tiger also separates-the li- quid from the prevalent Bhoto-Dravirian Jabial, and supports the deriva- tion of all the Dravirian terms from Tibetan sources, It is connected: with the Abor-Mishmi form da-ri,—ta, da, nu; sa, za, ja, cha &c. being mietely variations of the same Tibetan pretix, The liquid must-have beenjcarried across the Himalayas before it was coucreted with the other elements, apd f ETHNOLOGY OF THB INDO-PACIFIC ISLANDS &9 its diffusion isso wide that it must have taken place at.an exnedinsl-teg mote period. ‘The forms in which it is immediately preceded by the labial are probably West ‘Tibetan or Bhotiun, although at the time when they were transported from Tibet there were probably several Bhotian dialects, The Abor—Mishmi forms are probably East Tibetan, it’ vi be the-root, day ka &c. being an E. Tibetan prefix. The Gyarung ta-rha would be a si- milar form, and ¢e-ri or ta-li, da-ri, na-ri kc, may have existed in other _E. Tibetan dialects. _ But another view may be taken.of the Mishmi-Abor terms. In the nore mal animal nomenclature of the formation the liquid, as we have seen, was masc., and the labial in the forms ma, mi &c. fem, Afa-ja, ma-daa may have been currentas a fem. term, the root being ja, dsa&e. The superadded mase. postfix would make the term muasc. ma-ja-rt. So from ja-rt, the masc, form, the fem. may have been obtained by the prefixing of ma-, or ma- when prefixed may have come to be a meredetfinitive, That ja-ri was the proper masc. form and had its counterpart in the fem. ja-mi, is established by the adjacent Mijhu Mishmi retaining that form as its generic name, im ike manner as in sp. Bhotian the fem. si-me has superseded all other names. Wemay conclude therefore thatja, dsa, is the root and merely a variation of the Tibetan chi, chu, cheu, (cha, tsah, sha, ja &c. in other applications ; for ciger su, tsa, ja&e. areused). Inmen-ku-ri u ‘may/also be a variation of the same root (chu, tu &e,), (See Dog). The same combination with the labial definitive preposed (as in byi-la, pi-li)—tound in Manyak only am the known Tibetan dialects, (ma cheu)}—oceurs in the South in Khari Naga mo-chi (Thochu /o-chi), Bodo mou-ji, Joboka me-sa, Kumi min-vho, Karen tha-min-yo. The Dophila- Aka a-cthe, @-sa, is a contraction of an analogous form. The Joboka and Dophia root vowel in sa corresponds with that of the se Mishmi ja. The Bhotian form si is preserved in Lungke si-yo, the final also occurring in the Karen tha-mi-yo, tha-min-yo, and being probably a soft form of cho, jo, as it does not appear asa servile in other names of animals, The Horpa chu-/é and Thochu Jo-chi are not found in the south. ' ‘The Horpa form chu suggests that the Magar sv-thu, (sw-tum der Abor), ©. Tangkhul ta-mi, Maring tung, Manipuri haw-dong, Khoibu tong-kan, Maram tok—pu,.contain the same root in a dental form (as in the Deoris Chutia mi-di tor mi-li, pi-li &c.), It undergoes a similar range of varia- tion in some of its other applications. _ The Tiberkhad and Milchanang pi-shi is explainable as a slender form of the Manyak—Naga combinatien, similar to the sp. Bhotian si-mi. Gerard ives both pi-la and pu-shi as Bhotian forms, and if pu-shi were genuine Bhotian it would be hard to resist the conclusion that pi-shi is also Bho- tian, however much it would pernlex the enquiry as to the directions in which this combination had ropagated. The Dravirian pu-su, pu-cheha, pu-chche, Kol pu-si, and a eh to-pi-sa * are examples of the same Vocable, and it has also been carried to Asonesia, pu-so, bu-si, pi-tsa. dis Lie form pu-shi does not occur in any other vocabulary of Bhotian, it may bé safely considered as an exotic from Tiberkhad if it is really used * Inthe App. to chap, v. the Pashtu pishi, pishik has been displaced and eutered as Kapwi, and the Kapwi ¢o-pisaas African, Pisa may how- ever be yi-sa, comp. the Joboka me-sa, The Rotuma pi-tsa is the same. wWniety. 9° ernxoLoey OF tHe txno-PaorFic LANDS. 7 by the true Bhotians of upper Kinawar. Gerard states thatvin the North- * West of Ladak Bhotian becomes intermixed with Turkish, and if pi-shi be current in Ladak it is probably of Turkish derivation. The true general course of its urchaic diffusion appears to be clear. It is a primary Sey-+ thic term cognate probably with the Tibetan, butdistinguished from the current ‘Tibetan by the sibilant invariably following the labial, and in its most common form taking a guttural final. Mongol has mi-choi and the probably contracted mii. The Tungusian terms are not given by Kla- proth. Ugrian has mi-sak, mi-shok, ma-tska, me-tschik ; and Turkish qu-shak, me-shuk, mi-shik, pi-shik, ata-p-shik, ata—ma-chi, ata-p-si &e. (mouse shik-an &c.), With these Ugro-Turkish forms are connected, on one side, the Caucasian pi-shik (Chari) and Semito-African bi-s, fi-so-na, mu-si, mu-sa kv. (the Turkish ata is also African), and, on the other side, the Pashtu sbgey pi-shi, Sindhi pu-si, and Tiberkhad-Dravirian pi-shi, u-si, pu-sei &c, ‘ The Chinese miau, mau, biu &e. is found in Anam, Lau, and Kasia in the original form miau, and the Mon—Anam stream has carried it to Tengsa meyau, Songpu and Koreng myau-na, Kumi miyaung and Garo myou. The Limbu and Kiranti myong, Namsang miang, Muthun miah are probably also Chinese through Mon-Anam. The Kambojan china may be the same root. _ The Chinese wiau, (Hok-kien, Hai-lam), ngio (Teo-chu), is found in Singpho ngyau, Jili te-ngau, and Chass phiwag he-ngau-bi. The Toung-thu ngwai-pai, and Mon po-kwai, are probably related to it. Oss. The Bhotian root byi, pr with the liquid servile, only occurs in o few of the Manipuri-Yuma dialects, and the pretixual position of the ser— vile shows that the Ultraindian names are not derivatives trom the later eoncreted Bhotiun and Lhopa byi-la, pi-li, but were received when the root was separate. This is made still more manifest by the prevalence of the labial root in the Yuma dialects, either separate, with a def. prefix, or followed by a distinct root. Ifthe prevalent archaic Indian name be of Bhotian origin, it must be very ancient and derived from a glossarial cur- reut distinct trom those that carried Bhotian words into Ultraindia., It was probably prececed in the Dravirian family by the Scythic pi-shi &e; which is found in Litcaindia aud Asonesia, while uo examples of pi-ti, Lis li are found out of India. The sibilant is not found in the Horpa and Thochu mase, forms, but the fem. form current in Manyak is common. The form of the root is not Manyak, whence it may be inferred that the connection belongs to the eva when similar fem, forms were current in the Sifan languages, or Tibetan erally. The form sa, ja, da Mishmi Dophla, Abor, Joboka—is not ound in Tibet. The Horpa chu appears to be connected with the Magar thu; Kumi cho, Karen yo. ‘The sande Bhoto-Thochu si, chi, is Bodo ji, Mulung chi, Lungke si, These various forms and their distribution at- test an ancient aud general transfer and diffusion of the Tibetan names to the southward. The Burman k-roung is evidently one of the latest Sifan acquisitions, and belonys tthe modern Geyarung—Munyak current, The Chinese nanies, which do not occur in Tibet, appear to have earl spredd into the Ultraindo-Viangetic province. As they are best preserv in Mon-Aram voc#bularies, itis probable that they were received by the other dialects trom them. ‘I'he Lumi, Koreng, Songpu, Tengsa, Kiranti and RTUNOLOGY OF THE INDO-PACTFIC SLA NDS... OL», Limba names are all connected, mi arethe Chinese form with anasalfinal, , , ) : | . oa, 1, (a.) The Bhotian khyi, Gyarung and Takpa khi, has the same form in the South Bhotian dialects. of Serpaand Lhopa e-khi, Abor t-ki, e-ki, ~ Dhimal and Limbu khi-a, Newar khi-cha, Kambojan chi-ke. iy . (0) The most common, Ultraindian term is, in its full forms, khwi, khwe, kwiorkul, The Thochu khwa is age broad jorm and the adjacent Sokpa nho-kliwe has the same form with the slender yowel of Burman, khwe. This identity between the normal Ultraindian form and the Mun- golian, shows. that the former was not derived from Chinese (kiuen, hun, times from Scytho-Tibetan. That khwe is a distinct root in nho-khwe and the other Mongolian forms, no-koi, no-gul, no-choi, is clear from kui, by itself, being applied to, the shecp in Mongolian, koi, as in Yeniseian, koi, kay ; to the Duy in Korea, kai; in the Mongolian form to the Dog in Fin with a postfix or second root kai-ra, koi-re, koi-r; and, lastly, tu the Cat in Mongolian mi-choi, Korean koi, kui-ni, Japan mne-ko. | In its applications both to the Dog and Sheep, the guttural root has frequent~ ly atinal nor second nasal root, in the Scythic vocabularies. Thus for the Sheep Mongolian has cho-nin, cho-in, ko-ni, go-ni; and for the Dog Tungusian has nina-kin, nena—ki &e., and Samoiede wene-ku, ka-uany, ka-nak &c. The nasal is shown to be a distiuct root by the Fin fod nak, Japan inu, in, Aino inn, wa So ian nyin, | ‘The Thochu and Burman kl! wil, khwe, khwi, kwi, being thus undoubt- edly Scythic in their affinities, it is possible that the Bhotian khyi, khi, is a soitening of khwi, and not a derivative of the Chinese kiuen, That the Sifan-Ultraindian form is not a modern derivation from the Sokpa nho- khwe is evident from its wide diffusion in the Gangetic, Ultraindian and In- donesian provinces, aud from theSukpa distinctive root nho being absent in Thochu and in the southern vocabularies. it must belong to the earlier ages of Scytho-Tibetan connection, The Gangeto-Lltraindian forms areas follows, Anam khuyen, Mijhu Mishmi kwe, Taying M.n-koe,xo-kwe *, Murmi na+ nzi, Gurung vo-pyu, Changlo khu, Chepang kui, Newar khi cha, Ti, berkhad khui, kao, Milchauang kwi, kui, Garo kai, Mulung and Tablang kui, Singfu gui, kwi, Jili, Mru ta-kwi, Rakhoing khwi, Burman khwe, Karen thwi, tai, Toung-thu thwe, //-twi, Luhuppa thu, Sak ku, Manipuri hwi. The contracted forms aré hu, su; zu, z, hi, shi, si, wi ui, u,#i. Namsang Naga hu, Muthun, Joboka, Mikir hi, Sorgpu shi, es eth ae (u-Sl, ah ea tu-su, [ AAG Mee . Tangk uj, Sogaung az [=tu-zu), Tengea a=ti Khyevg, Kumi, , Kyfne Reyes, C. eae SY Khvibu, Maring wi or ui, RY ak bui, § Tangkhul, Shindu u, Ahari and Silovg ai. The Horpa kata and Manyak /-sha inay be the Tibetan prototypes of the Lepcha ku-shu, ka-zeu, Limbu and hiranti o-chu, Newar khi cha, Magar chhyu, Sunwar ku~chuny, Lodo choi-ma, chi-ma, sei-mu, Garo “® Mr. Brown’s form of the Taying-Mishmi word, neko, led me to believe that itand the Murmi nangi, nai, Guiung nagyu, were distinct from the Tibeto-Ultraindian root kli, kwi &e. aad ailied to the Draviro- Australian naya, nayi, nugi, nage, alay &c. in which the root is na, la &e, lt ia now ciear from Mr, Robiusou's form, nkoe-nokwe (Mijhu kwe), thatthe laying root is koe, kwe and. n-, no- the Mishmit nasal prefix. -, The-renwrks on the athnities of the Draviro-Austialian names of the Lug , (ch.v. sec, 11) must be so far mouified, , C76. ale “eiiwoiod® or: Hie =e WA. A yy » nis bg ih birtedt mung tohai”, iden} Be endl Chdogr ae bbatit is 2 qymore, os See ee rar ‘wansasipag: vob -kejy By ihirtied in i Ay Url h ai dite hat y Ch pu fj ite first: ce +4ltide me : nam n, mA, Ch rena he a me eeteelearatteae ‘30 the ingot he affinities i Sethi Fa pero Samoie » puny, pi a i aie ; ne rabies ee mea the ae Y 11 r names; Wo, animals. aM f rm Ks ie ‘the O ‘Ca i an eth bojan, chi-ma. 3. The Mon ha-la, < is ‘the liquid, used also for the Dior in Mon syn sameof he nga vocals. It) appears: te ‘be the conimon Pe Brooks viii wey enyatonnaft bepress aah ty eastids Nee r¥ ed e Chinese kn fis Deen Satrodueed {ato Anita ont.» a aa 10° pa! © » hon Tren i vant oe } qs : fh t" otk ous ; ble it ‘an be‘infe ones) ‘a ears + (Aueun { ] a 4 J i} I ) , ¥ 1 an ies : ae : + Pogel oe ry x : n LN: "Taagekhial “if ‘not Mon-Anami] 4 iy nally ‘vonneeted Th ; da a-t) TO a forms, § be iB I ea a8 Wir on vi Saher an vas oti serie : ith ia the e distri tt ion AE ; ne Kae poate denieo ation of the sautkary) originate *Y Sw S00" Part 1, ch. hates cis : upabe dado and Gare ng | mais Mr ilees dimers re yl 3 eet a mixed: OD, oallegeinte Dice ases Hows votre taste ba ae OE Fete an ETUNOLVOGY CF TIME INDO-PACITIC ISLANDS. 98 _ The ¢cuttural prefix of the Nipal varieties and the nasal tinal of Sunwar show that they belong to the earlier forms of the Gyarung—Mishmi— Yura band, represented by the Tiberkhad khui, Milch. kui and not to the emasculated Karen and Burman, The k prefix is still current in Mishmi Kumi, Mon, Toung-thu and to some extent in Karen, In the Manipurt and Naga dialeets the dental and palatal forms are more common, but ka a ae in several. (Champhung, Luhuppa, N. and C. Tangkhul, oreng &c,). - The Anam, Ka, Chong and Binua cho belongs to the same era, As the Mon fa-la is exceptional, its proper application being to the Tiger, it is probable that it possessed a similar name for the Dog at one time, 3, Toa, 1. The full Bhotian form phag, phak, is found in the southern Bhotian vocabularies, phak Serpa, phag-po Lhopa, in Limba and Kiranti phag, iu Chepang pak, Changlo phak-ya (Lhopu), Mikir phak. The Horpa and Manyak forms in v, w—ILorpa vah, Manyak wah, of which the older forms must have been vak, wak,—indicate that the most: common Ultraindian forms were of Sifan derivation, and as some of them have ta—, ka-, it is probable that Gyarung had fa-wak, ta-wak or ta-vak before its proper Tibetan vocable was displaced by ki. It had not received ki when it spread to Ultraindia, for that form of the Chinese root is not found there. The Magar wak, Jili ta-wak, Singpho and Mrung wa, Rakhoing wat, Burman and Kyau wet, Khyeng wut, weuk, Khyau vauk, Kumi and Khari Naga ank, Kumi au, Kami 0, (iaro, Namsang, Mauthun, Joboka and Sak vak, Mru ¢a-puk (? ¢a-vak), Nogaung, Mulane, Tablung and Tengsa ak, |? Songpu gh-ak}, Khari auk, Koreng ka-vak, Champhung a-yak, Garo, Maram, Mariug and Lungkhe wok, Kapwi bok, Luhuppa, Khoibu, N. and C. Tangkhul hok, 8. Tangkhul and Manipurt ok, Angami and M. A. tha-vo, the-vo, Shindu vo, Sunwar po, Lepcha mon, Lau mu, Bodo yo-ma, o-ma, Dhimal pa-ya (comp. on-hya horse, nho-ye monkey, pun-ha snake, hai-ya fish, khi-a dog, ji-ha bird, nar-ia elephant &e.), Anam often changes the labial into’ the aspirate, ita heo is probably from beo or weo (weuk Khyeng). 2, Mon ka-let, ka-leik, &-lueit, 4-lut, Taying Mishwi ba-li, Mijhu Mishmi leh, Gurung fi-li, Daphla a-rak softened in Abor to exe, e-ek, yuek [=Mon lueit], Kambojan eh-rok, ehe-ruk, Chong cha-ruk, Ka chu-r, Anam ¢-ru, lon. This application of the liquid root to the Hog is not ‘Tibetan, and the distribution of the names shows that they belong to the peculiar Gangeto-Mon current. The broad forms ruk, luk, rok, rak, appear to be the originals, let, li, yeg, being churacteristic of the later emasculated Gyarung-Mishmi phonology. They are connected with the Manipurian names tor the Blephant, lok, Joak, Tus 4 Saeei A simi- lar archaic form is. appliedsto the Monkey in some Muanipurian dialects and Silong, naky Jak, lait, rhu [=rhu Cat Gyaruny), The rvot is not 2 native Chinese or Mon-Anam one for the Hoy. 1 is evidently of seconda- ry origin, ancient as the form is. It is probably a contraction of one of the -names for the Boar, ha-luk &e. . The amplitied vocalie forms applied to the Buffuloe, Elevyhant and Mou- key in some vocabularies, lui, rui, ruai kc. appear to be contractions of forme like lueit, luak &e, . . 3. The Chinese root. chu, chi, tu, ti, has been received by Karen tho, Toung-thu thau, Murmi dhwa, thua, and Deoria Chutia chu. The Anam. of. efiivofoot Of tie MhoLbicterd WieeHe hed like the Gyarung ki, may possibly be a Variation of the Clinése “chiv’ Anam has also heo kui. hengl — "il 4. The Kasia sniang, (prob. sni-yarig, comp. shia-reh), Nicobar haon, - KOwh, appears to bé connected with the nasal name for the Horse &c., but it must left undetermined. a, Ly _ 5, The Aka kuk-pa, is Hindi khuk. | sulin) ‘Ons. The Ultraindo-Gangetic names are nearly all Tibetan: The Bhotian forms have asmall range. The Ultramdian names appear to’ be’ Sifan, and mostly archaic, thatis they were received before the loss of the® ttural final, The Mon-Anam names are Tibetan and sécoudary, one’ Gaines from a soft Angami form of a Manipuri variety of the rout, and the’ other being the Tibetan liquid masculine qualitive in an archai¢ form. ' Anam may possibly preserve a native root, but it is probably Chinese.” Karen liaa received the Chinese name and commtmicated it to Deoria® Chutia and Murmi. This is one smong many glossaria} evidences of ite’ pre-Burman iifluence, diffusion and Chinese relationship. 4. Goat. .1. II.) The, Bhotion ra is current in eps Lhepa, Murmi, Gurutig Magar ue Changlo, ‘'Tiberkhad has la for tl e female. Allied forms ae found in Garo py-run, Muthun ron, Joboka roan, ¢whence the Mulung yon, Tablung yun), te k-long, Kasia d-lang, Singphu pai nam, Anam hoi nam (hoi=2boi, pai), Lungke, Kumi ke-l, Kyau ke-vat, kie-ar, Mra “td-Ynu-a, Limba men-tla, Garo do-dak (Brown), Anam de. The liqnid 9 clearly the mase. root. 1t is current in the same forms as a sex word, and in the names of other animals, 2. (ITI.) The Bhbtian chang, Horpa’chhe, Thochu and Manyak tsah, Gyarung ku-so, are found in Anum su hoi, (= su bv1), Abor shu-ben, so+ fen, Aka sha-bum, Kumi su-b¢, Kathi tao-De, Lepcha sa-ar, Kirauiti guneng. vr, (Bhot. ae ts Newar cho=le, Sunwar cha-r-sye [See Cony hepang m-cha, 'l', Mishmi ma-dze (Brown), Songpu zya, Burman’ she-ik, tshi-et, chii-t. Fas pat ~The Anam, Abor, Newar, Kumi, Kami and Songpa forms in uy op resemble the Gyarung so. The Aka, Chepang, Lepcha and Sunwar_ sha, sa; cha adhere to the Bhotian, Thochu and Manyuk vowel. The Kirantiy laying Mishmi and Burman have the e of Hor But these va- jations are too slight to warrant any inferences, save that u, oy is proba 3 “ -* = 4 y the older ferm., | °° 3. °(1.) ‘The labial root is very prevalent asa name for the Goat, althotgh it may have’ originally been a Contraction of Tibetan forms in ‘Which the’séx detinitive was conjomed with the root, asin the Bhotian ya-ba, r-nia; Changlo ra-ba. The Abor, shu-ben and the cognate terme ‘nay be the sibilant Tibetan root with a similar posttix, Mijhu Mishmi ken-pai, Singfu jui-vam, Taying M. nea-biey Bode bur-ma, bor-ma, Tengsa and Nogaung xa-bung, Khari na-bongy, Mrung pun, Manipury Wu-weng, Koreng ka-mi, Maram kha-mi, Lalwppa me, Chimpheung a-mu, N-Langhul mi, C. I’. mi-k-re, 8. T. minsk-re, (ke=re male postf.), Kumy, “ine, mee, Khyeng, me, ma, Pwo Karen bhe, Mikir be, bi. In some of these terms the labial retains a qualitive power.” Sak ii-bi, meats ep kha-pa, ‘Toung-thu bay, ta-byu-pai. (Cat ngwai-pat), Bonyjuwoi, Kuki: BEWBQWOPT, QP, THE, PEPOTRACIEIO. ISR AN DRL oes hui, Anam swhoi, hoi nam, * Lau-pi, pe, Ahom pe-nga, Deoria Chutia li-pe—du-ru, Kambojan p ites Silong pet. T°"! Cs BE TYO aot seifostf The labial is ftlot uséd in Tibet is a name for thé Geet, the only truel Tibetan fadical minté being the aibilint (2). These labial names are obviously of secondary soitherm origin. The various forms are andi : the differétit ¢irrént mModifitations of the labial -masculine root, sulk fee, of them are also used as names, or elements of names, for the Gut, Comy Buffalo’, Blephant, Horse sid Morley. Many of the dialects ‘Wa tse the labial for the Goat retain the proper sibilint toot of Tibet for the Cum; and in forms and combinations identical with those current in othort southern. dialects, or in Tibetan, for the Goat. This Tenésa; Nogaunge? atid. Khari have Ae-bune, nd—bong for the Goat, Wut me-si, nisi, ma-suy for the Cow. The samme remark ds to the liqitid qualitive: Thasi while Kyau find Kumi havé otily Rear, he-vat, ke-l for the Goat, they retain the Tibeto-Gatigetic forms cha-ru, si-ra, téi-ya, shar), as names Me. ie. Chinesé "$e aan seus wocyeing 6. The Chinesé yeong, yong, yang, yu of shan yeuris, tso yeungy sian yang, fsau yang &e. (sheep tin veang &c.) appears to ba: fut dom Anim du-ong, dat (Brown). The common mative term is de: 4, Jili ta-khven, Natndang kien, Kapwi ken. & Khoibu hing-ngan. “i ) Aare 6, Angami fenu, My Ai ta-nu. This appears to be the fem. root used substantively. The Burtman nua, nwa Cow (Chinese) has some resem- blanté toit. ‘ ; i wie Obs. The Ultraindo-Gangetic names that are similar, to the Tibotat do not appear to hive been derived from any single Tibetan dialect. The reptodee al the Tibetan forms, and must be considered very, archaic; The Bhotian secondary form ra is only found in the Himalayan vocabus « laries, in which it is pega nly modern. The Kiranti chheng-ar_ is full Bhotian name slightly modified, the yowel being’ similar to th Horpachhe. The prevalent sibilant in the south is probably Sifan. The paucity of riames containing this root is remarkable when its pefsistenee in all thie Tibetan dialects is considered . From the peat eae fang of the labial names, the peculiarities of some, and the seryiles annexed several, it appears that the labial became current as a substantive naine tor the Goat aba very early period, and in dialects that ei ses al fluence, The sibilant root, in the names in which it survives, has the las bial ase. postf, in the forms wi, ma, be, ben, tim. With mi Chep md be Kumi, Kami, ben Abor, the moar Mafiporian mi, me, meng, uma me, bi, Taying Mishmi bie, Mikir be, bi, Keren blie correspond, sd that-ali may have been derived from dne East Gangetic dimlect. The Nat fa be iz, bong’, is probably da ee to the Aka bam, atid it appears to havé e ed i He oi n the original of the Bodo bur, bor, the older form of which,is preséry- ed in Mruig pun. The Mijhu and Singpho, Mon, Tangthn a df Yun pai, is a distinct form, perhaps derived from Tibet by the Iriwadi route, lik other peculiar Sifan—rawadi vooables. - It occurs in. Thochu in the fori -wai( Moukey mai-si), and in all the forms in other Ultroindian names, ; . The Lau and Kambojan names appear to have been derived from tlie Kavens¥uma me, be, pe See PTR alain 7 . = So-the Arung Naga name of the Mi-thun, bui saug (buffalo ga- bui) is hui in Angami, fy) x : Lepit4 @eAf — 50 ETHNOLOGY OF THE INDO—PACIFIC’ IsLaNps: 5. Cow. I. (a.) The Bhotian ba, lang, are found in Lhopa bha, cow, lang, dang bull, Tiberkhad ba-lang, rad, Milchanang lang, Lepcha long, Changlo wa. In the slender form itis found in Sgau Karen go pi, Dhimal bi-a, Sunwar bi, Lepcha bik, Limbu bit, ye-pi, Kiranti pit, Murmi mhe, mih, Gurung myau, Bengali pa-bhi. In some of these vocabularies os in Bhotian the labial has become a sub- atantive term, or it was originally received into them as such. In most of the Ultraindian dialects it is conjoined with proper substantive names, and has either a sexual or a definitive force. As our information respect- ing these vocabularies is too scant to enable us to distinguish those cases in which its sexual meaning is still recognized, from those in which it has become a mere definitive, and as it is current with both functions in Bho- tian and several other languages of the family, I have in all cases itali- cised it, in order to give greater prominence to the substantive names, The forms ma, mi, mu, appear to be always feminine. Ba, pha, bo, bu are masculine. But as the two forms of the labial are easily interchangeable, masc, forms such as pha become fem. in some dialects, and fem. become masc. Boi, woi, bi, wi, wa, would be masc. if the Bhotian masc, force of b, p, w, were preserved, but in some cases they appear to be fem. The i may have a fem, power in some Arianised siding sooty asin Kasia, (b.) The Manyak form we-mi may possibly be directly connected with the Shan wo, woa, Anam bo, Toung-thu po, Mon e-won-ban, Sgau Karen a-mo. A similar form is applied to the Goat mu, woi, Buffaloe moh Elephant mon, vu, mu &e., Hog vo, po, mon, mu. Bree Bap __ (¢.) A consonantal guttural form occurs in Maring muk, Manipuri and Tangkhul sa-muk, Champhung she-muk, Luhuppa st-muk, Sak ¢ uk, Khoibu na-muk. In these forms the sibilant is the def. pref. Comp, in Manipuri, sa-muk cow, sa-mu elvphant, sa-gol horse, ha-mung goat. The same form is applied to the Horse in Maring, Khoibu, 8. Tang- khul and Luneke puk, phuk, and Namsang mek, and to the Llephant in Namsane puok, Singpho mag-i (Manipuri sa-mu, Champa ta-mun). (d.) Namsang man. The same form is used for the Morse in Muthun man, aud Joboka mang. Il. (a.) The Bhotian chuk of pha-chuk is found in Serpa chu-ma, Taying Mishmi men-chu, ma-chu, me-tea-kru, (masc.) Abor sou, Pwo Karen tshu-men, Bodo ma-shu-jo, Garo ma-shu, Mrung ma-chou, Khari wa-3u, Angami and M. A. mi-thu, Tablung and Muthun ma-hu, Joboka mu-hu, Mulung ma-hu-nyu(fem.) Ahom hu, Singpho kan-su, Anam sung- krau, * Lungke tcho. b. The a and i forms of Thochu zya (du//) and Manyak nga-zi ult) are found in Aka shye, Abor sha-me, og eed rate Newar sa, hanylo ja-ba, Nogaung na-si, Tengsa and Kasia ma-si, eng shya Kumi tsi, Kyau cha-ra, Kumi si-ra, Mru tsi-ya, Khyeng sha-r/. The forms with the ep masc. postfix are similar to some of those for the Guat, chang-ra, Bhot., chheng-ar Kiranti, sa-er Lepeha, cha-r- sye Suuwar, cho-le Newar. The same roots ap to be contained in Arian names forthe Bull bri-sha, shanr, Goat chha-gal, chhag, aja. * Brown's dialect. MHAIMATEY STIIG4T-OU04I FHT To YHOLOuNTS ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, Hod fo 3. (ILL) Songpa wot-tom, K wi toa Fp ne m4- : ieee teat ebay med iets bin eee. sod dane ites eure’ for migle)) "Anat ’ hetre, "Ket si t ti) . ey (a Mic Mild lohoiratg ater: the eye has Seay cies TH Phe: pen ww noov Seuilt fo Snide fr i ah ee ne i Betas alec sega Esiourtanertt ued is ledesplois)) 8 ve At wine ae BALMER -SeyeoThe Ohinese root i tho Tok | ay ngaszi pis ddurid in severabof the woifther aes er ee HE a-dirertiShinose Seopsrenehin) Dine att Hab at ae he i vat tunteuted iy to! Pa i 7 pao aoe tener fori beitig Sun caren ‘bel Ti e..1 (era iio yguy non, ati ee -tiea, Bukinith ua, an 8 patie Bisa Kaken #-mn0/ Con net y nWhe DO Dull Ru aude) oct To eavini owl 90 ras ti amitiosenr Mabe ionD inecactmioaltes nalitive ku) kin} fe peti an reer wlio, » The. aeheear ib and’ ath a tai! ol ot Ts qs yqat ar, Se OF fad | thanks se Py sonia or Se. Dhevouly:Sifensforms of the: pair dtys athe y Hes ei ent shdivelomnionn Gly aftgiliny- syerfounit int sai fer “las if ti, arcartA tt oe (OW ila t). itnd-0'R =") 110 nye ou cal amen The Bhotian ba and lang have t mg limited ed gna m foe areonly fous in! the) sate Rot dates and! ve the conterminous ones 700! Fy oF & 4M OY A Jun formris'the Manyake fio tavTibetaww nin ma canal ‘ind “mecha Soe eS ale ats elie: its ‘opr ae nen ‘ lect. But as the mase, qualitive has the same forms if fo the: Bul and inohaotihe old Tibetan | rosa bao tien oe tion; thisia:doubtful.. _ the Briririn-awaenh Obie? nivalpmny, ne Siamese nguw or po key De inven terine robabl¥ ardentraction of a ‘terra’ ed!-the Burmie Forge morte te tlie Keven name: tor Mother and: ae orp most probable tira sire cit ans the tin gia a compat) rivative from DheeMani connie mere slot pion i ot Als Ses yes Gout) ‘Dhe Seino Sa Tai hae ad termenaoal i Sie ath ree oe it i F og a am Ser Chsrenionet Bd them of at least in their modern raga YD the east Galivétie tort pelt, ce Abort) dom) thi cabularies. The Singpho han-suy. distinict’ ihe fosn ant witha eis cing ma=, aso favours a repeal an gin. Lastly Sea cimmibailas tothe Thetis xin and Mumyake 2¥ate-found, 98 ETHNOLOGY OF THE INDO-PAOCIFIC ISLANDS. along with chu, tso in the east Gangetic group, shya, sha, ja, shye, —in Rack si,—and in the Yuma gr. shya, tia cha, ze ee ‘Tt is miobuble that the Abor-Naga broad form was received into the east Gangetic vocabu- lary from one of the earlier Sifan streams, and the Abor-Yuma from one of the later, after the slender phonology prevailed in east Tibet. It is clear that the sibilants used for the Goat and Cow in the Himalsiec glossary are forma of one root. The same variations are applied to both animals, shu, (with the variations zyu, so, tso, cho for Goat, and su, thu, hu, tso, teho, sou, chuk, chu, for Cony), cha, sha, sa, (also for Cow ja, sha, shya, zhya, and for Goat chang, tsah); for Goat the slender tshi, chhi, chhe, dse, she, chheng and for Cow zi, si, shye. Thetwo animols were therefore referred to the same species in the primary Himalaiec zoology, and they were probably distinguished either y the forms of the sex qualitives or hy separate attributives, descriptive of size, colour or other distinctive qualities, as in several of the Chinese names of animals. In the gradual concretion of the glossary in each dialect, each variety of the root would become a substantive name, ren- dering qualitives and definitives superfluous ; and in like manner, varieties in the qualitives and definitives, on acquiring an independent substantive meaning, would render the older suhstantive roots in the compound re- dundant and sometimes cast them off. For example in Bhotian the ra form of the mase. qualitive, may have become distinctive of the Goat, and the lang form of the Cow; and when the sibilant substantive name itself took the two independent forms chang Goat and chuk Com, distinctions in the qualitives were no ene necessary. In the latest stage of concre- tion and metamorphosis all these forms, chang, chuk, lang, ra, have ac- quired distinet substantive applications. The Manipurean tom appears to be a local modification of the sibilo- palatal root, from the form thu, Angami tcho (Lungke) &c. found in ad- Jucent dialects. The same variety is used for the Llephant in Mishmi da-ton. The Ultraindo-Gangetic names forthe Cow and Bull of Chinese de- tivation appear to have been first acquired by the Lau tribe, and to have been communicated by them to a few of the other vocabularies, when they spread to the west and south, 6. BurFraLor. 1. (IIi.) The Bhotian ma-hi, ma-he Lhopa, Lepcha, Murmi ma-hi, Sane me-shi, Sunwar me-sye, Newar, Deoria Chutia me, Gurung ma-i, Ma- gar bhain-sa, Taying Mishmi mo-ji, Mrung ma-shi, N. and 5, Tangkhul shi, appear to be Arian, Sanskrit ma-hish. But although the Buffaloe with its name appears to have been carried from India to Tibet, the name is probably pre-Arian and Himalaic in India. If the Arians found the Buffaloe there, they would be likely to adopt the native name. That ma- hish, ma-shi, &ec. is Himalaic can hardly be doubted when it is compared with the corresponding Himalaie names for the Cow, ma-shya, ma-shu, mo-hu, ma-si Ke, ) 2. (IT.) (v.) M. Mishmi ¢e-loi, Jili, Champhung xge-lui, Mikir che-long, je-lang, Muthun loi, Joboka lue, Maring lui, Manipuri 7-roi, Songpu woi-voi, Kapwi sa-loi, Luhuppa. si-loi, 8. Tangkhul se-lui, [se-loi is ap- plied to the Asl Gyal in Chittagong], Koreng a-lui, Maram a-ghoi, Angami and M. A. ra-li, Khoibu ra-loi, Kyau cha-la-we, Bongju f-se-loi. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 99 This is the common liquid mase, root. The form loi probably spread southward from one Himalayan dialect (Mishmi or Singpho) to the Manipuri-Yuma vocabularies, in which it prevails, or froma southern dialect northward, As similar amplified forms are found in the Yuma up and Mon applied to the Mog leuk Xe. and Monkey h'lait (Kumi) it may have spread from this group to Manipuri and thence to the Lra- wadi and northward. The thet element in the name of the Hurse whatever its etymology may be) takes the same form in Kasia, kalai and Jodo korai, gorai. e Kambojan name for the Zlephant has a similar form tam-rai, Chong ka-nai, but this appears to be aderivative from the Trawadi ta-loi Buffatoe. The form is probably of western Lrawadi ori- Dn. ob) Anam k-long-nuk, (Mikir che-long) Sak k-ro, Ahom kh-rai, Lau kh-uai, Burman k-ywai, k-wye, k-yue, Kam Bayan k-ra-bo. From the Ahom, Sak and Kambojan forms the original appears to have been a form of the mase. qualitive similar to (a), rai &e. with the guttural retix in place of the ng or t of the north lrawadi v ili, Mishmi). The bojan combination is similar to the Kyau cha-la-we. La Sec. 11 of ch. V. Lconsidered the Kambojun name to be Dravirian. If, as I now think, it is Himalaic, the question arises whether the Dravirian karan, karayai is not itself Himalaic. (c.) Namsang le, Kasia shin-reh, Mon pa-ren, p-riang, p-yen. Similar slender forms occur for the Goat in Karen and Newar le, C. and 8. Tan khul Anam de; for the Hlephant in Kambojan re; for the Hog in Mijhy Misha leh, and Mon let; and for the Morse in Abor, Burman re, The modification belongs to the later Sitan—Irawadi current. The Manyak ding-mi is perhaps connected with these forms. 3, (III.) (a). ‘Aka, Dophla men-dak, Abor men-zek, men-jeg, Mulung, Tablung tek. Possibly this is an archaic form of the liquid mase. root similar torak Hog, rat Goat, lok Elephant, and lak, nak Monkey. But it is closer toan archaic form of the dental and sibilant preserved in names for the Tiger tak, jik. In the Bhotian u form of the same root as ap- plied to the Cow tinal k is preserved, chuk, (2). Tengsa tyang, Nogaung chang, Tengsa chang, This is the form of the sibilant root found in the Bhotian chang Goat, Murmi chyan Tiger, Lepeha tyan Elephant. 4. (I.) (a.) Deoria Chutia me, Assam moh, Garo mat-ma, Khari a- ars Limbu sa-wet (Burm. wet Hoy), Kirunti san-wa. 5. Anam ngiu, Singpho nga, Jilinga-lui, Lungkhe na, Khyeng nau, Ku- mi pai-noh, pa-no, ma-na, Karen, ‘Toung-thu pa-na, pai-nai, With the ex- ception of the Anam ngiu these forms appear to be all variations of the north Lawadi nga, which is identical with the Manvak form for the Cow. Oxzs. No distinctive root for the Byffaloe occurs amongst the various names. They are the samesubstantive and qualitive roots that are used for the Goat, Cow &c. Some well marked groups exist. | The Gangetic form, including the Sanskritand Bengali, is the archaic Himalaic name for the Cow (ma-chu, ma-su, ma-hu, ma-si &c.), If any qualitive original- ly distinguished the Buffaloe from the Cow it has been lost. Another a is the large lrawadi one in which the mase. liquid root has been di in the form loj &e. from some single dialect, > A second and:smaller Ultraindian group presents the same root ina 100 ETHNOLOGY OF THE INDO=PACIFIC ISLANDS. later or slender form. It appears to be an [rawadi form (Mishmi—Mon tiog),and was probably communicated by Mon to Kasia aud by Kasia to Nenang Naya. he Abor group preserves the substantive root in an archaic Tibetan form and the Nogaung and Tengsa forms appear to rank with it. These forms were probably used originally tor the Goat and Cow in the east Gangetic dialects. The labial names, with one exception, are confined to Assam and its bor- ders. They appear to be remnants of the old Gangetic ma-hish, me-shi &e. (Ni i, the Deoria Chutia me being identical with the Newar, The «.imbu sa-wet and Kirantisan-wa appear to be archaic, for they have the qualitive postfixed uud not prefixed as in the prevalent concreted Gungetic word. ELEPHANT. a 1. (IL) (2.) The Bhotian wr. «nd com. Tibetin g-lang-chen is Lhopa Jang-chhen, ‘The sp. lany-bv-chen, Hurpa la-mo-chen is Serpa lang-bo, Murmi lung-bwv-chi, Changlo lang-pe-hi. o.) Anam ti-rang, Kambojan ¢wm-rai, dwa-re, Chong ka-nai, Ka ruai. futhun loak, Joboka luak, Mulung, Tablung lok-niu, Burman a-ne, Rakhoing nin, Kyau ni, Kasia iny-nar, Dhimal nar-t, 2, (i.) Mijhu Mishimi man-yong, T. M. amiang, Garo mong-ma, Sing- pho may-wi, Namsang puok, Manipuri s¢-mu (Cuw sa-muk, horse sargol, goat ha-meng), Songrpu yui-pong (evw wei-tow, buffialoe woi-rhoi), Kas wi ¢a-pong, Koreng eha-pong, Maram m-pong, Luhuppa ma-vu; . Tangkhul ma-phu, Anam voi, woi, Champa ¢a-mun, Khying mwi, These are similar to forms of the masc. labial current, a3 names for the Cow, Goat, Hors &e. (’. Garo na-y-lo, Champhung p-lo-bi. ’ a The slender form is applied in Mon to the Bujffialoe p-ren &e; Comp. also Songpu wei-roi. 3. (Ill) T. Mishmi da-ton (Brown), Lepcha tyan-mo, teng-mu. 4. The Chinese siang, tseung, sioug, ch’hio, sio, tiang is found in the Lau dialects chiang, tsang, sang, Jili tsang, Burman, Mon shen, Mon chuvin, Rakhoing san, Kyau sane-hung, 5. ({f1.) The Karen ka-tsho, An- gami and M. A. tsu, 8. Tangkhul, Maring sai, Khoibu ka-sai, C. ‘Taungkhul sa-ka-tai, Shindu mu-shey; Kumi /v-shai, Lungke tsai, Newar and Chepang fi-si, may either Chinese or Himulaic. The form of the root, the pretixes, and of the range the terms appear to show that some of them are native application of the root for Cow &c. G. ‘The Sanskrit sita, site iscurrent in Abor, Sunwar (soda), Tengsa suti, Nogaung shiti, Khari sati. Both thisterm and gaja (=ya-ja) ap- to be Himalaic. . 7. The Hindi hati is used in Aka, Bodo, Limbu, Kiranti, Murmi, Magar, Gurung, Mikir. 8. Sak u-ka. , Ons. ‘The Tibetan chen is probably of modern Chinese origin. The most common native term is the masc. Himalaic labial, in the archaic u, 0 form (Bhot. po, bo, bu &c.). The —k forms appear to be very ancient, reseinbling those for the Hug. The nasal were probably formed from them. ea “The Anam voi, woi, is the Songpu prefix with the root elided. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS... 101 - The Naga loak, lok, isa remnant of an archaic masc, form. Similar forms suryive in names for the Hog, amd this may indicate a special con- . nection. 8. Honse. 1. (ILI.) (@.) The Bhotian r-ta, ta, Takpa te, is eurrent in Serpa, Lhopa and Murmi ta, tah, tha; and the Tiberkhad shang, shumy, is the same root. Karen ha-the, ka-se, thi, Khyeng tsa, Kyau sha, Kumbojan se. 2. (IL.) The ro, bo-ro, b-rd of the Sifan dialects, rhi, ryi of Horpa, is the most common Ultraindo-Gangetie name,—buv-re Abor, sa-la Newar, se- rang Chepang, rang, rong Milehanang, kam-rang Singpho, m-rang, m-yen Burman, rang Kyau; Mishmi ga-re, g-rue, Abor gu-re, Tengsa ku-ri, Nozaung ko-r, Khari kung-ri, Angami ki-r, M. Angami che- — kwi-r, Manipuri, Champhung sa-go-l, Koreng and Maram cha-kon, Song- pe and Kapwi ta-koun, Kumi kaungo, Luhuppa si-kui, N, and ©, Tang- hul sa-koi, [Rakhoing k-ray, Burman k-re, Mon k-veh, Kasia ka-lai, Bodo ko-rai, go-rai, Kiranti, Magar, Gurung, Sunwar, Bengal, Hindi apse Changlo ko-r-ta, (ta 1), Sindhi go-ri, Kashmiri gu-ri-ri, Tirhai uera. P 3. ah) Maring puk, Khoibu sha-puk, 8. Tangkhul sa-puk, Kumi, Lung- ke ka-phuk, Namsang mok, Muthun man, Fo beck mang, Mijhu Mishmi kom-beng. (See cow, elephant, hog). Tablung and Mulung ko-wai, 4. Dhimal on-hya, Lepcha, Limbu on, (? 2). 5. The Chinese ma is current in the Lau dialects and in Anam. 9.- TicER. © ; 1. (II1.) The Bhotian and Horpa tag, tak, is Lhopa tah, Se jik, M eho asi Tibarkhad tar, Nis ‘ i a tag 2, (ILL) (a.) Abor si-muo, su-myo, T, Mishmi (Cat, si-mi Bhot.). (b.) Bodo mo-cha, mi-sah, ma-sa, Garo ma-tsa, ma-cha, Sunwar u-p-aa, Chepang ja, Karen bo-thao, bo-sa, bo, Jili ka-sa, Namsidng sa, fulung—Tablung sah-nu, Joboka cha-nv, Muthun chia-nw. (c.) Lepeha se-thong, si-tong, Newar dhun, Murmi chung, chyan, Gurung chen. This is anarchaic form for cat Manipuri gr. tong, tung, tu, &c., Magar thu, Horpa chu; and the Maram tok, Anam sok, show it to be a variation o the ee tak, awe - . d.) Anam ho, ong-kop (Brown), Lau su, sua, Maring hum-z, Song- Ry sy 8. Tangkhul ham-pu, Khoibu hom-pwi, ») 2 3. (l V.) (a.) The guttural root (primarly applied to the dog) is very com- mon, Tengsa khu, Khari a-khu, Angainiand M. A., Kapwi ¢a-khu, Koreng cha-kwi, C. Tangkhul sa-kwi, Maram khn-bui, Chemaphage a-khu-it, Lubuppa sang-khu, N. Tangkhul sa-khwa, Shindu chu-kom, . Nogaung kayi, Manipuri kai, Kumi ¢@-ki, tw-kae, ta-gain, Kyau kieh, Lungke tchek-ke (? ehe-ke), Mikir ti-ke, Limbu ke-wa, ke-ba, Kiranti ki-wa, Sunwar gu-p-sa, Anam ong-ko-p (Brown), Ka dea, Silong p-nuk. ; 4. Changlo kai-la, Kambojan k-la, ki-la, Mon k-la, k-ya, Kasia k-la, * Many of these names are evidently Sceytho-Dravirian (Hindi.) and not Scytho-Himalaic, but it is difficult to draw the line. Those with- in brackets appear to be Hindi. See ch. v sec. 11. : 102 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Magar rang-hn, Singphio sa-rong, Lau si-ra, Burman ni-ra, Chong ro-wat. (See Voy). 10. Monkey. 1.. (a.) (IT.) The Bhotian wr, s-p-re-bu, Gyarung she-p-ri, Takpa p-ra, is current in Lhopa p-ya. (b.) Serpa rhu, [ta-rhu Cat Grarung), Chepang yukh, Kumi h’lsit, Bodo wa ee Garo ma-kh-re, Silong k-lak, Muthun mat-nak, Joboka me-nak. ung-thn ¢av-lam, Lau er. ling, Kasia sh-ri, The Newar mako is Avena derived from the Bedo mokhara. 2. (1LL.) The sp. Bhotian tvu, Horpa zum-dé, Thochumaz-si, is found in Mulung and Tublane si-ma?, Tengsa su-chi [L/ephant, su-ti], Changlo ja-la, Kiranti he-la- wa, Nogaung shi-tsu (Llephant shi-ti], Khari ki-sha, Kapwi, Maram ka-zyong, Koreng ta-zyong, Manipur, 3. Tangkhul yong, Maring yung, Luhuppa, N, & C. Tangkhul na-yong, Khoibu ha- yong, Khyeng, Lungke yaune, Burman m-youk, Kyau juang, Kambojan Bun, 3. (I.)} The labial root is found in Mijhu Mishmi muh, Taying Mishmi ta-mium, Gurung ti-myu, Murmi mang, Hindi mai-mum, Drav.-ma -mange, Sunwar mora, Lepcha sa-heu, Anam won (Brown), Abor s2-bie, si-he, Aka le-be, Singpho we, Jili tae-we, Namsang veh, Garo kon-we, ‘Limbu so-ba, eho—ba, Mikir hi-pi. ‘ 5, Angami and Mozome A. te-kwi, Songpu a-koi, Anam khi. The saine forms are used for dog,, goat, tiger, horse. 6. Mon ka-nwe, Dhimal nho-ya. 1. Frise. 1., Nga, nya (Bhot., Takpa) is very common inthe southern vocabu- laries. The Gyarung form neyo oecurs in Abor e-ngo, Tengsa and Noganng a-ngu, Kumi, Khyeng ngo, Kyau ngwau, The Angami kho, Mikirv-k are related to these. The change of ne to k iscommon to Angami with several of the Manipuri dialects kha, khai, cha-kha, a-khai, khi &e. with Anam, Mon and Nicobar ka, and Kasia do-kha (o-k Mikir). The loss of the nasal occurs in Pwo Karen,—Szau nya, Pwo ya. 2. izha Thochu. This Seythie word is only found in one of the pub- blished Gangeto-Irawady vocabularies, but it is Indonesian. The ? she of di-she has the same root. 3. yu Manyak. Chinese yn Kwan-hwa, yue Gyarung. The Gyarung usage of preposing the word for water—chu-ngyo—is an archaic Asiatic one, Seythie, Semito-African &c, It is preserved by Magar di-she (di water). See also Snake. The Murnti ter nya, Gyarung tan-nga appears to have the dental prefix in one ofits Gyarung forms, . The Taying Mishimi tan, ta, may have lost the root, or only preserves itin the n, ta being a common pret. in this vocabularye The Toung-thu , da-dan, Mra dum are evidently related to tan, and suggest its being a form of the la, ran root. Lau reverses the Tibetan application of nga, ngo, ngu and la &e., using the former tor seake and the latter for fish, p-la Ahom, Siam, p-la Khamti, Laos. The Kambojan group has the same root for fale t-rau, t-irai, ¢-rei ant t-re Chong, me-l Chong. Anam follows the Tibetan > > usnve ran, save, (t-ran dye). Thea form of the foot is peculiar to the Moa-Anum group—the Tibeto-Burman having u. 1tis probably related ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 103 toa Manipuri form, ma-run Kapwi, phu-run Khoibu, the normal form being rul. The Anam luon gel is a similar form. The Mru form ta-roa resembles the Kambojan. The 1, r, root appears to be that for river, wa- ter, in archaic forms. See Srake, 12. SNAKE. 1. The broad wr. Bhotian s-d-rul, Manyak }-rui, Serpa d-rul (whence d-eu Bh. sp., b-eu Lhop. —_ is preserved in Takpa ga-rm, Mikir phu-rul, phe-roi, Maring ph-rul, Khoibu phv-run, Kapwi ma-ran, Tengsa pha-la, Luhuppa pha-ru, N. T. ph-rn, C. T. ph-rai, Kuki rol, Lungke rul, rui, Sakhoing w-rwi, Kyau m-rui, mi-yuk, Burm. m-ywe. Anam ran, ¢-ran, (eel lugn). Mon ¢ha-run, Mru te-roa, Singpho la-pu. Champhong ri-nam, Maram sa-na, Koreng Aa-nu, Moz, thi-nhye, Yerukala tu-na. Manipuri lil, Champh. ri-nam, 8. Tangkhul ma-ri; (tu-lil river Maring, ri water, nim water, revi’). 2. The labial is found in Dophla ta-bug, Ako ta-buk, Garo du-pu, Deoria du-bu, Gadaba bu-di-bu, Bodo ji-bou, ju-bu, Dhitnal pu-ada, Taying ¢a-bu, Sak ka-pu, Singpho la-pu, Namsang and Mulung groups pu, Nooauny pu-r, Magar bu-l, Khari ¢-ha, Aneami tho-fa, Kumi pu-wi, Khyens pwa, Mrung fse-bu, Sunwar bu-sa, Muarmi pu-ku-ri, Gurung bhu-gu-r, Abor ta-bi, Garo cha-pi, Newar bi, Kol bing, Kamboj. po-s (Sunwar). S. Drgvirian pa, ba (root). 3. Mijhuzhu. 4. M. Kumi ma-khui, ma-kwi, 8. Karen gu, P. K. w-gu. 5. Limbu o-sek, Kir, pe-cham, Savara ja. 6. (ond ta-ras, Kondh so-raso (prob. 1 with a postf. sa—ra-so, comp, tan Anam, bu-sa Sunwar). 7. Lau nga,ngu, (See gfe Allthe names for snake, with the exception of 2 and 7, appear to be forms of the common roots for river, mater, raul being an archaic form; that is, the root has been lost, and the descriptive or qualifying word only reserved, as in many other current vocables, including names of animals. ‘he Tibetan root was probably the labial, the forms aud “distribution of which show it to be radical, and not merely the Tibetan prefix with the r rootelided. The Murmi puku-ri, Garung bhagu-ri (¢e-bue Dophla), are examples of an archaic form of the root, followed by the form of the liquid root for water, common to Burman, Magar &e. The Singpho la-pu muy be a similar combination. In the Garo du-pu, Gudaba bu-du-bu, Yerakala tu-na, du, tu may be water ae not uae a prefix, 12. Birp. 1. a. The old Bhotian byu is now an exceptional form. It is pre- served in Indonegia, pio Sambawa. The » form is found in Lhopa bhya, Tukpa pydfTaying m-pia, Milchanang pia, pea, piatsh, Murmi ni-inya, Newar w=-mya. The Gyarung pye-pye is the only slender Tibetan form, Comp. weng Kapwi. With the old Bhotian form are connected the Singpho wu, Naga the- vu, vo, 0, Kumi ¢a-wu, Limbu ba, Lepcha pho, Chejaay moa, To the forin in a are related the Mijhu wa, Yuma wa, ka-wa, ta-wa,- ha-va, ta-va, Toung-thu a-wa, Sik wa-si, Chepang, ful, wa. 2. The Thocha mar-wo (wo, from the analogy of other vocables, be- ing probably the def.) has direct Seythic affiuities. It is an archaic é Od ETHNOLOGY OF THE UNDO—PACIFIC ISLANDS. Seythic liquid form of the labial root like the dental form, and like it is also Iranian and Dravirian. The pure root appears to he radically JSeather, wing. Both the liquid (-n,-/, -r) and the dento-guttural series -t, -k, -s &c.) are current in Seythic, Indo-European and Dravirian. eather pal Korea, pil, pul-an Yenis., pil-ge, pil-ga, pyd, pnd, bud-al Ugrian, (pul Tibet, mun Singpho &e.), pli-na Lat., wot Armen., par, al-ah, pad, pakha, pakh-na &c. Sansk., Beny., Hind., puru-kw; bot-c« v. Wing bar, bol, pank, Hind. &@. &c. Bird German, vo-:-cl, Lat. avis, Eng. bir-d, Sansk., Beng., Hind. par-indu, pata-ka pakh-ern, pakh-yi, Drav. par-va, para-vet, pul, paki. pita: The Angami para, pera appears to be Dravirian and not Thochu. There is no other example of the Thochu voenble, and other Dravirian vovables are preserved in the Ultraindian- vocabularies. Allied vocables are current in Malayo-Polynesian—pio Sambawa (byu old Bhotian, pia Tnkpa &c.), bau Kissa, pao Mille, Batan fow/ u-pa, Polynesian fowl moa ichenencs: e Tugalo i-bon, Murray 1. a-bor, Erub i-bu fire probably contractions cf the Malagaso-Polynesian vuru, vurn-na, buro-ng Ke. which is related to the Seythic: pul-an &e. . The Tasmanian muta, Lampong puti are Dravirian, Paser has piata _ fowl, in the Abor form putah. ) The Binua pake is Dravirian or Bengali, | The same root is current as wing, feather, and egg, the specific conjoin- ed roots having been dropped. In many of the smaller vocabularies these words are wanting. But the larger ones furnish undoubted affinities. 1. The labial is yg in Abor a-pin (old Bhot. byu, Sambawa pio, dird), a-pu, Dophla pupu, Aka pa-puk, Kambojan pung, Koreny pa- bum, Murmi phum, Gurung phung, Sunwar de-phu, Mon ka-pa, Male kir-pan, Thochu ki-wost, (Naga-Nipal vu, bu, wn, va, wa &e. bird, Polynesia moa furl, Malayalam pui fowl, Gyami 38-phui bird). tis eather in Chinese mo, bo, man, Burman mui, Kumi a-moi, Singpho mun, Chonglo /hesphu, Mikir ar-weng (with mo bo comp the Tibeto-Ultraindian byu, bu, wa &e. bird 3 with mui, moi the Gyami s-phui, djrd, Pol, bui, toi, egg; with ar-weng the Kapwi weng bird). Wing does not occur even in Mr. Robinson’s vocabularies, avd as Feather is also wanting in Mr. Hodgson’s, the root will probably be found to be common with both of these meanings. Siamese has pi wing. The » Indonesian bang Matlura, (tir-bang to fly Malayu &e.), pai. Bugi , Balig- nini, baka Kissa, pak, pako Philipine, appear to be partly: Tibeto-w itrain- dian and partly Dravirian. iekmetee of the Scytho-franian and Dravirian forms current for Bird, Feather, Wing, have already been given, and it will be seen from the forms now cited that the Tibeto-Ultraindian labial for bird is immediate- ly connected with the Chino-Burman forms for feather, Seythic vocables. For Duck the dental form ‘is Scythie pot, pont, bata, Semito-African bato, bit-at, ma-bata &e., Indo-European pat, bat, bat-ak &e., Dravirian bata, bud-uhe, Indonesian patu, bati-ki &c,, and) Mon-Anam vit, pet, the Jast heing probably an archaic Scythic form, as itis also Semitic. The duplicated radical is found in Turkish papi, baby-sh, babu-sh. A simi- lar form is common in Indonesia bebe, pipe, bibi-ko &e. It is oauad Tibeto-Ultraindian, The word is not included in the gmall yoeabularies. Bibiko, bebek &e. were probably formed by a common Indonesian mode d not with the - ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 105, of reduplication from bik, bek, in whieh case the form is Mon-Anam, vit, pet. Naga has asimilar double form pak—mak. 3. The current Bhotian chya is Chinese tsioh, chiau, tiau, chio, tiy, &e. The Horpa gyo appears to be another variation of the Chinese, e Mishmi tsa, Mon kha-ten, Naga o-zah, w-so, u-zu, Manipuri mag bu, witi-tsa, wa-cha, a-ta, o-ta, ma-te, nyu-the, Karen tho, Kiranti chong- wa, Dhimal jila, Serpa jha, Newarjhango, appear to be all variations of a Chino-Tibetan vocable, which is probably of later diffusion than the abial. The Abor patang, petang, Dophla pata, Aka putah appears to be the dental and sibilant root (3) with the labial prefix. It may possibly be an archaic Dravirian form of the labial root not derived from Tibet, but having’ direct Ugro-lranian affinities. [See App. B to chap. V, Bird]. 4. The Jili ma-chik, Manipuri w-chek, »-thik-na, and the Sunwar chi-va, Anam, Binua chim, Mon ke-chim, Kasia /ca-sim, Gond sim, Si- long sisom | Komreng sisu|, appear to be archaic Chino-Tibetan forms, distinet from the preceding, and of earlier diflusion. Kwang-tung pre- serves the final k in its tseuk. , The Manyak ha, Naga au-ha, Khyeng hou, is referable to blrya, wa, va er tochinu, chya, sa. The latter was probably its original form, The Irawadi thik, chik is foandin Indonesia, tiku Tilanjang. The Sa- sak tiu preserves the Chinese form tio, tiau. The same root is found in the Gond ite, titit, andas Duel in Dravirian ite, Burman ute, and Indonesian itik, iti’, ite, titt. As Bird the root is archaic and widely diffused. Seythic doi, tirte, tori, tschir-pw &c,; Semito-African tauir, dide, dea, diury &c.; Sanskrit ati. ‘ The final m is included in the Scythicrange of finals, ziaf Aino, sibe-chu, shobo, shub-wn &c. Monyolian. In Samviede the same form is duck, shibu. For Feather the -—k form is common to Turkish and Tibetan, @ Turk,, shuk-po Bhot. The pure sibilant is Japanese asi (Sansk. ati bird) and Korean 20. : 5. The exceptional Maram a@-roi, Songpu v-roi, appear to be N. Dra- virian, ure Mundala, orak Uraon (urak &e. duck Turkish), 13. Ayr. The Bhotian g-rog-ma, Gyar, ko-rok, and Takpa rhok—po, preserve the same archaic form. The Abor ta-ruk, Aka ¢u-rak, are referable to the Gyarung brapch. Sunwar has the Bhotian rog—me-chi. In the Burma-Gangetic dialects the guttural final is nasalised and the prefix is uerally the labial as in.the Manyak, da-rak. The a vowel, variable ?, ¢, isalso more common than aor wv. The Mishmi a-ruang is an am- lified form found also in Burman pa-rwak-—chhit, parwet, ya—yuet, the First of which preserves the guttural. The form lang is Abor (trom rak asin Aka, raz Manyak) Jil, N. Tangkhul and Maram, he slender ling, leng, is Manipurian and Yuma, mi-ling, ma-ling, Ja-lin, pa-leng. The more prevalent Burtha-Gangetic term is a distinct root, chi, tsi, tsip, tik, chu, teng, ching &e- cha, tak, tang &c. 106 ETHNOLOGY OF THE INDO-—PACIFIC ISLANDS, pest iug Worns or Arr. Pi: House Village Roatl Boat Arrow [ron Salt - Bhot. wr. khyim yul tsho lam g-ru =o m—dah -chags tsha s— sp. nang thong lm | koa da chhya — chha syen Horpa hyo rha va che ‘gras L-das chu chhd Thochu ki’ we-kha gri phya jai sor-mo_— che Gyarung chhem wo-khyu tri bru ki-pi shom chile tu-khyu corde whet sha-b-ra Manyak nyé hao fri gu = m-as shi ché Takpa khem yu lem-dang g-ru m-la lekh tsa 1. JTouse. 1. The Bhotian khyim, Takpa khem, preserves the full form of the most common southern name. It appears to be one of the primary native roots of the family.” It is not Chinese, and although the tho- Iranian glossary like Chinese has the guttural root never takes the m final, the common forms being k-r, k-l, k-t, k-d, k-k, k-sh,,k-th Ugrian, Kamsch., Pasthu, Hind. &, The Milchanang and. Tiberkhad keum, T. keung, Cars Bhotian kung, Serpa khang-ba) suggest a connection with the : By heung Village K-t., hiang lu K-h. Chinese for house has uk, kwei &e. 2. Thesp. Bhotian nang, Manyak nyé (=nyek) may be remotely connected with the liquid root found in Chukchi, Japanese, Caucasian, Dravirian, Iranian and African. Southern forms. 1. The guttural is the most common of the Southern roots, and it takes many contracted and softened forme. The full form is retained by Jilikim aud Kiranti khim. Milehanang, Magar and Abor have archaic u forms, ¢-kum Ab., keum M. T: This appears to be also a local Bhotian form, one yoo. giving si kung, The Namsang, Kumi and Aka forms ei have been similar to the Abor; and the Mulung, Joboka and Mikir may be referred to the group. The Southern Irawadi varieties have n, Sak retainiyg the full form kyin. Khyeng and wr. Burm. pre- serve m. The guttural is lost in some of the Manipuri dialects, most of the Yuma, Borman, and Pwo Karen. F Hee final consonant is lost in several of the Naga dialects, Sgau Karen and Mon. The Newar chhen appears to be referable to the Gyarung chhem. The Magar yum belongs to the Abor-Milchanang band. It may be from a local full form like keum Milch., or it may be derived from the eastern extremity of the band where similar forms occur, Tengsa has yum vi/- lage. (See Village). e-kum Abor, hum Namsang, um Kumi, wu Aka; ham Mulung, Joboka, hem Mikir. . kim Jili, yim Manipuri; shim Luhuppa, shin N. C. Tangklvul, tsim Khoibu, chiin Maring, yin 8. Tangkhul, in Kapwi, eing Burm., Khyeng, Lnung-khe, Kumi, in Kami, ing Kyau, im Burm r., Khyeng, yen Pwo Karen, hi Sgau K., he Mon, kin Mru, kyin Sak, cha-ki horeng, , ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 107 kai Songpn, Maram—final ai for i as in other words,—ki Tengsa, Nocau Angami, a-ki Khari, khim Kiranti, dhim Murmi, khi Soman hing Limbu, chhen Newar, yum Magar. _ keum Mileh., Tiberkh., keung T,, kim M., khang-da Ser 2. The Manyak’nyé and sp. Bhotian nang are found in nagou, nak Garo, noo, nau, na Bodo, nok Mulung, Tablung, nya Deoria Ch., nha, ya Anam. The Anam and Deoria Chutia forms are connected. The others are more faithful to the archaic Tibetan forms. The root is also’ used for village (see Village). 3. 6-li Mijhu M., li Lepcha, lan Toung-thu, a-ru Champhung, ren Ahom, reuan Siam, heun Lau, hun Khamti, hon, ong Taying M., ron: Gond, ora, oa uraa Kol, er-pa Uraon, arra, ar-shyToda, illa-m, illu, ila Tamil &c, (See Village). ’ This root is used for village and it appears to be the same as the pre- ceding. The variety of forms and the mode of its distribution, show it to be very archaic. The Toung-thu lan is closer to the Bhotian nang than the other forms. The Khyeng nang village has the Bh. form, and in Anam and Magar it becomes lang. Abor has long, lung, Sin¢gpho reng, Angami rana &e. The Siamese reuan, Aliom ren, preserve the full form of the Lau fam. The vowel connects it with the Singpho reng village, but the Rakhoing rwa, (Horpa rhava) is probably more faithful to the original form. The Champhung a-ru, and the Mijhuand Lepcha li, appear to be both referable to the Lau reuan, ren &c. 4, ka ting Kasia, teng Kuki, tin Gurung, (?dhim Murmi). Thisis a Chinese word for rillage, town, and it is current with that signification in Taying M., several Naga dialects, Sak, Kiranti and Newar. (See Vellage). 5. pe-tah Kambojan, »ta Singpho, cha Dhimal, sang Manipuri. This appears to be a distinet group from the last. It is probably referable to the broad Chino-Tibetan form for v7dlaye, town, tsho, thong, so, in Gu- rong sa. The Gyami shhang and Sok pa-syang Aowse resemble the southern forms. » »% Village.,: 1. The Bhotian tsho, thong is Chinese, tsun, hiang tsun Kwan-hwa. 2, yul Bh. wr. (yul tsho), yu Takpa yul Serpa. Mongol. ail. 3. rhava({rha-ya) Horpa. Comp. ala, ola, ula Ugrian (Cheremish), ail Mong., and the liquid root for house. 4. Thochu we-kha, Gryar. wo.khyu, tu-khyu, Many. hu. Probably the guttural is the same root that is used for house. ~ Southern forms. 1, The broad form is very rare, song Garo, nam-so Murmi, na-sa Gurung. A slender form is prevalent but it has a distinct Chinese sourco. ma-tyung, ma-ting Taying M., ting Muthun, Mulung, Tengsa, ting~ khuaJoboka, ching Mulung, thing Sak, a-ti-gu Deor., teng Kiranti, Newar (kousé tin Gurung, 4a ting Kasia, teng Kuki). Chinese town, erty, ehing K-h., shing K-t. pis) ates ae ee 2. 3. ma-reng, me-reng Singpho, nkha-yeng Mijhu M., du-long, do-lung Abor, dung Toung-thu, rong Mikir, sh-nong Raa, nang Khyeng, lang 108 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Anam, lang-ha Magar, rana Angami, a-rame Moz. Ang., rwa Rakh., ywa Burm, ‘ ram Champhung, C., 5. Tangkbul, ram-khu Luh., rabang N. Tang- khul, nham Songpu, nam Kapwi, Koreng, Khyeng,?-nam Maram, nam- pum Aka, nam-so Murmi, na-sa Gurung. Py. ° These forms have several distinct affinities with Tibetan, The Khyeng nang, Anam lang &c. resemble the Bhotian nang /jvuse more than the forms now current in Tibet for village. , ; The Moz-Ang., Sak me Khyeng, Aka and Murmirame, ram, nam &c. is a group referable to the Horpa rava, of which the Rakhoing Twi is a contracted form. a The Tengsa yam, Nogwung yum, Khari a-yim, Khoibu yon, might all be referred to the liquid root, But as they are identical with soft forms of the guttural-root used for hose, it is probable that they take their eae Wy them. Comp. hewse yum Magar, yim Manipuri, yen Pwo ren &e, . The Maring yul is identical with the Bhotian yul, and the Khoibu you is probably connected with it. 4. ting-khua Joboka, khul Manip., Champh., ram-khu Luh’, khui C. Tangkhul, ko Kyau, koo Shindu, kwa Lungke, Mra, koh Mon, a-ti-zu Deoria C., go Dophla, ha Namsang, n—Kha-yeng Mijhu M., ke-p Male, ga Chentsu, yiu-go-ma Gadaba. kyong Lhopa, Lepcha. gaun Sunwar, gang Newar, The Mijlu kba, Namsang ha, may be archaic forms, of direct Tibetan origin (kha Thochu), ‘The others have the u of Gyarung and Manyak. The Lohja kyohg and the cognate Nipal forms appear to connect these pee with the guttural names for house, Comp, keung Tiberkhad, khang rpa. 5. muang Siam, mung Lan, ban Ahom, Lau, Siam, man Khamti, vang Kami, a-vang, ¢-wung, wang Kumi, te-wun Pwo Karen, tha-wo Sgau, nam-pum Aka, bane-yhe Limbu, m-ba-t Jili.. The lab. is Dravirian pa—da Uraon, ha-tu Kol, ym-tti, ha-tti,, pa-/li, ha-lli, mor-¢, mo-d, man-de, man-du, man-d S$. Dray. (House manei, mane, pei, vu-du ui-du, vi- da), The Jili ba-¢ preserves one of the Dravirian. forms, and the more com. man, ban &c. the other. The root is probably cénnected with the Chi- nese fung tsz house K-h. Samoiede has ma, mat, men house, mar villuge. 3. Road. , “1. The liquid root is Chinese lu, tau lu K-h., lau Hok-kien. The Bhotian Jam, and Manyak rd, are variations of one form {[=lam, lak], from the com. range of the final cons, It is probably a distinct archuic form. The root is one of the widely diffused ones of the ancient Asiatic lossary. Semitic do-rom, o-rom &e., African ji-ra, sv-la &e., Malagasy ae lala-mbe (path-great); Georgian sha-ra, Pers. sa-rak, rah, re, Pashtu lar, Kol ho-rah, ho-ren, da—ha-ri, Drav. sa-di &e, The Thochu g-ri’, aud Gyarung tri appear to be connected slender varieties. ‘The Takpa lem isa modern one of the Bhotian lami. The -m form thus appears to be the distinctive one of Bhotian, and the -ng and -k of the Sifan ialacts, The last is probably archaic, rak Many., nk, Tho- chu. The Persian and Georgian rak, rah, ra are referuble to such a form. 2. ché Horpa. Unless this rout be Scythie it is exceptional. The word is not given in most of Klaproth’s Scythic vocabularies. 1. The liquid is the common southerp root, STHNOLOGY OF THE INPO-PACIFIC ISLANDS, 109 ’ ‘The Bhotian form lam is found in Bingpho, Burman (also lan sp.), Kumi, Khyeng, Kyau, Kami, Sak (lang), ia lan-ti, Mon ga-lan, Ma- hipuri, Maram, Champhung lam-pr, Kapwi, Khoibu lam-pai, Koreng m-pwt, Maring lam, the Naga dialects (save Khatiand Angumi), Nogaung lem-ang, (Takpa), Tengsa ung-lan, Abor lam-beu, lam-te, Dophla lam-bu, Aka lam-tau, rama, lam, Bodo lama, Dhimal dama, Mru tama; Serpa, Lhopa, Limbu, Kiranti, Magar lam, Sunwar la, Lepcha laum ; Che liam, Mijhu Ca Newarlon, Jili tang—long, Siam thang-don. The slender Thochu and Gysrung form ri’, ri resemble the Khari ndi, ig Karen k-le, (Thochu g-n’), Toung-thu #-lai, and Taying Mishmi a- yim » ~ Takpa dang, Anam dang, Lau fam. tang, Siam thang don, sa-non, tha- thang, Jili tang-long, Toung-thu 4-lai-tan-tha, Pwo Karen phun tha, Mon dan. It appears to be Dravirian, tang-ora Savara, da-ha-ri Uraon ry Karnat., sa-di Tuluv., sa-ri Gond fey do-va Tulug., da-da Irula, i Telugu, ea Leet ork aie sibilant Songp' chang, Luhuppa song-vu, N. Tang! som-phu, C.'T. som-dut, — gun chah, Deotia Ch. tsa-gu, Mikir toar appear tobe referable to the The Murmi ghyam, and Gurung kyan, are probably reluted to the Gadaba kung-oru, Yerukala ye-gi. Pwo Karen phun tha, Toung-thu ta—phu, Mon kha-pan. Dravirian, ‘va—zhi, ba—te, pa-de, be-ie, ba-t, pa~ho-v1, mo-7-y (varying to the aspi- rate and sibilant ha, sa). The Kambojan chirada may be Arian, Ons. 1. e Bhotian —m form is the most prevalent, but the n forms eannot be radically separated from it. _ 2. The labial . connects the Abor and Dophla with the Manipu- rian forms. . 8. The Thochu and Gyarung slender form is only represented by a Karen name. _ 4. The Anam and Lau names are connected with the sibilant and dental forms of a group which ineludes Manipuri and Naga dialects as well as Jili and Takpa. ‘5. The & prefix is preserved in Thochu, Karen, Mon, Toung-thu, Murmi, Gurung. Me aa - oat. 1, The principal root isru. The guttural pref. is common to Bhot., Morpa, Manyak and Takpa; the labial to Thochu and STRUNG, The only variation of the root is im the substitution of a for u and the softenin or eliding of the consonant. The Horpara preserves the original form o: the Thochu of Fa ph-ya (for opp | 2. ‘The h. koa may be a contraction ofa form likef-rua. 3. e Phot. sp. syen is Chinese, chuen skip K-h., shun ship, boat, K-t. i . The southern terms have been yen in Chap. V. Sec. 11. 1, The prevalent forms of 1 haye the nasal final. The vowel aug- ment of the Bhotian koa is found with this final in Taying Mishmi ro-. wang (rua Brown), The Lau fam, has rua, reua, The Ka duak is the same*form with the final gutturalised ; and the Kambojan tak, tu-tuit, tup, pe ie dok are variations of it. Khyeng also preserves a k form, r 110 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, The Lungke, Khumi, Kyau, Kami, Mru laung, loung, Sak hau; have a different form of the double yowel. The Anam tau ship» is probably re- ferable to it, the Kambojan group eying also the dental for the liquid. It is. possible that these au forms have a distinct origin from the ua ones. They resemble the lrano—Dravirian and Asonesian nau, plava, falau &e. Aslender form of the root is found in Singpho, Jili, Karen, Kapwi, Moring, Ti Rakhoing, Maram, C. Tangkhul thi, Burman lhe, Mon leng, , Garo ring. The ht prefix of Bhot., Horpa, Manyak and Takpa is found in Jili ta-, Karen, Mon, Khyeng ki-, n-, Mikir t. In Aka it becomes hu-, in Nogaung Naga sv—, in Khveng ha—; Lhopaand Changlo have d-; the Jabial of Thochu and Gyar. occurs in Kumi, Kyau, Koreng, Khoibu, N. Tangkhul, Champhung, and Toung-thu. The sp. Bh. koa is found as a distinet root in the Nags groap khoa, khuon, khung, kho, Manipuri gr, khong, klo, kong, and Limbu khom-de Anam has ghe. Abor et-ku. 4, The Chinese sam pan K-t., san pan K-h., is found in Siamese for * ship kam—pan, sam-phao, ta phao, Anam s/oop tam ban, Burman song pua. 3. he Chinese chuen is found in Anam thuyen doat. Brown gives ding which must be the Chinese ting. Mulung has ye-sang and Ta- blung ih-sang, which are referable to the Chinese san. Ons. 1. The slender forms of ru, ra are only found in the Irawady branch. ‘The u form was evidently the original’in the South, and was. communicated to the Gangetic dialects, the same form being found in the Nipal, the ang 0 oe north ath bx pisces eres 2, The Tayin ishmi. rowang [=ruang] anc ugk, a to preserve an aathale form now lost Re Tibet hes presence in the Kambo- jan group in one form andin the Lau in another, accords with the com- arative antiquity of the Mon-Anam glossaries. The Gangetic forms onga, dunga &c are referable to this variety. The Dravirian and Arian Hoar: have distinct. terms, but none of these are found in the Mon~ = 5. Arrow. 1. The Bh. dah, and Thochu jas st that the archaic form was dak. The southern forms are similar to the softened Takpa m-la, aavek m-a). Comp. Jili ma-la, Singpho pa-la, Toung-thu and Kate p-la, pa-la, Burm. m-ra, m-ya (found also im Gurung, Murmi and Magar, and fur- ther contracted in the m-e of Kiranti, ph-ee of Garo unless these be me, a: hu Mishmi has the variation lo (lo wat), and a similar form Ju is ” found in Angami the lu, Songpu lu, Champh. ma-lu and applied to the bom in Tengsa lu. The Siamese luk (luk son,—son, /a-sun is bon’) has this vowel with the guttural final which archaic Tibetan appears to have had. » Kambojan has i-ruen, Mon leou, lay, lau. The other forms of the Lau family have the m final, lem Laos, Ahom, lim Khamti. ‘The Kasia‘’-nam is related to these. A slender form is also current in Karen /A-li, Lungke, Kami li, Doing-nak and Mrung le. Sak has the double term»to-li—ma-la, the second belonging to the Jili- Burman-Takpa-B} otian variety, and the former to the ola aren-Yuma, Kumi has li-ta-i. * The Thochu ja is found in the Khari le-jak bow. ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. Lil . 2, The labial root—found in Gyarung only in the slender and contract- ed form ki-pi—is the prevalent one on the north bank of the npper Bra- mahputra, and in older forms, probably Dravirian (a-bu, am-bu &c.) m-po Taying Mishmi, e-pug, e-puk Abor-Miri. The Garo phee and Kiranti me may be this root in slender forms similar to the Gyarung and to the Kondh pin-ju. It is found in the Khamti lem—pun, M. An- gami thi-wu. 3. A very common southern root applied to the bom in some dialects and to the arrow in others appears to be of Chinese origin. Chinese has for arrow taien _K-h., tsin, chi K-h., ten, chen, dian &c. in other dialects. Kambojan has ting, (bow) Anam ten, A broad form is more common Ka tong, Siamese son, ka-sun, Limbu tong, Lepcha chong; Kapwi than, Namsang la-chan, Muthun, Joboka san, Mulung, Tablung Ja-han, Teng- sa la-san, Nogaung la-sang. Angami has po-si dow (M. A. Porbhy; Mru sa, Kumi ta-i, tsa-koi, Khyeng thwa, Deoria Ch, a-ta, Mikir tha-]. 4. A guttural root occurs for arrow in Khari éa-khaba, and for bow in Namsang doa-khap, Muthun and Joboka hap. : Ons. 1. The older lrawady forms—Karen, Yuma &c., have the Gya- rung # pref.; the later Singpho, Burman &c, have the Takpa and Bho- tian m pref. The Burman form has spread to the Nipal vocabularies. 2. The Mon-Anam forms are older than any of the Tibetan, —_s 3. The Abor preserves an archaic and probably Dravirian form of the de root. . 4. The slender form ofla, luis characteristic of the Karen-Yuma “oup. ce 4 The broad form of the sibilant root connects the Lau group with the Naga. This form has spread to Limbu and Lepcha which have the Siamese vowel and not the Naga. Possibly the a form of the root is Arian and not Chinese. , 6. The same word is applied to arrow in one dialect and to bow in another. In some dialects both have thesame name as in Nogaung la- sang (properly bum’s arron’). 7. The name is frequently compound. 6. Iron. I. The Bh. wr, chags is a broad full form of the sp. chhya, Manyak zhi. The Horpa chu, Gyarung shom (prob. sho-z),bave a distinct vowel. The Thochu sor-mo appears to be connected with these forms, Chinese has the sibilant root tit K-t., tie K-h. (comp. salso sik fan K-t., se K-h.). But the archaic broad Tibetan form is more immediately re- lated to Seythic forms, thu-mar, tu-mur, also the-zaar Mong., tup, tip ae. soi, suy Korea, The s-r,s-l form of Thochuis probably connect- cL with the similar Scythie forms applied to irvn, sello, selle, zhilla Tun- gus., but more commonly to gold sor, shor, son, sir &c, The older ap- plication of the sibilant root was to silver, salt &e., the root. itself bei that for white. Jron was afterwards distinguished as black-silver an silver itself as whtte-silver, and this led to the form for silver acquiring a ying meaning (metal). See chap. v. sec. 11. Chinese hak kam K-t., 1¢ kin K-h. for iroh i. e. black metal; and pak kam K-t., pe kin K-h. - for silver i. e. white metal. Gold is kam, kin, with or without the quali- tive for yellom. Some other examples in addition to those given in chap. v. may be 112 ATHNOLOGY OF THE TNDO-PACIFIC ISLANDS. taken from the adjacent Scythic family. The Yakuzki Tungusian ho- hese old is . variation - He yet for peel haba i, the Lamuti sree aty—s $c dame ae 0024 I and the common Se altyn, nitan altun, appears to a waorition of the form for red, the full root being kula, hola, {ula-tyn=kula-rin], in Ugrian gor-d, gor-le, kel-ban &c. Some Turkish dialects have kysyl- kumyos red-silver, or simply kysyl; Yakuti pretixes the word white to silver uryon-kumus. In some of the Lesgian dialects the same root arats, araz &c, is both white and silver ; and it recurs with the former meaning in Nilotic dialects arsa Woratta &c, 2. I\ekh. This exceptional Takpa form is probably a variation of the Tibeto-Burman root for bluck nak, lak, reg &c. &e. (See p. 25 et seq.). Southern terms. 1, The broad form is found in the Burm. than, Mee Kumi ka-dang, Songpu n-tan, Nams. jan, Muth, jian, Job. Mul., Tab. yan, Angami wir i (M. A. je), tha Karen, tai Toung-thu.. : archaic Bhotian guttural is preserved in the Khoibu sak-wa, Ko- chag-hiy Anam sat (comp, Chinese sik tim, tit aon), Tiberkhad Slender forms are common, teng-gri Mijhu M., tsi, si Taying M., thin Kapwi, tin Luhuppa, thir Maring, thiar 8. Tangkhul, Ri-ying Kyau, ey Tengsa, Mogae. ean Khari, je Moz. Ang.; tir Kyau a thi Khyeng, ings chie Mikir, shein Kami, thein Sak, say-thi Toung-thu ; Dhi- Ir, n-jing". . The u, o ah oF Hove, Thochu and Gyarung is found in Garo shur (Thochu sor-wo), Bodo chur, Mrung tcho, Angami the-ju, Deoria Chutis sung, Mon pa-soe ka-sway, po-thway. . 2. The Takpa lekh is found in the lek, lik of the Lau fam., dek, dik Kambojan. Some Irawady terms which 1 formerly considered to be Dravirian, appear to be variations of this root. wa-ru N, sr an ma-ri 8. T.,mp-ri Singpho, a-ruk Champhung, rung, run, p-ron Mile — Rok, luk, lok are current forms of the Tibeto-Bur- man root for black. The same pee aero occurs with the guttural pref, in the Abor yo-gir, yo-gid, ya-gurah, Mijhu teng-gri, Sunwar wa a-kli, (Singpho brass ma-gri). e analysis is probably g-ri, X-li, gu-rah &c, 3, ha-pha Maram, ta-phi Jili, ¢’-mau, ta-mo, ta-mhu Khumi; Lepcha pan-jing, Limbu phen-je, Kiranti, Magar pha-lam, Murmi pha-i, Gu- rung pa-i, Sunwar wa akli, Ka mam. Kiranti, M pha-lam. The labial is Dravirian, panna Uraon, karu-mban, ka-bina &e, 8. Dray. 4, nar Kasia, na Newar. Ons. 1. The broad and full Sifan form sor, sho(r),chu(r), is found in Bodo and Garo; the more prevalent slender forms in r—alsq— referable to it—are chiefly found in the Manipuri and Yuma group. The a form distinguishes the Burman, Karen and several Naga dialects; it may be Bhotian, chang, than for chag. ; 2. The liquid name appears from its variations to be archaic. The slender full form connects Takpa and while the contracted form, with the rutt. pref., is common to Singpho, Mijhu, Abor and Sunwar, and, with the labial pref. is com, to Binkybe and 8. Tangkul Broad forms like the Champhung a-ruk, and Milch. ran, p-ron—found so far apart—sug- t that this word, derived from the native root for black, was the earlier iffusive Tibeto-Burman name. It may have been that of the first Hi. ETHNOLOGY oF THE INDO—FACIFIC ISLANDS. 113 tmalaic vocabularies that were carried south (Mon-Anam or Lau). The names for the other metals are not paves in the short vocabularies, and any comparison of those known would be very imperfect, But I vive those for silver and gold, as some repeat the rvots that are used for irons 7. Silver. 1. The Bhotian word for silver is Chinese, K-t. ngan, K~-h. yin, Bh. ngui. The Lau fam. has neun, ngon, Namsang, Joboka ngun, Mathun newun, Toung-tha yun, M une nmin-mane, Tablung toi-nan, Anam nan, The Lau word and the derivative Naga have the archaic u preserved by Bhotian. The root is probably mhite 2. ' 2. Khari has a-tsun, Mikir tana, Changlo tang-ka, Mon thaun, This is one of the roots for white. A different form preserved in the Siamese sit white is eee to tia in Chinese, sik K-t., se K-h.; Anam thiek (K-t.); Siam di-bnk (buk white). | 3. Angami has roko, aka. This is the liquid root for white long &c, The Kambojan p-rak has the same root, but it is probably from, e Malay perak, in Chuinpa preak. The Angami raka throws doubt on the Semitic origin of the Indonesian perak, pirak; and the silaka of Javan &e. more than strengthens it. The root of buth words appears to be the Angami raka,—pe-rak, si-laka. The prevalence of this root in Indonesia and its rarity in Ultraindia’ may be attributable to the native ‘Tibeto-Barman term having been generally replaced by the Chinese. ihari preserves a native name, a-tsun, which is ulso 9 root for white « engsa and Nogaung have adopted an Arvan word, rup, lup; while the other Naga dialects have received the Chinese name, 4. Singphokum ph-ronyr (metal white). 5, Anam bae (white, buch), 6. Abor a-mel, Nara mang, Murmi mui, Milch. mil, mul, Drav. vili&c. (See ron 3, and chap. 5. sec, 11.). 8 Gold, 1. For Gold Bhotian has ser, si, Changlo, Mikir ser, Kasia k-ser, Hindi gar, Pasthu sar, Muthun sien, Joboka sian, Silone sin, Karnataka chine ni, Rakhoing ahwi, Burm, sui, Karen tu, Mon to, tha, thau. | ‘Both the Tibeto-Burman slender ser, and the Pashtu sar, are Scythic, ser-ne, sirene Wolg., sar-ni, sor-na Ost, &c. The Aryan hir-na, hira-nia are evidently from sir-na,—sona from gorena; snyar-na is an amplified rm. . The Siamese thong appears to be from the Pali sona, like the Milch, guns. 2. Mikir dor-d: (rok-dor iron Dophla). This is prolnbly an archaie form of the Scythic sor, derived from aSifan dialect. The Sifan names are not kuown. 3. Abor a-ngin, a-ina, Doph. a-en, Angami li, Khari ta-ru (ron 2), » Siugpho ja (Jron 1). 5 The Chinese kam K-t,, kin K-h, is found in the Lau fam. kham, thon Kham, Namsang kam, Mulung, Tablung kham, Anam kim, Toung. thu khan-ni. ~ 6. Acam has vang, Laos wang (yellom Ch.); Ch, hwang kin K-h., wong kam K-t. yellow-metal. 7. The Kambojan niias is Malay, mas. } , - Ops. The southern names for the metals are mostly Tibetan. A few w+. 114 ETHNOLOGY OF THE INDO—PACIFIC ISLANDS, Dravirian terms are preserved in the North Gangetic dialects, |The Mon- Anam vocabularies coutuin archaic and modern Tibetan, and archaic and modern Chinese, names, The Chinese names appear to have dis- placed the native ones in several dialects, 9, Salt. All the Tibetan dialects have the same word,—Horpa, Bhotian and Takpa having’a broad, and the Sifan dialects a slender, vowel. ‘The abrupt ac- cent of Horpa, Thochu, Gyarung and Manyak indicates an archaic gut- tural final. The root is not Chinese, but Scythic. It is common to nll the Scythie families, and in most of them it oceurs both with the liquid and dento- guttural finals, and also without them, Korea sok-om, suk-vim &v.; Ye- niseian chhyg, Sumoiede shak, sak; ser, sir, si. Ugvian sich, sek; chal sal, sol, sal-va ; sot, sow, 80 ke, ‘The rooteppearsto be thesame thatis used for mite, andapplied to various white or bright objects, moon, silver &c, &e. (ante p. 2H). ‘Thus Samoiede has sirr, seri, kyr &e. white, sir salt, servan-bire silver ; Ugrian has sor-ny, sar-ni, sair-em whit’, sol-ni &c. sult, sur-yn silver. The Yeniseian chhyg sait ovcurs for white in-tyg-bis Yen., tyag, chaga &e, Samoiede. The Armenian ag salt appears to be the Uprian and Turkish ak, agh, white. The rvot is one of the archaic and widely diifused ones of the preto-Scythic vocabulary. Caucasian preservesa primary funn shug, chush. It basalsof rms in m, n, and vocalised,—zam, zyam, zon, zun, chin, Zio, ze, p-su Leese. Mingrelian has zhuini and Lazian chumo. In Cherkesian the same root 18 white and salt,—chush, kush white, Chush sa/t. The sibilo-liquid form is foundin the {udo-European family,—sol, sul, sal, ealz, zout .&c..—and in afew African languages,—sira, Sinn Malagas sy, singa Shangala, gi-sili Ilausa, gi-sile Karekare, dv-gu-sin, ge—seli Bode er., go-da-sin Baehermi, i-si Mandara, dsi-sem Kandin, [sale Kabenda, ndzolu Mimboma, prob, Jurop.|. An om form occurs in Gonga shum-be. Au older Seythic torm occurs in Penin sek, Ham tok, Gouli sap—lele ; anda sibilant root without a final cousonant is com., cha—wish, til Silotic; chua, cha-ai, le-dae, m/e-i, mo--0, un za, m-ta, n-to, wu-si, w—wo-tls, a-dsi, 1s, 1-20, esi, Ja-se, Ysa, mi-si-s. The Semitic root isditferent,—melach, mileh, melh, walbut, an itisfound sporadically in Afiica, mrh Egyptian, miluk Dishari, mer Landoma, Baga. The more-commen Atnican roots ave the liquid, v-uun o-nuk, ¢-non, ba-no, a-n0, e-ro, bu-ro, ju-ro-ne, pu-bamn, lum-dam &c. an the guttural koro, kiri, kon, a-koli, a-kan, mo-nyeus, pe kua. From the Georgian mirili, and the weueral elossarial connection between ‘the Semito-African and the Caucasian vocabularies, it is probable thatthe mel, mil of the Semitic word is a distinct rout, and 4hat the Chaldee mil- chu, Syriac mel-cho, Assyrian wil-cha, preserve the compound best, If 80, the second element would appearto be the Seythic, Tibetan, Cauca- sian and African root,and the first the Semito-African, Dravirian and Bcythic root tor white that is so prolific in names of white and brizht objects (ehap, v. sec. 11, Iron, Silver), The form of the second element chu, cha, cho, isthat which the sibilant root retains in Tigre, Agau and Gonga, cha-wi, chu-a, shum-be (or shu-mboe). In the south the Tibetan root is very common, Most of the forms ap- ear to be more archaic than any of the current Tibetan, and to be refera- le toan-m and not toa-k variety. Possibly tue Chinese yam, im, - @ MTHNOLOGY OF THE INDO-PACIFIC ISLANDS. 118 yen—which would otherwise be neither Scythic nor Tibetan—is a soft form of an archaic Chino-Himalaic form, sam, sim, as the yum is of sum. The -m form is rare in Seythi¢, and the Caucasian zain, zug &e. favours the archuic existence of a similar Chino-Himalaic form. Tue Mijhu tue m-yin is probably a modern Chinese form. P - Siazpho sum, tsum, jum, Jili cham, Manipurt thum, Namsang sum, Muthan, Joboka, Mulung, Tobluag hum, N. Tangkhul n-tsu, Nogaung méa.tsu, Bodo shyung kare, sayung kri, Deoria Chutia sun, Sak sung, Limbu, Kiranti yum, Lepcha vom, Sunwar yu si. Maram n-chi, Songpu n-tai, Koreng ma-tai, Avung tn-chai, M Ang, ma-tse, Ang. meé-tsa (by invers.), Kapwi, Luhspie, G. Tangkbul, 8. T.; Tengsa, Khari ma-chi, Kuki cli, Khoibu mé-ti, Maring ti, Mikir img-ti, Pwo Karen thi, Lungkhe she-te, (nghet-te ant, wat-ti egg &e.), Kyau, Khyeng ma-tsi, Dhimal de-se, Sunwar yu-si, Newar chhi, Savara dba-si, Gadaba di-ti (= mi-ti. Khoibu). | ~ Champhung ke-sam, Changlo im-cha, Burm. chha, sha, Toung-thu - ta, ta-thah, 8. Karen i-tha, Mru wi sha, Magar cha, Gurung, Murmi cha- cha, Tiberkhad, Milch tsa, Garo syany, kara sam, (the guttural is the Mindi khar, Bengali khyar potash). 2. Siam k-leua, k-lua, Alom /-lu, Laos k-eu, #-em, Khamti s-u; Kasia m-luh, Kuti ma-lwe, pe-loi, Kami wa-loi, Abor a-lu, @-lo, ela, Doptrla a-lo, T. Mishimi p-la, Pwo Karen la, Anam loi. This root is Chi- “nese, iu. bt appears to be the com. ‘Tibetu- Burman liquid root for whi‘e, whic occurs with similar forms (}. 28, 29). The Bengali nun, Hindi lon, Sindhi, Asam lun, Singhalese lunu, Kol, — fbu-laig, bu-lans, bave the sime root in the Mijhu, Garo, aod Yuma forms for white, lug, jung, nung &e. The African liguid names may have an Tndiay origin. ) 3. Anam muoi, Mon bho, Kambojan am-bil (? vom Lepcha). These words have no wfiinity with the Tibetan or Chinese roots, and they must be relerred to the Draviriun element of Mon-Anam. The 8S. Drav, dia- » Jevts have uppu, uppa, upp, Kol bu in bu-lnne, Male be-ke, Uraou be-th, All these Draviro-Anam words are connected with the Thochu, Mon- Aimm and Dravirian labial root for white. The Kambojan am-bil hus the Drav. form found in white, silver, moon, otar &e. The others resemble the Kel, puredi, pun-ta, Tuluva bol-ane (the other $8. Dray, dialects have the sleader form vel, bil), \agw a-po, Suawar bwi Ke. white, und some of the Gangetic forms tor stlmer, wul Mileh., uui Murmi &e. The root is Scythic, white boi-nat Yukahiri, woirkan Upvian ; silver hopia &e, Min, um Korea. From the forms of the Draviriau and Mon-Anai root in ita different applica!ions, itis probuble that it beluuys tothe Seythic basis of the archaic Deavirian glossary. The —k forma tor m/ete in Lau &e. are Chi- no-Tivetwn, (Uhocha)., The use of the labiv-liquid root tor suit is com- mon to Georgian, Semitic and Dravirian, The Goud sa-har (sa— is pret. in some other words) appears to preserve an archaic full form of the labial root hke Kaimbojai (bal, val is the Kurzi and Tuluva form, marin mar-sal dgit Wol.). The Pashtu mulga may be the same root, mal-ya, The Sanskrit la-van may have the same liquid root. The postfix oc- ¢urs in Seythic and African words also and may be the Scytho-Dray, rogs for white. Australian has wil-buu wh.te (wil Drav.). ne ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. * Sec. 7. THR FORMS AND DISTRIBUTION or TE Curxo-Hiatatc Numerité in Cucva, TIBet, [xpi AND ULPRAINDIA, CONSIDERED AS ILLUSTRA- TIVE OF THE ANCIENT RELATIONS AND MOVEMENTS OF THE TRIBES OF THIS PROVINCE, AND OF THE SECULAR CHANGES IN THEIR GLOBSANIES, Any further remarks on the distribution of the numerals must be chief- ly supplementary to those which have already been offered. But some general points of interest are brought out by a closer comparison of the southern forms with each other and with the Tibetan and Chinese, It is clear that each of the primary numeral elements—detinitives—has assumed a great diversity of forms in different eras, or in different dialects, The most prevalent Chinese and Tibetan forms are not those which ap- pear to have been so at the epochs of the tirst diffusion of the Chinese: numerals in ‘Tibet and of the Tibetan in Ultraindia, But as the vowel’ probably varied from a very remote period, the subject does not admit of our attaining positive results beyond a certain extent, a Upon the whole evidence, direct and collateral, it may be concluded that broad forms of the guttural or dental definitive were used as earliest unit, and in higher numbers formed from the uwuit; and that the Chinese series was— _ (TIL) 2 +1,..............contracting to 1. A eS oe Ae ee > op ae (V.) 3 + 2,ie.2+1Land2 ” yy 2s EVES OR res in 2 Si (VIL), 6 HA senccteccene » 9» 1. ( Vill.) 10-2 (10 being 1) .. ” » 1 or 2, or 2, 1, Path) Se pare ene eee Lana ae ere ae 1 ( X.) SS pr a Ar er ae a Tn sev. 4, I considered it probable, from analogy, that 5 was 1, aly, though differing in form from all the definitives used in the lower numbers, save one of the varieties of 2, That itis really 2 will appear on a come parison of all the forms, The root for Gappeared exceptional, and it wag considered as being an archaic N. B. Asian ternary term, 3, for 3,3. Aw _8 isl, the root, whether considered as § + 1 or 3 + 8, would be the same _in form, solong as the two definitives used as 1 and 2 formed higher numbers by simple repetition without agglutination or contraction, IL now think that luk 6 can be explained as a normal Chino-ibetan form of 1. The Chinese 7 was shown to be 1 (for 6 +1); but the Tibetan, with the exception of Thochu and Manyak, to be quinary, The Chinese, § was shown to be 10 ue 2, 10); hut the ‘Tibetan 8 was considered to be 4 2, for 4 the 2d several of the Southern dialects the name tor 8 is a root for 10, that for 2 being lost. rome mu-tsat, Bodo jat, Avrung fi-sat, Namsung t-sat, Tablung thath, Kbyeng sat, Toung-thu ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 17 that, Khari sa-chet, Muthun a-chet, Garo chet, Burman shyit, Nogaun® te, Deoria duqu-che, Chong ka-ti, Ancami the-ta, Newar chya, Tengsa the- sep. ‘Chis is the Chino-Tibetan 10 (itself'a form of 1). In sec. 4 the identity of ther, 1, clement of 8 with 4.was pointed out, and the inference drawn that8 was 4, 2, i.e. 2d 4,thesecond element resembling some forms of 2. But these common southern forms of 8 appear to be reconcileable with the northern. In the remarks on the Chinese and Tibetan 2 and 4 the radical identity of both was shown, and it was pointed out thatthe Tibetan liquid form of 4 preserved a variety now obsolete in the Chinese 4, although praserved in 2, “one of many illustrations ofthe great antiquity of the tirst diffusion of the Chino-Tibetan numerals.” The probably denary orizin of the Chi- nese 8 and ) was also noted. In mv first comparison of the Chinn-Tibe- stan namerals (App. C.), [ remarked the accordance of some forms of the liquid element in 8 with liquid forms of 2. If we consider this element as in all cases representing 2, the southern numerals of the Tibeto-Bar- min family are reconciled with the northern, and both with the Chinese; and this view I now consider the right one. According to it, the Tibetan Sand 4 retain forms of the Chinese 2 distinct from the current nasal ones, The common southern 4, w-li, pr-li, ba-li, b-ri_ p-re &c. is an an- cient Tibetan form of 2 and 4, corresponding with the Chinese li, liang 2. A Tibeto-Burman full form with the Jabial prefix, asin 4 and 8, is also reserved in one of the most archaic of the southern dialects, Mru, which as p-re 2, (comp. Bodo bere 4, &. &e.). This appears to have been re- laced in 2 of the other southern dialects by the later prevalent Tibetan 2. ‘he second element in the Tibetan.8, gyud, gyet, yet, eye, é¢, must be» form of the unit representing 10. The Mru ri-yat 8 ‘has the two pure roots 2, 10, as in Horpa and Gyarung. * ; From the preceding ‘Table it appears that the definitive nsed as the unit is found in 1, 3, 6, 7, LO, and, in some dialects, in 8; and that the definitive used as 2is found in 2, 4,5, and,.ia some dialects, in 8. These two detini- tives may be termed the unit and the dual. The unit beiny in its origin a demonstrative or definitive used quali- tively or discrimiuatively, its vocabulary wis propably coextensive with that of the definitive in the older «lasses of language. In -the Chino- Himalaic defiuttive system the normal unit def. was the guttural, varying to dental, sibilant-and aspirate forms, ag the same def. and unit does in all other formations in which it occurs, Scythie, Semito-African &e. It also varied to the liquid as in other formations, but this form was eom- paratively rare. In Tibetan this def. is now applied to inanimate substan- ces, the labial being the primary animate «ef. he Chino-Himalaic nait, in itsearliest cognizable stage, had two forma, one hiving the lubial final, and the other the guttural passing into the dental, the vowel being broad in both, u, au, a &e, Extant examples, gyud, kyok, kyot, kat, at kang &c,); sat, tsat, chat, jat &e.; zum, sup, sin, tham, sap, chap, (dun, sun, tun, song, san &v.); luk, rak, lat (ram mod.,rang, nung’ &e.). Of these forins those with tle rutturalini- * As the Chinese y of yat, yit = 9a. to have been formed, not hy a conversion of ch a haf by a hurdening of the vowel, from an am- plified vocalic form similar to 9, it is probable that in @yud, gis the ra- digal initial, and yu merely the amplified vowel. Comp. the Lepcha kyo img, kyok in 8. ‘ i f1s ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. ° = tial are probably the oldest. Forms in t, d, ch, z, s, 1, r appear to have been formed from it at an, ancient period, and hefore the system spread into Tibet... When it was first carried into the Himalaic province broad forms only prevailed, and these are still the most common in it. In the next or 2d stage—that preserved in the southern Chinese dia- Jects,—slender vowels were developed, but the final consonants were re- tained. The Bhotian 1 chig, chik, is referable to this condition of Chi- nese, In the 3d or latest stage, or that of the modern Kwan-hwa, the vocalic and elliptic tendency setin strongly, In the Kwan-hwa phonology all the consonantal finals, save n and iy, have disappeared. The ouly nume- ral forms referable to this staye that are found in Tibet, are some of 1 and 10 (sih, chi, che, thi, ti), Its influence is chiefly marked in the con- tractions of the native vocables. Many of the broad archaic forms are still preserved, although vocalised by the loss of the final consonant. In others the vowel las become slender. If the formation of the Kwan- hwa phonology only began in the Tsang dynasty (A. D, 620 to 907)—as Mr. Edykin believes—and the loas of the final consonants took place after- wards, its influence on the Himalaic province must be very modern, Probably it dates frem the conquest of eastern ‘Tibet by the Chinese in the 12th century. ‘The 3d stage being brought down to a period so recent, the 2d will not require to be placed at a very great distance behind it, A second Chinese def. used us the unit was the labial, It appears to have been distised at a very remote period, and before the Chinese sys- tem was carriei to Tibet, as it is only ‘extant in the Chinese 8 pat &e. (for the unit of 10), and 100 pak &c. This is on archaic form of the masc. labial qualitive and definitive of the Himalaic system. Rs normal dual def. was the liquid in n, ng, |, r, variable to the al t. ForMs OF THE UNIT DEFINITIVE, Broad Forms. Ist, u, o forms, ITand X. Thearchaie broad forms have, tn most of the Chinese and Tibetan dialects, been changed into slender ones. Some of the Chinese forms retain final k, t, (in 10 p.) The oldest forms appear to have been kuk, kut, tuk, chuk &e., kak, tak, chok &c., and these are current in N. E. Asian systems, Bhotian in 10 preserves a similar form }-chu, Serpa, Gurung chuh. The great antiquity of this form is corroborated by its re- tention of the archaic labial def, as in the Dravirian pa-du, ba-d &e. 10, Seythie be-t 10, bui-ke, ji-to ke, 1. A similar broad form is found in the 1 of Sean Karen tu, Pwo Karen ka du, Tengsa kha-tu, whence the con- tracted Kuki, Magar aud Lepcha ka-t. Thochu retains this form in 10 ha-du, a Tibetan form of the root equally archaic with the Bhotian d-chu. The archuie lrawadi forms clearly associate themselves by their pretix with Thochu, which probably preserves an older form of the Gyarung ka-ti,—u in the slender phonology becoming i. The Lau nung, (ling in Ahom), is referable to the Tibetan d, r, 1 forms, Il. The same archaic form of the unit is found inthe 3 of Bhotian.. ~q-eum and Horpa su. Itis the most prevalent form in the south and ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 119 probably the oldest, being that of the Yuma gr., Karen, Toung-thu, Bur- man, Singpho, Dhimal, Nipal&c.—sum, tum, tun, sung, thun, thu, song, su, Win, OD. Vi. The Chinese luk appears to be a liquid variety of the fall archaic u form. ‘The u, o vowel is retained in all the Ilimaluie forms. In the ‘Tibetu-Ultraindiau dialects the prefix is the guttural, dental or sibilant, The root varies to ru, ro, rau. The dental and palatal varieties are found in the Bhotian duk, tuk, thu, Gyar. tek, Horpa chho, The last ia probably a primary form of 1 similar to the Bhotiau chu of 10. The others may also be more full primitive forms, but the lusmediate source muy be the Bhotian druk, and druk may be the Chinese luk with a Bhotian prefix d-ruk, Possibly dr is an archaic intermediate fori between the deutul and the liquid. VIL. Manyak s-kwi, a form of the unit preserved in 9, Lepeha kyok (=kyot of 9), a remarkable archaic form. The Bhotian and Lhopa dun, Serpa dyun, Changlo zum, I now consi- der native, for the reasons given elsewhere. Clianglo lus the arehaic form of 1 preserved in 5. Mijhu has nun, a form corresponding with the Lau nuny f. Vili. Ghotian gyud, Sunwar yoh (? Mijhu ngun). IX. Both Chinese and Tibetan retain sehats forms, kyeu, kin, gu &e. Bhotan, Tukya aud Horpa preserve the inanimate pretix, d-yu, du-gu, r-gu. lu the south it is found in Garo sh-ku, Bodo s-ku, ch-ku, Singpho tse-ku, Bhindu cku-ku, Noganng ta-ku, Tengsa tha-ku, Khari te-ku, Kumi fa-ko, Ikumi fta-kau, Milch. s-vol, Karen khwi. The Lepeha ke kyot isau archaic Chinese form similar to the Bhot. gyud of 8. The (rya.ni chyu is an instance of the change of the archaic guttural into ch, which has taken place in the Chinese 1 and 10. X. Bhot. b-chu; Thochu ha-du, Kami ka-su, Sak si su. The Kuki sum-/ca preserves the full archaic form. 'Tenysa has the liquid variation the-lu [==the-log, ta-ra &e. in 6). C. chum-wari Kumi, ro-k-ru Nogaung (10 X LO), ta-loyen Toung-thu. 2d, a forms. The a form is also preserved in several dialects, and is probably weval with the u form. I.. ra Horpa, a Thoehu, ta Manyak. This was evidently a common Tibetan form at one period. In the south itis Yuma, Burman, Karen and Toung-thu. From tho distribution of the a forms they appear to be of the same ave with the u forms. Comp. Tengsa kha-tu, Nog. ka-tang, Ssantuor ta &c. The Burman wr. tach [=tany Nog.) preserves the archuie guttural final. ILL. The Chinese form is sam, varied to sang’, san, sa, ta. In Tibet it is only found iu Gyarung, 2a-sam, whence it has been transterred to the south, being the form of Mijhu, Mikir, Garo, Bodo, the Naya groups (save Angami), Dophla, Changlo, Lepcha, Sunwar. 1t probably belongs tu the latest Gyarung current, In some of the Naga dialects the sibi- Jaut is hardened iuto z, r, 1—zam Muthun, ram Nameany, lem Tablung. ” Kasia has lai. VI. The a form is very rare. Thochu has /kha-ta-re and Mijhu ke- than, the Gyarung 3 form, VIL. Chinese in Kwang-tung has tsat, Thochu s-ta. 120 ETHNOLOGY OF TIE INDO-PACIFIC ISLANDS. VIII. The same broad form is found in several of the southern sys- tems, representing 10. It appari to have been the old irawadi form— Toung-thu, Khveng, Naga, Singpho, Bodo, Newar—tsat, that, snt, ta &c. The Takpa ya, Mru yat, Komi and Kumi ya, are from an a variety of the archaic form preserved in the Bhotian gyud. : 1X, The a forms appear to be referable to the archaic amplifications ofu. Thus the Dophla kayo appears to be a broader variety of the an- cient form preserved in the Lepeha kyot. The Kwang-tang kau is a simi- Jar Chinese form, perhaps more archaic than kiu, kyeu‘&c. The Lau and Kumi kau are referable to it. The nang of the double Abor term ko-nang-ko, ko-nang-e, ia an arehais ‘10 found.in the Dophla rang 10, and corresponding with the 1 of Horpe ra, ane Lau nung, ling &e. The Taying ko-nyong :9 has the broad Low yowel, ‘Dhimal has similar double form ko-ha-long, ond Taying has ha-long. X. Chinesechap. The Horpa s-ga, s-ka, Garo s-kang, is a vocalised variety of a still more archaic form. The Lhopa cha of cha-tham has the ‘Chinese form. Newar san-ho, Khyenge ha, Kami ha of hasuh. The Bhotian tham-ba_ appears to be an archaic form corresponding with the Chinese chap, Kuki sum. Dophila rang, Kiwri ta-rah, Shindu me-rha. (See I). Kumi has hau, an amplified form:corresponding with kau 9, C. geva-tham-/a Bhot., gya Serpa, cha-the (10 X 10) Nams., ya-kha Shindu, ta-ya Sak. . Bunwar s-wai ka (1 X 1 for 10 X 10, ka 1 Sunwar, chi-wai 10 Murmi). thya Horpa, lat-sa (10 + 10) Singpho, ra-sa Kuki, #-laat Khyeng, ta- ra. Kami, Burman wr. (ta.ya 8).). Slender Forms. I. Chinese chit, chek, yit, it, ih, i’, i. Bhot. chig, “Gyar. ti. These slender forms have made little progress in thesonth, and must have been received subsequent to the inigration of the broad forms preserved in Manyak, Horpa and Thuebu, Ill. ‘The only slender forms are the Thochu, Manyuk and Sak. VI. There are no slender forms. VII. ‘The Chinese:forms are generally slender, Asan archaic broad ‘form is found in 3, it may be a whether the Chinese 7 did not retain its full form 6, 1, until after the development of theslender form of 1 f VIII. Garo chet, Muthun a-chet, uri -sa-chet, Burm, shyit, Nogaung te, Deoriadugu-che, Chong ha-ti, Tengsa the-sep. These ure all referable to the modern 10, IX. The moderiu-Chinese kiu, kyeu, are not connected with the mo- dern slender, 7 or 10, but with the archaic form of 1. They Tibeto~ Ultraindian forms ave all broad. X. ‘the Chinese slender shi, ship, sip. &c. has been: received into Gy- arong Manyak and Takpa. tn the south it is found in conjunction with an older broad form, equivalent to one ten. (20 being in many two ten, 30 three ten &c.). Burman she, Toung-thu tah-si (ta in 1), Sak si su (su 1), Namsang é-chi, (Mikir ¢-chi 1), Singpho si, Bodo ji, Abor u-ying, Chan se (also sone), Lepcha ka-ti (L ka-t, Gyarung ka-ti), Kasia shi pon, Lam- a Lous, Murmi chi wai, Lau tam. sip, Mikir kep, Kiranti kip, Chep. gyib. C Ancani here. —_—=—n3— OO OO - ETHNOLOGY OF THE INDO—PACIFIC ISLANDS. 121 Forms OF THE DUAL DEFINITIVE. Il. A. Amarchaic Chinese form is preserved in the Mok-kien no and ‘Tie-chin no, B. The next form evolyed appears to have been liang, leung, ni,— probably from an archaic form of no [e. g. ngok, nyok or niok, niak, niang, liang]. Niisthe colloquial Shanghai form, and appears to have been also the Kwan-hwa. C. The latest form has been evolved since Kwan-hwa converted ni ‘into rh, in Gyami ar. Some of the segregatives used to indicate pairs were probably ancient words for.two, Kwang-tung has tui and sheung, Kwan-hwa tu and shwang; Shanghai has song “a pair of shoes.” The o form.is not found in the Himalaie province in 2, save in the Ku- mi nhu, Sibsagor Miri ngo-ye and Singpho gutturalised nkhong. A cog- mate broad variety ia preserved in Thochu nga, Manyak na, Takpa and Bodo nai, and, less contracted, in the Burman wr. nhach, nhar’ (sp. nhit), Tengsa a-nat, Nogaung a—na. The Bhotian nyis, Gyarung nes, are slender varieties formed from a similar archaic Tibetan form [nhak or nyak, ngak, ngik, nyik, ngit, ngis; nis]. The Changlo ngik preserves an older stage. The sp. Bhotian nyi, Horpa nge, are not derivatives from the Kwan- hwa glossary, but loeal variations produced under the influence of the later Chinese phonology. The contraction of the Sifan broad forms is to ‘be explained in the same way. ‘ In the south, the prevalent slender forms nhit, nyet, ngi, ni, ne &e. are not, in general, Tibetan importations, but local assimilations to the later Chino-Tibetan forms, induced by the modern phonology. (See p. 17). The Mijhu and Garo ning is referable to nik. There is no example in the Himalaic province of the modern Kwan-hwa form. While the Chinese dialects use both the nasal and liquid (e. g. ni, liang) forms as distinct words for 2, it is remarkable that only one of the known Himalaic vocabularies uses the ee It is found in one of the least mo- dernised of the Yuma dialects, Mru, in the form p-re, the vowel being that of the nasal form in Toung-thu, Angami, Khari, Dhimul, Lepcha, Limbu, Gyarung and Horpa, This isan archaie Chinese vowel—leung Kwang-tuug,—aad as it is preserved in the 4 of Manyak, Bodo, Burman, Angami, Tengsa and Sunwar and inthe 8 of Kiranti, Murmi and Gu- rung, it was jobably at an early period current as 2 in Tibet. The Kasia arisa covnate broad variety, of which the full form is preserved iu p-rah 8. Tn 4.and 8 li and not niis the common form. -Itis clear therefore that li, ri, re, were used for 2 in the Tibetan system before it was carried south, and as then form, now almost universal for 2, has archaic fo which could hardly have been derived from China subsequent to li, it is robable that in Tibet also both forms were current at one period. The Fibial prefix indicat’s the wreat antiquity of the |, r, form. Itmust have beeu disused in 2 before the original of the prevaleat Southern systems was transported from ‘Tibet. The Lau sang, song, Changlo ching (ngik ching) appears, like nung J, “F207 ETMWNOLOGY OF TYE INDO-PACIFIC YstANDs. to be an archaie Chinése numeral. It is similar to some of the current Chinese words for pair. * | . IV. The oldest forms appear to be the Horpa dha, Kiranti Ta, Arung dai, Angami da, deh, which are probably examples of aréliwie Chi- nese forms early current with nha, or mia. The Chinese lang 2is a similar form. Archaic forms ure also preserved in those names for 5 which prepose the word for 4—lia, Takpa, alia, Dhimul, rai Bongju, Tei Mijhu (liin Abor.), See p.19. The‘Takpa and Dhimal lia, are evidently from liang. The Jrssage to the sibilant was probably through the sonant ‘forms j, z&ec. In Kwan-hwa the archaic sound ni becomes r andj (=zh). T hig change would convert the nin into zha Ser i and the current Chinese az, 8s areanalogons forms, si, ti being probably later. The 1, r forms are probably older than the sibilants. The interchange of ni and Ii is so eas and common that forms in li must have early been current in Chinn an with n forms. The current liang, leung show that the n was eonima With ] before the final consonant was lost. The Tibetan and Southern li, di, ri, le, deh &c. are of & later type than Yha, ra, but older than the sibilant. As the Horpa Tha adheres to the nor- mal form of the Thochu zha, so the earlier form of the Dhotian’ 2lryi, zhi is preserved in the Takpa li, and the Gyartng di is referable to a sitither form. It must have prevailed in Tibet when the numerals were carried suuthi. VY. The older Chinese ngo, go, pe are broader than the old forms of 2 fo, nd, and similar fo some archaic Hlimalaic wipe A nga, In Tibet these archaic forms are also current a8 5,—ngo, nga, nha. In the Bouth nya (Bhotian, Manyak, Thochu) is the most common form, The o, u vowel of Chinese and Gyarung is found in Khyeng ngau, Angami, Tengsa, Nogatng ngu, Abor, Dophla. and Sunwar ngo, Lepcha ngou. The slender vowel of Horpa gwe and Takpa lia-nge is not found in the south. — ; VII. In Tibet the riasu] form with the e vowel is found in the Horpa and Gyarung 7 asin 2. Both e and i are common in the South, The older broad form is found in Toung-thu nwot, Kumi sa-ru (as th pa-lu 4), Limbu na-sh. _ The a form occurs in Muthun a-nath, Abor ko-nang-e, Dophia ka-rag, Chepang cha-na—zho. VII, The Thochu ra of 4h-ra-re, Gyarung or of or-yet, and Horpa rhi of rhi-eé are examples of the three forms which the root has acquired in the liquid variety. The Bhotian dr of br-gyud appears to be merely a double prefix to gyud as in br-wya 100 where the guttural can only te the unit. (comp. r-hya 100 Horpa, par-ye LOU Gyarung). Wien thre form of 8 was produced, the 1, r form of 3 must have been current im Tibet, or at least in the dialect which originated such a form. The Manyak a (in 40 zyi) appears to be a sibilant form of an older ri or li similar to the Horpa rhi. Jn the South the oldest group preserves some fall forms Mru ri yat, Kuki ri-et, rik-t, Shindu eha-ri-a, Similar forms without the 10 crits remmant occur in Kiranti re-ya, Murmi, Gurung p-re (=p-re 2 of Mrn). * Possibly it is from liong. In the decaying Chinese phonology ni passesinto zandj. Thus the Shanghai colloquial niun mon is read zub, dud in iwang-tuig has becume fin; nich fish is zoh uvdjuh, . ETHNOLOGY OF THE =NDO—PACIFIC ISLANDS. 123: The only broad forms similar to Thochu are p-ra-p Chepang, p-rah ia. The only nasal forms aré those of Abor and Mikir, Sala a RESULTS. (1). Thearehaic broad form ngo, nd &e, is current th 2, atid 5 of Chinese and several of the Tibeto-Ultraindian dialects, but in ‘tiost eases with the a vowel, and in several with a final consoniut, k, t, s, ng. From ita great persistenty in 5 throughout the Chino-flimalsic province, it appears that it was'a eurrent form of 2 when 5, by throwiig off # word for 3, acquired the character of a distinct root, and was thus ekempt- éd from participation in the later changes which 2 suffered. (2). Liqnid forms of 2, both broad and slender, appear to have euty been evolved, and to have been current along with nasal ones. The u of two or more nantes for 2 is common to the Chino-Himalaic with other nuiheral systems, The Chinese liang, leune 2 is an ancient form. It was probably the most common name in the diuleot that first gave tlie numerals to Tibet, for it appears to be the parent of lia, li, Ja, Iu, 1, ra, ru &c., the form of 2 which must have fo enrrent When the exist- ing Tibeto-Ultraindian 4, 8 and 40 were produced. The ClHinese, like the Himalwie, sibilant 4appears to be referable to this form. The Chinese form of 4 then current must have beén liang, leuné, (? liang-liaug &e.)or a similar form which afterwards became sibilant, erkd-disenieed the character of an independent root, (3). The current Hinalsie slender nasal forms of 2 are local varii- tions of the archaic broad nasal form, The ancient ra, li cannot be dé- rived from ni, the most modern forin of ngak or nyak, The relation 6f the Himalaic ni to lia, li, Iu &c. appears to be similar to that of the Shanghai ni to liang. Both havebeen derived throngh different chat- nels from a primary naso-guttural root nguk, nyok &¢. Inthe Himit- laic province the 1, r, numeral appears to have fpllen into disuse in 2 in nearly all the dialects, the current 2 being in general the nasal. The 1, x form has lost its binaty character, and in most. of the dialects has concreted with the archaic labial pretix. The occurrence of the later prevalent 2, and not of the earlier p-li &e., in 7 may arise from 7 having been 6, 1 while li was used in 2, asin Chinese, Horpa, Manyak &c. That thie quinary 7 was not formed till p-li ke. was disused for 2, appears from its not only taking the later current furm of 2 but the later prefix, ee To complete this review of the elements that enter into 4 comparison of the Himalaic numerals and throw light on the history of their diffu- sion, it is necessary to adyert. to the various forms of the prefixes and postfixes. These are simply the archaic detinitives of the formation usell possessively or qualitively, a3 in the numeral systems of other fortnations that retain a possessive servile, ‘ The labial (animate, masculine, fem.) is well distinguished. The others t some difficulty. The most important is the guttural, indnimate m the Tibetan systern, As an archaic prefixed def. it his the variations gk; d,t; 4 8s, hj 7, J], 0. : In the secular progress of glossarial mutation, the definitives lave acquired variows specitic functions, aud different forms are now in dmany cases equivalent to distinet roots. ‘The dialectic variations are also considerable, From the coummututivu ol ull the cousuniuts save tlle ‘124 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. labial, it is hardly pinay to ascertain the number and forms of the primary Himalaic definitives. The history ofthe labial is clear. It still ‘retains its primitive and earlier secondary applications, human, male and female, animate &e. Its transfer to imammate objects was probabl wrimordial, because in primitive science all things are living and sexual, ‘The guttural was probably also animate and sexual, as in Chinese. In Anam it is fem. and inanimate, and in Bhotion inanimate. How far the other definitives are merely phonetic variations ot the guttural, and how far they are primitively distinct, it is difficult, if not impossible, to deter- mine with complete accuracy. In many cases they ure clearly referable to it. But the prevalence of the liquid la, ra, na, lu, ru, nu &c. &e,* as an animate, a masculine, and, to a less extent, as a fem. root, in the Chino-Himalaic vocabularies, makes it possible that the liquid def. is re- ferable to it in some cases, and not to k throught, d. in many of the names for man and the lower animals it is a sex definitive or qualitive ; and as the labial in all its forms ’was early ‘transferred to inanimate ob- jects, it is probable that the liquid was so ‘also. The sibilant appears to nave been a very ancient, os it is one of the most-extensively diffused, hu- man and sexval—generally feminine—forms of the definitive, und as it oecurs with its primitive substantive meaning in the Chino-Himalaic vo- cabularies, it may have early been applied to inanimate objects. The easy interchange of t, th with s, h, and of J, r with z, s, renders the his- tory of the sibilant as a distinct prefix very uncertain. But as the pas- sage of the ¢, k, d, t into the th, t, s, z and h prefix is free from doubt, and the sibilant is not common asa ‘substantive sexual root, it may be concluded that, in general, the sibilant prefix of the Himalaic glossaries is a secondary form derived from ‘the guttural'through the dental or the liquid. The common series of mutations'‘are g, k, kh; d, t, th; z, s, h, j, ch, y; t,'d, n, 1, 1, %, 85—1, or r, y. It is probable that in the earliest stage of the definitives the labial was applied to animate and the guttural with its variations to inanimate ob- jects. In thatstage the numeral and the qualitive probably-took the det. of the substantive as in Zimbian. But before even the oldest prefixes now extant in the glossaries conereted with the roots, they had either acquired an absolute use, or life and sex were attributed to inanimate objects. After the preposed definitives became prefixes, they tended to merge lin the ‘root ; but definitives being still current as separate particles, they ‘were used with words which had concreted prefixes. In different dialects ‘the common prefixes, or their forms and applications, varied. Thus while cone ‘used the labial chiefly or wholly, another used the guttural or one of its forms. The same dialect varied in different eras in this respect. A further source of diversity and irregularity in the prefixes of each dia- lect has been the mutual action of the dialects. Thus when one which atiected the labial prefix came under the influence of another which used the guttural, and adopted it, some words might have the old and some the new prefix; but the old would, in general, either be disused or concrete with the root. ‘Thus da-ri would become d-ri, and then ka-b-ri. In the same way the archaic k, t, s, |, r prefix might merge in the root, or be retuined as a prefix, while the later current detinitive labial or cuttural ‘was superadded ;—du-ru or d-ru would become ka d-ru,—ki-ni, ti-ni, si- ior s-ni, would become ka s-ni, ma s-ni &c. In some cases the concreted pre- ETHNOLOGY OF THE INDO—PACIFIC ISLANDS, 125 fixed consonant appenrs to be the terminal and not the initial of the defini- tive, definitives sometimes having final m, or r,n, ng. Thus the Thochu r of r-ma name may be fromtar-ma, Gyarung haying tir-ming. The Bho- tian r-na ear may be from tar-na, Gyarung having tir-ne. But the final r &c. may itself be the remnant of ra, ro &ce. as Mr. Hodgson believes. While in one dialect an archaic prefix has thus concreted with the root mon syllable, in another the full form is retained. This has been abundant- ly illustrated in the course of the miscellaneous glossarial comparisons, 1 shall only adduce one or two instances here. The 1,r root fur bone has the labial pref, in some of the urchaie lrawadi dialects (Manipurian, Sak), ang ka, kh, g, ta, te, sa, thu, se, in the majority of the Southern diuleets. In the Gurung nv-g-ri the guttural pref. is concreted and the nasal super- added, unless nu be a later Bhotian form of the root (ra). Garo has g-ring, Jili kham-rang, Maring ki-ru &c. Inthe Gyarung tri read, t is the current secondary form of the guttural preserved in the Thochu g-ri. In the Gyarung ti-d-ri, skin, ri isthe root and dan older torm of the dental refix, of which the archaic euttaral form is preserved in the Horpa g-la, ve sk gti. The Thochu ra-pi gives the pure root with the labial postf. So gha-da, star, Thochu,is archaic; g-ra Manyak, an archaic conereted form 5 s*7-re Horpa, a slender variety of the same archaic form with a superadded prefix, in its turn concreted. The Gyarung ¢s/-m gives this later pref.in a full form. The Bhotian d-ro- 9 hot is preserved ina more archaic form in the Takpag-ro-moe. Gyarung, in ku-s-man ripe, superadds its current pref. toan archaic coucreted form of the same pret, similar to the de- of the Manyak de-mi. Bhotian has the Gyarung a with its current qualitive postfix s-min-bo. Thochu has the pure root, min, In the south the archaic form of de-mi, s-min, s-man is preserved ; Bodo ga-mang, Tengsa ta-man, The Khari ¢e-nhing (for ming), Anga- mi ke-me, connect the dental Manyak form of the prefix with the euttural. The archaic labial is alse found in the south, pap-man Garo. The names of animals afford several examples of the concretion of archaic pretixes and the adoption of new ones, e. g¢. Monkey she-p-ri Gyarung, s-p-re-bu Bhot. (both the primary and secondary prefixes coficreted); Snake m-rui Takpa, s-)-rul Bhot., kha-b-ri Gyar.; Ant ba-ra Manyak, tu-kh-ra Tho- chu, s-kh-ro Horpa; Crow wa-lo Tengsa, a-lok Lepeha, a-b-lak Bhot. ta-b-rok Gyar,., ha-tha-rak Khoibu, ka-luk Serpa &c. The roots having now been identified in their various forms, and the rela- tive antiquity of these forms so far ascertained, the numerals are found to reflect some historical light. The oldest units were the labial and the guttural. The former was disused in China before any form of the numeral system spread into 'Ti- bet ; but its retention in the Chinese 8 and 100 shows that it was current when the full denary numeration was attained. The guttural was proba- bly current from the first along with the labial as a definitive and unit. (Sec. 3). At all events it early became the principal umt. It ap- pears to have been the only one current in the Chinese dialect that tirst * gave the numerals to the ‘Tibetan tribes, and even in the existing Chinese system the unit in 1, 3, 6,7, 9 and 10, is the guttural root under different forms. When the system was first carried into Tibet the pure guttural was current as 1 in China. It isnow only preserved, contracted, in 9; but some full forms remain in the Himalaic¢ province, not only in 9 but in 126 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 1 and other numerals. These [imalaice names are a distinct f of the archaic prevalence of the guttural unit in China, and of the Chinese sys- tem having been carried to Tibet Vefore the era when dental, sibilant and liqnid forms replaced the guttural.* This purely phonetic change produced the existing Chinese system, in its oldest form, It is based on two typical forms of the modified unit, (Ist) chyuk, chuk, chut, dak, chak, dak &e., whence chhoat, tsat &c. 7, chek, chit &e. 1, and luk, lak &e. 6; (2d) chum, chup, chap &e. whence chap, sip &e. 10, sam, san, sa é. The history of the changes in Chinese pho- nolozy, has made a great advance in the hands of Mr, Edkins, but the full elucidation of the development of the numeral system must await further pro- reas. The tinal consonant bas some dependence on the vowel, and especin|- y on the tone. Most words ending in -k have the Jabial vowels u, o (Ed- kins’ Grammar of Shanghai PP. 59, 60). An examination of the phonetic characters shows that au, a, 6, 41, prefer -k, while ai, fi, 6, 1 prefer -t; iti also taking —p (Edkins on Ancient Chinese Pronunciation, Tr. Chinw Branch R, As. Soe. Part IV p, 52). The vowel also influences the initial consonant, Thus in Shanghai k has a tendency to be pronounced before i like t, dy or dj. The reeular tinal consonants taken by words in the long tones are —n¢, —n,-m. The short orabrupttone does not admit of these, but takes the corresponding tinils -k (or g), -t and -p. Thé yassave of —m inte —n and —ng, and of —n and —ng into -t and -k, conse- quent on gradual changes in the vowel or its tone, would explain the exis- tence of the same word in different ages or dialects in such forms as kam, kap, kan, kane, kot, kak, kag, The changes in the initial consonant may have some influence on the vowel and final. In compounds the different ~ words exercise an dnfluence on each other, and the Chinese numerals above 2 were originally all compounds. In these compounds the same unit and dual occupied different pésitions with relation to each other. Thus in 1 the unit stood by itself; in 3 it followed the dual; in 5, it had the dual both before and behind it; as the tinal element in 6 it followed it ; in 7 it followed itself. As the last element in these compounds distinguished the number from the one immediately preceding, it would most readily yield to phonesic influences inducing a change of form; and when thus changed, it would be considered as the distinetive element, even before the preceding numeral ceased to be repeated. The Milehanang sum 3, s0- rum 13, tuk 6, so-rukh 16, may be examples of an euphonie change of the initial from s tot underthe influence of the consonant of a preceding element. Such changes are common in all harmonie formations; and Mr. Edkins has shewn that Chinese is much more plastic in its sounds than has hi- therto heen supposed. The investigation of this subject must be. the work of a Chinese scholar, and in a more advanced stage of the science of Chinese phonology, The direct change from the labial to the guttural final is well illustrated by many of the Himalaie glossaries. Numerous eurrent roots have had * I do not overlook the possibility of the system having originated in Tibet and been thence transferred to China, The full diseussion of the primary relations of the Chino-FHimalaic tribes and dinlects demands a section by itself. The community of the numeral roots in their older forms was probably as ancient. as that of the pronouns. But the deve- loped system appears to me to be Chinese. ETHNOLOGY OF THE INDO-PACIPIC ISLANDS. ‘7 both forms from-remote periods, but the most archaic, as we have found in many instances, is the labial, Thus the oldest forins of the liquid root fur white, air &c. are lum, Jom, rom, dum, dom &c., whence lung, lug, lang, lak, rhot, lut &e. For water, river &c. the forms nam, nak, rang, rak, rik, rit, ri &e.; for mountain lum, (Kasia) ram, nom, lung, dung, rong, nong, rok &e.; for Aand lap, lak, duk, chak, yak, let &c.; for boat tup, dok, dong &e.; for iron shom, sung, chur, chak, ehat &e—are all current. Archaic labial forms of the common Himalaie roots are most common in the older Southern vocabularies—the Mon-Anam. In one of the later, the circumstance of its having been reduced to writing, enables us to trace the recent progress from labial to guttural finals, In Burman tup fat of the era when the alphabet was acquired, is now pro- nounced tok; nhup mouth is now nhok; a-rup chho vg/y is now a-yok shu. The change of m to n, and that of k to t, are also common, The two types of the unit may have been coutemporaneous in the same Chinese dialect from a very remote period, and even in the guttural era. The present system may have resulted from gradual changes in the same diulect—the different forms of the uit in higher numbers, to sume extent representing the forms used as 1 in different ages. A very slight change, in the tone, vowel, final or initial, would suffice, nm a monosyllabic and richly* vocalised language, to raise the unit in a higher number to the rank ofa distinct vocable, and enable it to dis- pense with the other member of the compound. The current 1 being used as an article, and much more. frequently than higher numbers, - would be more liable to phonetic changes ; and the units of higher num- bers, when their genealogy was lost, would nut share in these changes, But it is more probable that the azency of more thun one dialect is to be recoynised in the different Chinese forms of the unit, as it so clearly is in the Himalaie province. The archaie existence of the guttural root with both dental and labial finals is rendered probable by comparing the Bhotian forms for 10 in 8, gvud, gyet, gve, keu, khya &e., with the 10 of Mijhu kyep, gvep (in 30), Chepang ¢yib, Kiranti kip, Lepcha tip (in 11, 12 &e.), Limbu gip (in 100), Mikir kep,—these labial forms with the Chinese and Lau sip, chap &c, 10 ;—the Tenaserim tsit, §, 10 and the com. Lau chit, chet. 7, with sip 10 ;—and finally the current Chinese chit, chek, cha’ &e. 1, ch’hit, ch’het, tsat, sit &c. 7, with ship, sip, chap &c. 10, and sam 3. The Mon—Anam double form lag, lak, lat and dap,—dam, lam, rom, narn,—also shows that both finals were current ata very remote period. The —-ng, —n, -l, tinal is less.common than —m and —t, —k, and in some cases it may be derived from —k. But there are instances in which it is certainly a variation of archaic —m forms, and it may often have been the imme- diate parent of those in -t, -k. Eor example the unit of the Bhotian 7, dun, further modified in the Changlo thar 1, is from an archaic dum, zum, preserved in the Chanelo zum 7. In 3, sui, sani chanyes to song, sang, son; in 10 it has the forms sum, song, tham, san; in 6 dong, rong, luk. The liquid form has the variations ram, rang, rak. The archaic k form with final—m is very rare, but the form kun, gun. ken &c, ig widely preserved in 1, 6, 9,10, 20 and 100, Mijhu has a variation in ngun 8, in which the initial k or gis also uasalised ; and the Bhotian gyud 8 may be from a similar form. As the labial form is not now found in 1 of Chinese or any of the Ilimalaiec ’ 123° ETHNOLOGY OF THB INDO-PACIFIO ISLANDS. | systems, it probably preceded the other form: as the principal eurrent 1, In Chinese it became tixed'in' 8, 10;and 1000, when the currrent 1 changed ‘to duk, tuk, chyuk, &c, or itowas received from another’ dialect, Of the later current Chinese forins, uk, lak 6—referable immediutely to duk, chuk—is'older than the current 1 and 7, and-itprobably therefore — lost its identity as the unit, and became fixed in 6, before chat se. beewtme the current}. 0 9 raxpetenye=000F QOL OF sait— preven) Of all the higher nambers 7 must have been’ the last ito ‘become’ a” simple concrete numeral) It must’ have remained “a compound, G6) 1, after'the names of all the other numbers above 2 had become independent of the currentunit. 9) | frre OL wig ds oftges As the initial consonant is the most-essential part ofa root, the princi- palphases of the whit ivy be distinguished with reference to it, as prima- ry and sécondary,—the former embracing both the older g-m, k-m, k—p forms, and the later-k-k, k-t, @* Miri ¥-t&rd [he-r6 TO: ae 2 py The Murmiglrk amb Gurung iti, ~* formerly referred immediately tothe Bhotian g-chik, wre of uncertain ave. © % Forins less ¢learly referableto that with final mivare common in bigher « numbersy!'The Tengsd ela wid Nogy ¢a-rw'10 isa contracted: forthwres oor tuintag the Him pref Por’ 10) Angomi ‘hus’ hw-ro, hue, eerds jreory 0 Mike hore (in! 12,/125" 18), ‘Natnsing tusk (in 20, 80 foi) Memneheston tp (conffi-she-+ruk' 6) ooThe Tetesa group hasan a form in’ Khari ta-rah, © ane Mawshanle stort in the Desh nego apg The’ ér-eroupy has’ or 10 int phla 8 p-lag nag (10, 2),.rangein ‘the Dophlil Oy tang ti tlie Abbot 9:komang-ho 41, 10) Fling or 10in’ 20 of) onetiditvlect ir-litioehé, vine in’ the others im 20and m 10. Taving- has logy » lo prot hie 20 atid vongin') (kon“yong’).* The Mijhusnuti'Z 6, Dp is Psifhila’ fornPofthewiit! Gayo has tung for score in 20wnd 100mg | bo-nga} scores ®)oa form similar te the Mon 6-kasvung. “Vhe wre Bur- matrhich oS? appears to correspond with the *Dophla ky Beth « are probably frome ram. Phe'sime form’ of the unit otéerssin’) 100~ of Karen ita fap Sinphe lnésa,) (10% 10),0 Kuki (racke k-latye Burman) Kamhi tray Angami beads here, DS ogsiung ro-k-ra'(10 9010), Rharirusk- rely corresponding with the dani; thoiyeKambojan reay:Yoe; Aham rant, |’ For 000: Angwthivhasderiekesr (10010). 0 oro ee) Boe bhopa: has Satie forlO m0; Khe phedani (score, ten), sandyphe- danky in 50, klie phesdane-stin (score, teny \three)) forms: prdbably-corres- +) ponding with the Abor rang, tang) 6 Ty i Dutmixtsjae + Thereure some other and rarer remnants of the liquid unit, That dia- lectof the: Bast'Gangetie group which retains the» strongest Mon-Anam element-Kasia—hasian arclinte and \peculiar combinntion of numeral - names, 1 and? ure Vindyanas in the: Mon-Anam- dialects: . But while. thes latter have alsov adopted the Vindyan 3, Kasiaretains a Himalaic unifin3, lai, which is evidently a variety of the liquid unit of the Mon- Anam tomily.cclt recurs: ine 9-kan-da, which’ I formerly considered trittuly butowhich is'more probably denary (1 from 10) like the other forms of the Chine-Himalaic %& | Chong preserves the some form, in 10, rai, and Lewin 100-voi. dn the Kasia 6 the unit has wlsooa:somewhat pecu- liam form: hin-raiy) ta-d) 10 may also be Mon-Anam, but. its -resem- blance to the Yuma Gangetic ka-t &e. makes this doubtful. The Kasia fort of SandO isretained asl inthe Play Karen lay(Maplu na), The archaieprevalence of a liquidunitin 3 explains the otherwise anomalous ©. liqnidanthe compound’ of Bongjnand:'Takpa. Inithe Bongju rai-nga- kar, rai agrees in form with the Kasia lai 3, and as nga is the Chino- = Himelaie 2): the names the foll archaic: compound «Byo28 da the Kuki ru-ipur/e, the 8 hrs the form of the unit'that is common in 6 (ru-Ae’): 4 while another dialvetlias ra-nga. Mijhuhas ka-lei d,'ng-run-si 50:(5,10). In the Takpa lia-nge, Jin would also appear-to be'the unit andnet the dug. The Aborupe-la-ngo-ko, pi-li-ngo-ko, despite the accordance of pi-li with (edie dwaed) clieate ‘ a qe Weber et Aad op kali, Uae td .\RmA BSB SMA R 6 eRe ° > ec t _ = -- 2 ee ea eo e.¢ rr a $ ETMNOLOGT OF THE INDO-PACIFIC ISLANDS 153° the eommon dual, must now be classed with the Tukpa, Kasia, Bongju and Kuki names, andthe li, la referred to. the liquid unit preserved in ling, rang, ying 10, and nang 9,—the slender form being also, as we have. eeen, that of the Ahom 1, ling. This identification of an archaic 1 in the disguise of a common form of 2 leads to another important correction. We have, seen that the liquid was one of the most ancient and widely diffused nnit forms in the South, anid that it, occurs with the archat labial prefix,—p-ram, p-rap, p-lag,,, » pr-la, pi-li&ke. In Tibet it is still current as 1 in, the Horpa ra... From the analogy of all the other Chino-Himalaic names for 100, :hose of Gya- rung and Horpa, parye, rhya, must be the unit; and their true elassitica~. tion would new appear to be with the secondary liquid form and not with, the primary guttural *, howeverstrongly the Bhotian dr-zya, the common softening of gya, zye to ya, ye, and Mr. Hodgson’s orthography (Par-ye, r Hya) may be considered to support my former analysis (par-ye, r-hya). That these names are pa-rye and rhva is contirmed by the Mikir pha-r, corresponding with the more common ta-ra, ta-ya of the adjae cent dialects. This recognition of a liquid: form. of the unit Tibetan, names for 100, necessitates the recoenition of an vbsolete liquid. form in 10, and this throws anew licht on the liquid forms found in 8. The Bhotian and Manyak 8 are clearly 10, and it may now be inferred that the Thochu #h-ra-re (ra 1 Horpa), the Horpa rhiee, and the Gyarung o-ryet (=rye 100) are also 10... This inference also involves the similar. southern names, p-rah Kasia, p-re Gurung, Murmi, re-ya Kiranti, ra‘ Milchanang ae 100), cha-ria Shindu, riet Kuki, riyat Mru, rbach Burm. &c.+ Taviny has the same amplified vowel but the —m_ tinal in its e-lvem 8 (comp. e-khing 6). It this conclusion. be the correct one, the only names for 8 in which 2 is preserved are the ancient Dophla p-lay nag. itself a strong illustration of the mode in which the forms of the unit sind ual approximate }), the Abor pi-nit and the Mikir nir-kep. | The Gyarung and Horpa’comn.unity of numeral forms found. in- 100 and 8, vecurs'also in 7, the Gyarung ku-sh-ves being au old Sitan tormy similar to the Horpa z-ne, with the eurrent guttural prefix saperadded, Both also retain a guttaral anit, Horpaii 10 and Gyarung inl. The Sifan dialect which had the form pa-ryet &e. in 10, 8 and 100, muy have.’ originated the similar southern forms found in the older Yuma. dialects, Mru, Shindu, Mikir &e. The Gyarang element in the general glossary of the older Kast Gangetic tongues. we have seen to be strong, The Mon-Anam dap, dam, ram and its derivatives belong to. an older moves ment, but amongst the contracted forms it is difficult to separate those of Mon-Anam from those that inay beof later Sifan origin, The Kast Gangetie * But these r forma may have been from the g¢ forms directly, and not through the dental or sibilant. G and a guttural r are phonetically close to each other, and the passage of gy into r and that of rinto g are coinmon. _ + The other lan sin which 10 occurs in 8 with the liquid forar are Changlo yen (for ren ), Dhimal ye, Limbr: yet?, Sunwar yoh?, Kami, Kumi ya, Kyau ruet, Shindu ria, Kuki rai, riet, get (g for r). tf. The Nogaung li-ri 40 (pha-li 4 Khari, ru, lu 10 Nog., Khari) is a similar example of an euphonic assimilation of slender forme, In the Khavi li-rah, 10 retuius its proper form (¢e-rah), 134° ETHNOLOGY OF (THE INDO-PACIFIC ISLANDS. liquid 10 appears to bea Mon-Anam remnant. Thus Khari has namin 50and ' ‘rah in 10; and¢the Burman fa-rai0-can hardly have a different origin from the Afiam ?-ram; or the Nogaungta-ru 10, ro-/-ru 100, drom the: Mon k-lom. The only forms thatire possibly liter Sitan, appear to be! those which’ have both the final tind the amplitied vowel which is 2 remrant of the rh-, ry-[Seh, gy }fortns of Horpa and Gyarang, ‘and: those winch are” celdarly derivatives from the former. The Takpa linet liange domay be ‘one of these'renmnants, as itaerees with the shindy ria ines, | He riya’ 100," hiée 8, but the Taying lyem renders this doalittal, | The Palepa livat + ‘Kha-li'20 may betrom a eorresponding obsolete bOs > But onthe whole the majority of ther. forins appear tobe those of an archaie Sif systeny which hid thenrinT, 3,5, 6, B, 10 nnd 100 as moditivations of the older deiital avd “diiliitfornis, the latter also remaining ewrrent in some muinbers or dia- Ydcts.” To this 'system the-remnants iw Mon-Aniam,\in amy ef the Abor- Yumaedivlects and in someof theleurrent Sifu areydm general) attri ba table, Ifthe Abor-Yuma liquid 10 belongs to the early Mon-Ariam era” Cand Not toa leter Sifuny the Abor- Yuma 8 musb be associated with it, and owitiethe remnants ini3yided&ely tint Tl carmel one ot bo anemone) ; i Te is not always’ clenr whether the vocalie forms are-contractions of ‘those in nv cor of those inky) ‘But it is certain | that \both were early current. While the former-agree with the: Chinese 3 and dO) >the duttar ‘ores! with the Chino+Himalme 6, Their full aveltie forms yaaky» lonk, Tiley ruk, mk, rak; lok, lat sresimilar’ to the wiit preserved inthe com, © ‘ChintoAimalsic 6 (1 for, 1), luky lak, lok, rok, (whence-ruy ri, ni, ne’ We.) | A‘ yiasalised form similar to nung, louny, liny,is preserved in the Gof Mon ke-rondy Chony tasdong and perhapsinationd dialect sarong, a distribution which proves’ its currency im the era when the Mon-Anem family possessed ‘the: Gangetic valley. Phe Ahom slender form ling is identical with the Wbor ling, ying of 10y 20/000. i ai 'Thediqnid forms extant in 1; 4), 6, °8, 9, 10, 20 and 100 of different diniects may be reenpitulated,-sdap, dam, rap, ram, nam, low, lyem, dong, rane, Toney Nang, rung, nung, nyong, nun, nul, run, loung, lon, ling, ying, rok, rak, ruaky rak, rik, Ink, lak, lag, lak, lat, dai, rai, lai, roi, Iei, rou, rha, ra, ya, Ta, Yo, V1, Te, r, la, lu, li, na, riyat, ret, ryet, rye, rhya, rhiee, ria, lia, xiv. While some of these are localand of uncertain aye, many are clearly referable to the archaic migrations of the province. ~ “The later secondary forms of the unit present considerable difficulty. Jn several dialects they cannot be referred to any of the extant older forma, and it is certain that there has been some transfer of these latest forms trom dislect to dialect. But the question how far such transfers ean be clearly traced is not easily answered, for the same phonetic series ‘of mutations has taken place in all the languages of the province, however = + = — _ * The direct, passage of the sibilant into the liquid appears to have characterised the old Tibetan phonology more Bik: than I-had ascer— tained when comparing the miscellaneous vocables, The minuter compari- sons [have since made in grouping the southern dialects have shown that several of the liquid roots are only variations of the sibilant, and that ex- . ceptional sibilant and dental forms which [ had donbtfully referred to the prevalent ig Ye roots are, in reality, remnants of the older phase of these roots. In Manyak, Namsang &c. s, z, ch, j have a strong tendency to be- come ror]. | ; GT et ah . ETHNOLOGY) OF THE INPO+PACIFIO. ISLANDS, 135° irregular its operation on the different: vocnbulavies, has, heen, in degree, in-extent and in the particular yacables, subjected, to it. In the Chinese numerals the older secondary sap Of 10, sam, ot 2, appears to be the immediate parent of the currént lL wnd-7 . in, thei older, form chok, chat, teat &e.,—thatis Lamb?) like 3. an 10, had the unit in the furin chap, sap &ev before it changed to clink, chat&e.. 4) _., _). ‘The common form: inG@ luk, lok, is diatinynished, by, its vowel from 1, 3, 7, and 10, and.its elder form duk, was probably from dup, dum, cor- responding with the Bhotian dun from dum 7,(sum3) and with numerous othee Himalnie forms ina, o.. Inthe older Tibetan system the form of the unitia 6, khum, chum &e:, appears to havecorresponded with thatin 1, 3,10 &e, The Goynruny tok, and the original of the sp. Dhotian thu.as preserved inv tuk of Milehsy Serpay dbimba and Kiranti,..audcdok of-Garoy must have been derived from OChinaat a inter period, and, when. dup—probably the reiinant of a distinet dinleet from thatin which « forms. preyailed—bad taken the forin dul in Chinese. Ly Syhas une Dro d The Gurung tu, Lhopa and Munni dha, Dhimal ta and Bodo do, are contractions of the same form. ‘Their diffusion in the, south, appears, from their distribation, to: be not, older thin the later bhotian migration. The sontherm Bhotian dialect: of thit period, must shaye agreed with the present spoken dialectof Lhasa in having adentalformis 6) 45 The we. Bhotinn has ia distinet form d-ruk agreeing with the eurrent Chinese, aud-toit the Manyak ¢-ru is referable. The common Gangetic 6 isulsoderived fromit. This form is certainly separated bya very slender phovetic boundary from duk, tak, echuk, suk, but as it is also the Chinese form, luk, aud its creat Southern diffusion attests. its antiquity aud persis- tency, there seemsno reason to doubt thatic was-the form used. bythe southern Tibetan dialect whielr originated the predominant. Gangetic svs— tem. The Gangetic vocabularies’ combme Sifan with Bhotian words, generally in older forms than the eurrent or even the written, Bhotian, so that G tay be referred to the Bhotian element in the parent sonth Tibesau dialect. The form d-ruk, is from du-ruk (comp. d-eu 8 with the ‘Tak— po dv-ru), and dwis.a enphonic secondary form of the eutmural (4a-, Ye, ta, da, leu-, gu-,tu- du-&e.),'The southern forms are 4-rnk,, k-ra, k-ro(Siuypho, Gare, Chepang, Takpa), da+rok, ta-ru, the-rok, elu-ru, so-ru, éu—ruke, | ) rt The preservation of other varieties of 6, in which the unit has older forms both primary and secondary, and the wide prevalence af the Chi- nese form, make it evident that the latter was carried westward by distinet Mig from those which gave primary aud, secondary —n forms to Tibet. The dissemination of the later forms of the Chinese 1 and 7 is much more doubtful. The Bhotian ehig, chik has certainly a close resemblance ta.the Chinese chit, but it may be from a nativechuk, chum. Possibly ehuk is Of the sime ave as tuk &e. 6, and was & Chinése form ofl iu ‘the dia- lect which gave tuk to Tibet. fl barat In the South there are no forms clearly referable to the Bhotian chig ehik, save the Kinawari aad Serpa chik, Limbu thit, Newar chhi, and Lhopa chi. | ; The common form of 1 and 10 in the latest diffusive Gangetic ¢ystem whs similar.to the older secondary Chinese forms ia’7 andl, bat 1 ap- peirs nevertlieless to have been local. Lt is well preserved in 8 of Toung- 136 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. thu, Khyeng, Sak, Tablung, Namsang, Singpho, Burman, Bodo, that, sat &c., and in later slender forms in several of the allied dialects. Simi- Jar forms occur in 10 and 1; but under the influence of the later phonolo- ¥ they have, in several dialects, become slender like the | of Chinese, Bhotian and Gyarung. Both broad and slender forms are sometimes found in the same dialect, Thus Boro retains jatin 4, but in 10 has ji, in 1 ehe. Garo has chet in 8, sha in 1, chi for LO in 11, 12 &«., both being combined in chi-sha 11. Namsang has é-sat 8, i-chi 10, van-the 1; Sing lio ma-tsat 8, si 10; Burman tach, tit, ta 1,shyit 8, she 10. The slender ie is evidently borrowed in some of the diulects. It is clear that the broad forms have not been derived from the current Bhotian chik. older Chinese form of 7, identical with some of them, ia not found in Tibet. It is probable therefure that, like these Chinese forms, they are directly referable to the native labial form tham, tsain, sam, sap é&c. The Tengsa group preserves sep in Tengsa, corresponding with chet in Khari, while Mikir retains a primary form kep in 8, 9 and 10, anda similar form is found in the Kiranti, Chepang and ha 10. The No- gung tang is an intermediate form between tam (a Himalaic form of the unit still current, as we have seen, in 10 and other numbers, both in Tibet and the South) and tach, the old Burman form. The com, tsat, sat, chat &ec. like the slender Burman tit, are but luter variations of tak, chak, This form is the distinctive one of the latest East Gangetic (Bodo-Singpho) band, and has been communicated by Burman to some of the Yuma dia- lects, as it is found in the 8 of Toung-thu that, Sak tseit, Khyeng sat, These dialecta have received numerous other Burman vovables. Whe- ther the common vocalic forms in 1 and 10 were contracted from —m, =—p, or from —k, —t, forme, is uncertain. Tie labial forms may have be- come contracted in Land 10 before the t forms were evolved in 8. For example sha 1 of Garo may not be from shat, sat, the current Bast Gan- getic form in 8, but from the older formsap. The same uncertainty attends the Tibetan vocalic forms in 1 ta, ti,ra, In the South the evidence is in favor of many ofthe vocalic forms being from sham, tam &e., through shang, tangy &c. In 100 Arung has chang and Kuki shang, shan; ia 20 Dophla has san, sang. The coznate dialects have sha, cba, tsa, tha, sa in 1, score or 100, (Novaung, Tabluny, Muluny, Joboka, Muthun, Numsanz, Singpho, Manipuri, Bodo, Garo, Dhimal,) and itiselear that they are re- ferable to the current forms in -nz, -n,—of which Nogaung, as we have seen, preserves an example in 1,--and these to the widely prevalent labial forma common to the Mon-Anem and Tibeto-Burman systems, Theu forms appear to be chiefly of ancient Bhotian origin—Changlo being an example ofa highly Bhotian systen—and thea furms wo be chiefly Mon- Anam, Sifan and Chinese. Notwithstanding the examples of the passace of guttural into dental forims im the Suuth and the possibility of this-having happened in many cases of which no evidence remains, I think there ean be no doubt that the most prevalent Himalaie forms of the unit are not local varintions of the guttural, but were derived trom the Chinese system utter seconda. ry furms had been evolved init, The prevalent forms of 3 preserved in tiat numeral in libeto—Burman and in the 5 of Mon—Anam, agree with the Chine-e 3 and must be reterred to it in its existing « form and in older t, d forms, It may also be held as certain, from the abundant rewuius of this typeof the uuit,—uotouly in 3 and 5, butiu 1, 6, 7, 8, ¥, 10, ETHNOLOGY OW GHE INDOPACIEICSLANDS = BSE 20 and)100,-+that when the secondary Chinese system wag transferred to ,, the Himualaic province, sm, s—p bad not Pecanen fixed ee aa LO, bué *. that, aii i-m, 87m fame were the, A i unit,” * + ma AH sions.at w we seem jus ving, ire thefore as follow tt "The . typical Me cof, Seunoe prevale a ett rman anu on-Anam Pn ims, dum, cane sara, sum om S¢ ib, nition qu unit of both families, had therefore the same, tani re source ap pears, to have been.the Chinese system in i ahi deny dary. form, t that is when 1and,?as well,as 3 and, 10 had the | he meets pine when — 3, Zyand.10 must have still been, con aie se isera may also have had the fat packarmgt pan “the We, “Gane ¢MonrAnam rete ie slim een esis ‘| ee etre pe: oe! Pe Ad From the Tater Farr while.it, was affined Anope alder, grou Bun Eien the J tendency,.to lyr, nfo Toh in, - ference , ws aL i modain = reall its farms. p ano oth ifr are, neh xed es the, pai a he nan r res...Wwi e, prope oe yrins, i uy th 1 the laws Wy 100 & ‘ea fannie ait se i a a i id * " tendency, in which, se origina arpa ewe HCA is shove. Lat or from &. common, Aaa cor si ansource. | Lb is.c u ar 4 aaa the- older Fant ie di ee ee na trom ¢ pat prqcatent fo ravirian aah BUN ra ut is al an in, some, of these diulects,—Manipuri ih val 10 (in Bae Se 9), one 9, the Khyeny, nest aan 1, 4 iu ue Ais Ak pha tt 4 (bomAnam),, Avgami,po 1, aint) We fin lung aud ‘a 10 :ban, + Lin mi .in_pi-ra, mi,2 nO ua spuibre m-phai-re 30, Wurparlu-re) 40 Tea re ‘tL) th Wwi-] ach) (pa 9) a— 5),.va in chum-wa-rd or waiee, L0\), ad op h a hte ‘ eee bong 10, bha-g-y4, 8 (bo-/c.Mon,,.vogtt ‘ bial yee zu 2. ee: ere Murmi ch’ ehi-wai- 10, Suawar s-waj-ka, 100, Ca i! An 4,1 in- tha-ful. 6, Car feun 4, ta—-fod %, a—wera, 8. Limbu and (Granti cor (10.for 1y.10)... oe | The priradem ottieed forine rebels in y differenti pwinberad in; eat t dialects,/and also passing directly into thesibilant, iidicateaneldentrans-(! fen’ of the’Chinese’ system to. the (westward than, thatwhichy produced», the Mon-Ananiand current Tibeto- Burnians. The'preservation wf khung 9 in & is itself strong proof ofa distinet migrationyprion to, the, erarwhen dental, sibilunt and liquid formsalone prevuiled. - Botham the Mon-Anam and. Tibeto- Burman dislects 6 is either the current Chinese form lak, dukpor its t immediate parent dik, tuk, which is probubly a derivative from dum, tums, through dungy tangy heiMonzAnam family. in its Gangetic tera hab » pr rung’ in 6 Sy elated with luk, and front the: wvesermmblance 4 eas wy x * ‘Tn many of the southern Ainlecta (Singpho-iBoda. gre &Q)) the si pie } bilant and dental of 1,8 and 10 is referable,to the, puttural khumy khip,, klting, khat, &c., but inotherstothe secondary Cline-Tibetan sun, SMC,» + For exanplethe old or written .Burmanethaghy, 8} ata, (1.00, is, evi-y., dently a distinct andan Burman-an elder form of the wnit,of (10 than,the current 16 chhe, corresponding with the eurrent-8 shyitand 1 tit., Doph= lathas the same-8 lag: and 10 rang, and ‘Barman, Thust atone. time. lueve : hatl a similar 10, Nawsing has the susie type in.d,.8, L0-and 10Q,as the CU's 4 rodit. Burthaii 1; Sand JO, -bwtdir 20 preserves cuales ou: jo 94 (eid tocuie nas ~ > 138 ETUNOLOGY OF THE INDO-—PACIFIC ISLANDS. of the former to current varieties of dam, ram) &c. an early Mon-Anam dislect may have had its prevalent-m unit in 6 as well as in other num- Ta. The Abor and Yuma dialects preserve many of the most archaic forms of the Ilimalsie vocables, A large number are leas contracted and soften- ed than in the other groups, both Tibetan and Southern. They have been Jess affected by the laver phonetic and glossarial modifications, whether spreading from Tibetan or from Southern sources. The guttural 1 and 6 of Abor &c., the 10 of the Mru 30, and the common guttural for se: re muy therefore be considered as remnants of the first Tibetan system that was carried to the South, and an evidence of the very vreat antiquity of the migration, How far this early form of the Chino-Himaluic system wus disseminated in the South cannot be ascertained. If it had been that of the first Himalaic tribes that became predominant in the Gangetie valley and in Ultraindia, it is probable that some distinct traces of it would have been left in the Mon-Auam dialects. Althoueh not found in the existing much mixed Mon-Anaim systems, it occurs in others which retain some Mon-Anam) numerals, Of the three hypotheses, Ist, that a Himalaic system having this anit preceded the Mon-Anam, 2d, that it was contemporaneous with it, one dialect or number having the euttural andun- other the dental and liquid unit, and 3d that it immediately succeeded it,— the 2nd is the most consistent with all the faets. The Mon-Anam voca- buluries are largely and closely connected with the old Tibeto-Burman of the South, but the Dravirian and the peculiar native traits of the former as well as their distribution, show them to have been earlier. The pre- sence of numerals of Mon-Anam formsin the older Tibetan systems of the South is in accordance with the general character of the vocabularies. But as the Tibetan glossary of the period when the Mon-Anam migration took place must have differed little from that of the first Tibetan tribes who followed them across the Himalayas, it is probable that the guttural unit was from the first coeval in the South with the sibilant and liquid. At present, however, we have no direct evidence that it was, the only native oi-Anam unit that has been preserved being the dental and liquid ; and itis therefore possible that while the more eastern parent of Mon- Anam had a secondary Chinese unit, a Tibetan dialect retained the ar- chai¢ guttural unit of the Chinese and afterwards carried it south. In some of the Sifan dialects the guttural unit appears to have been suc ceeded by the liquid, and the earlier dialects that were carried south probably ee both forms. The Gyarung—Horpa series still combines 4 in 1 of Gyarung and ga, ka in 10 of Horpa, with a liquid form in 1 of Horpa and in 8 and 100 of both. The older Abur-Yoma systems appear to be referable to the Gyarung-Horpa. The Dophla system with the gute tural in 1 and 6 and the liqmid in 8, 10 and 100, must be an Saaiaitta of the earliest Sifun systems thut were carried south. The differenve be- tween these and the cognate Mon—Anam appears to have been, that the Jatter bad entirely lost the guttural unit while in the Sifun system it was uscd ulong with the liquid. The most remarkable point in the history of the dual is the prevalence of a secondary form in 4 and ofa primary one in 2 and 7. The ques- tions that remain to be answered ure, when and how this difference arose,— weve the same forms ever current in ull the numerals of the dual series,— ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. 139 and how many versions of this series were carried tothe South? Ona cursory glance at a comparative table of the numerala, it might be inter- red that, in all the dialects, the common broad form in 2 and higher 1um- bers, ngat, nhat, nya, ngo &c. preceded the slender current form nyik, nhit, mi &c.; and that the latest and most contracted nasal form of 2 ni &e, waa the parent ofthe liof4. This was doubtless the phonetic order of the mutations, if li be from the nasal root. Butit by no means follows that because ni is now the current 2 in several dialects, it is the immediate parent of the li currentin the same dialects. The historical succession of the forms ofa rovtin a particular dialect, has no necessary dependence on the absolute philological succession. A slender form may be contemporaneous with a broad form in one group for thousands of years before it is evolved in another group; and it may be communicated, in a special application, by the former and re- ceived by the latter as a substantive vocable while the only native forins continue to be broad. Ina numeral series it is of course possible fora unit root or a dual root common to several numbers, to suffer phonetic changes in one number while it adheres tothe old torm in the others, And it might, at first sight, appear that the li of 4 was an instance of the kind,—ngok &c. having, by segregation and successive phases, attained that form in 4, but stopped atan older segrewated one in 5, while a still older remained current as2in 7. But many facts concur to show that li, lu originated in one dialect or group and that it was received as a substantive name for 4 into other dialects which retained or acquired older forms of 2. Amongst these we need only at present refer to the almost universal prevas lence of the 1, r form in 4, its retention of the labial pretix where 2 and 7 have the guttural, aud the traces of an archaic labial prefix in the unit series. The frequent passage in the Himalaic voeabnlaries of the sibilant roots of Chinese, Scythic, Chino-Scythic and Himalaic itself, into liquids, ap- ars to afford the true clue to the history of the numeral 4. The sibi- ant forms, I now think, in accordance with my first opinion *, must be regarded as the primary ones and the liqnid as the secondary. All the Chinese dialects preserve the sibilant. in Tibet the passage into the li- que is illustrated by the Thochu zha and Horpa Jha, The Southern arms appear, with afew exceptions, to be all referable to one dialect. The first great mivration must have brought the form pw-li, bu-li, ba-li&e. + which became all but universal in the South ; and its Tibetan type in the primary form is preserved in the Bhotian &-zhyi. The highly Bhotised Gyarung has the same form in 40, but liquid as in the South, p-li, As the identification of the Chino-Himalaic 4 with the com, dual of 2, 5 and 7 was founded on the hypothesis that the sibilant form was a modification of the liquid, 4 must now be considered as involving « separate root, for the primary form of the dual, ngok &e , cannot be derived trom si, zhyi, 2lia &c. As the liquid elements in 8, formerly considered to be 2 in the 4 form, have been found to he forms of the unit, itis not strictly necessary, tor the purposes of this Section, to pursue the enquiry why the Chiuo-—Hima- laic 4 has a distinct root from 2. Af si, ri, li &e. of 4 were really a distinet * App. to ch. vi“ Comparative Table of Chinese and Tibeto-U1- traindian numernix” pp, 28, 24. Bee also See. 4. + Oritsolder form bu-lu, trom a Bhotian bu-zhyu, t Unless the Lau song, sang be the same root, 140 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. root for 2, it must have become obsolete im that number when the system be- came that of Tibet], and no remnants of it are preserved in other numbers, The latter circumstance is hardly consistent with its being a primary dual root of the system. On the other hand if it be a unit it is iden- tical with the common sibilant and liquid forms of the Chino—Himalaic unit, and the Chinese system must have had a trina] basis throughout,—-that is 4 must have been 3, 1, in like manner as 5 was 3, 2, and 7 was 6,1. This mode of forming 4 1s rare, but examples of it occur in the N. E. Asian and African provinces ; and the Vindyan 4 is also trinal. As the Chino-Hima- Jaic sibilant 1 and 3 has forms precisely similar to 4, both in 3 and in other numbers, I do not hesitate to transter 4 from the dual to the unit series, When the first Himalaie dialect was carried south a broad 0, u form of the dual was common to 2, 5,7, and, probably, to 8; and it appears to have had the labial pretix throughout, as it preserves it in 5, and ina few diulects later forms have itin 2. The unit associated with this form of the dual must have been one of the oldest carried South, and as the later form was contemporaneous with the Tibetan gyud, ryat & it is probable that the ngo, nhu, dual was that of the system which had khung &e, os the unit, In the Abor group both khung and ngo (2) are preserved, When the next great migration took place the predominant Tibetan dia-— Ject had a numeral series which had the form nag &c. in 2 and inthe quina- ry 7,—associated with tam, sum, ram &e. in 1, 3,10, and in the denary 7; and the labial prefix had given place to the guttural. Dophla retains ha- nag in 7, nag in 8; and in 2 Burman wr, has nach, Tengsa a-nat (for ha- nat), Changlongik, Witha few exceptions the Southern Yuma-Gangetie forms are modifications of this form. In Tibet the broad vowel is retained in Thochu and Manyak nga, na and in the Horpa 2 of 20 na, In the other Tibetan dialects and in most of the Southern ones slender and contracted forms now prevail. The change has generally been local. The Horpa z-ne and Gyarung sh-nes of 7 are from a form of 2 similar to the Gyarung hi-nis in 20 and Bhotian g-nyis 2. Both are probably from ene dialect, perhaps Horpa, which has a similar form of the prefix in 9 s-ga. _1t has nge in 2 which may have been /i-nges, whence sk-nes, z-ne, The Gyarung s/i—nes of 7 is not the current ka-nes 2, but a concreted vocable, as it takes the current prefix /w-sh.ues. The luter southern forms have many varieties, and some similar to the Tibetan, but they are all of local growth. Thus the Bodo and Garo s-ni 7, although so close to the Horpa z—ne, is from si-ni preserved in the Mikir hi-ni 2 (Singpho si-nit 7); and s?-ni is from ia-ni, gi-ni Garo, Inthe Yuma 7 n becomes r, but the prefix identities it with these Gangetic forms. Comp. Tengsa ta-ni, Suk ¢ha-ni, Shindu xha-ri, Kami, Kuki sa-ri &c, The associated form nit Ke, shows that ni, ri belongs to the later ngak, ngik, ngit series in 2. The Himalaic form of 2 brought south by the Mon-Anam family appears to have been the broad form with the labial prefix pisces in the Tibeto-Ultraindian 5. In the Bongju, Kuki, Tapka and Abor 5 and in the Dophla 8, this form of 2 is conjoined with the liquid 1 (for 3 and. 10), whenee it is clear that when that form of 1, @ and 10 was current, the Chino-Himalaic form of 5, ngo, nga, was current as2. Ifthe Lau song, sang is from an | orn form it was probably one of the varieties of the Mon-Anam 2. ‘The current2 in all the other diulects ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, i4t of this fumily is the Dravirian labial, Although the traces of the original Mon-Anam 2 are not very strong, they are decided enouch to lead us to the conclusion that the dual had a form similar to the prevalent Himalaic 5, or, in other words, that this was the form -enrrent in south Tibet at the era of the first great Himalaic migra. tion, Although the labial prefix was that of 2, 4nd 5, and the unit also had it, the Mon-Anam 6 and 100 (k-lom, ¢-ram) show that a unit having the guttural, passing into the dental and sibilant, prefix was early current. ‘The Bhotian d-yu, Thochu r-gu 9, is another form of the unit with the same archaic secondary form of this pretix. The later slender forms, both of the prefixes and roots, connect many of the southern systems, although the movements and special borrowi: disturb the agreement thus induced. The progress of asjirate, slender and contracted forms of the guttural prefix is illustrated by the dual in 2 of Garo gi-ni, Mikir hé-ni, 7 of Singpho s7-nit, Nams. ¢-ngit, Kasia hi-nian, Garo and Bodo s-ni. This group or movement centres in the western extremity of the Garo-Singpho band. Another well marked rroup, Which breaks through this band or has been broken through by it, js distinguished by similar changes in the broad forms of the pretix. In 2 Angami has #a-ue, the adjacent Tengsa group a-nat, (Gyarung ka-nes). In 7 the Yuma dialects have tia-, sha-, sa-, the Tengsa gr. tha-, ta-, Chepangy and Sunwar eha-. The original #a- is preserved in the broad Dophia ka-nag. In 9 the Yuma group has tha-, ta-, Chepang ta-, Shindu chu-, Singpho tse, Garo sk-, Bodo ch-. In 6 the Yuma er., Ten gr. and Lepcha have ta-, Shindu ehu-, Ang. so-, Mikir tho-, while Singha: Garo, Takpa and Chepang retain k—- and Mon ka-, The unit 7 (6, 1 or 1) is found in Chinese and Lau; in Thoehu Manyak and Bhotian; in Changlo, Lepcha, Milehanang, Mijha and Mikir. The dual form (5, 2, or 2) is found in Tibet in Horpa and Gyarung (2), and in nearly all the Southern dialects. It was the form of” the dialect that gave the Prevalent numeral series to the South, and it corresponds with the 2 of that dialect, thus proving that 7 was still qui- nary or dual at the period of the creat Tibetan mivration, or immediately before it. No example of the full form is preserved. The quinary Kam- bojan p-ram pil (5, 2) belongs to the earlier Draviro—Himalaic systena of the South. The principal inferences bearing on the historical relation of the Hima- laic to the Chinese numerals are as follows. Ist. The earliest Chinese dia- lect that gave numerals to Tibet had the primary guttural form of the unit either alone or with later torms. It is still current in 10 of Horpa, ina 1 of Gyarung, in 7 of Manyak and in 1, 6, 7,10, 2U of some southern dialects, ‘here is no direct or conelusive evidence that it was ever cur- rent in 8 or 5, either in Tibet or the South; although the close resemblance of some of its forms in 10 to the Chinese sibilant 10—and of the latter, and similar current cis-Himalayan sibilant forms, to the Chinese 3— make it probable that the earliest Chinese system of Tibet h d gut- tural forms throughout. The 8 and 100 must have had similar forms of the unit, and not the lubial of the current Chinese. The «tual in this system ~ must have had the form ngong, ngang or ngok ngak &e. 2d. The later Chinese phase, which was communicated to Tibet betore the great mi-« gration to the South, is very distinctly marked by the furms of 3, 4 and 6, 142 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, Each of these isa peculiar modification of the primary form ; and the deviae tious from it and trom each other are so considerable, that they must be- long to an age long subsequent to that of the exclusive currency of gut- tural forms and when the system had become partly concreted. Whether these forms originated in one dialect or in several, their combination in a single current system cannot belong to the earlier eras of the formation. The form in 3, sam, sum, &e. must have long been the unit exclusively cur- rent in a Chinese system, The form in 4 isalater form, as it wants the final consonant ; the initial, however, having a stronger aspirate. The current form was probably preceded by one closer in the vowel to 3, like that preserved in Mhochu, Horps and some Southern dialects. The still later liquid 6 may have originated in a western Chinese dialect which had a similar form, }um, lam, lung, lang, luk, lak, in ita current unit, and was associated with those ancient Sifan or East Himalaic dialects which pos- sessed a similar unit and gave it to the South, but it is more consistent with the other forms of the system torefer it directly tosum,suk &c, What- ever may be the history of the production of the Chinese numeral system which ultimately had these forms fixed in its 3, 4 and 6, it is clear that, after they were sotixed, it became the most influential eystem first in China and Tibet, and then, through the great Tibetan migration, in the South also. The prevalent 4and 6 a pear to have always been concreted and substantial names in the Himalaic province, the connection between them and the unit having been lost before they were received from Chinese. The current form of 3, on the other hand, appears to have retained its unit power, after this late Chinese phase became that of the Himalaic province also, as it undoubtedly remained current with thia power in 7 a) 10, and pro- bably inland 8also, The slender form of the dual current in the Ohi - nese 2 in a contracted form is so widely prevalent in the Himalaic provines in 2 and 7, that it must be associated, in the form nging, ngik &e., with the later Chino-Himalaic form of 3, 4 and 6. The vocalic ut broad 2 of 5—which is almost universal—obviously belongs to the same phase, The current Chinese 1 and 7 appear to be later in form. The labial 8 and 100 were probably not possessed by the western Chinese dialects, until after the period when they gave the secondary forms of the unit to ‘Tibet. They have no connection with the other Chinese numerals and no representatives in the purer Himalaic systems, It is probable there- fore that this unit was preserved in a northern Chinese dialect—perhape the Kwan-hwa itself—which became that of the predominant Chinese nation. The Himalaic systems present many examples of a similar per- sistency of a native or older form. Thus the Kwan-hwa system, in 3 late form and embracing this labial 8, has been widely spread over Ul- traindia by the conquering Lau tribes, but the 1 and 2 are wanting, in the Lau dialect, native names taking their place. The northern dialects hate the Chinese name in 100, but it has not established itself in the southern, os * As the forms of the same root varied from era to era in different numbers and in different dialects, and as these forms did not attain the character of independent names at the same period, any attempt to re= produce the system as a whole in its successive phases would be liable to the risk of assocint.ng some forms that may never have been current together ETHNOLOGY OF THE INDO—PACIFIC ISLANDS... 145 Further light wil] be thrown on the history of the Chino~Himalaic numerals when we compare them systematically with the Scythic, Caucasinn and African. But our examination of the Chino-Himalai Semitic and African must be followed up by that of the Scythic an Caucasian, before we can enter on such a comparison, All the nu- meral systems of the Old World are more or lessconnected, and point to the diffusion of its earlier arts and civilisation, or of the races themselves, from one family. For example both the primary and secondary forms of the Chino-Himalaie unit are found in the other East Asiatic and the connected Western systems. The most important of the secondary forms sam, sum, sap é&c. is so widely diffused that, on am examination of a few numerals, it might be thouwht probable that it was evolved from kam, kap, ata very remote period and before any extensive dispersion of the ancient Asiatic system. Both the primary and second- ary forms certainly appear to have been circulated toyether over other provinces besides the Himalaic. But a comparisou of all the Old World systems leads to the inference that the connection of the Chino-Himalaic with the Sevthic and more distant numerals is through the primary guttural forms, and that the recurrence of identical secondary forms ani types in different provinces is, in general, attributable to the same cycle of phonetic change having been independently repeated in each family of language. There are doubtless examples of a transfer of later forms from one family to another, But it requires strong evidence to establish in a single dialect. If we were to assume that, at one time and while the naines still remained compound, a dialect existed in which the primary form of the unit had been lost and sam, sap was its only current form, the trinal system might be thus presented. 1st series. a sam nga lil 2,1 nga-sam 2d series. IV ,3,1 nga.sam-sam -) V,3, 2 nea.sam-nra Vi, 3, 3 (or 6,1) ng'a,sam-nga.sam 3d series. VII 6,1 Nga.sam-nga.sam —sam In all the compounds of the 2d and 3d series, the last word would be the distinctive aud permanent one. The office of the first common term of each series would be merely to mark the series, and when w slight phonetic change was induced in the last, this would itself’ suffice to mark the series, and the first term would become a needless incumbranee. Thus if, from the action of the preceding elements of sound, or by acqui- - gition trom another dialeet, the name of 1 became sang or sak, the sam of 8 would be enabled to dispense with the nga— not only in 3 but through- out the higher series of which it forms the radix, So, if the distinctive or final sain of 4 bocame sum, su or si, the initial nga-saim, or its remnant sam, might be rejected. In like manner achange of nga 2 to ngi would enable the distiuctive nya of 5 to reject the prepused term or its remnant. 144 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. the foreien orizin ofa form that can be deduced from a native archaic type by the ordinary phonetic evolutions of the tamily. he most widely prevalent of the older secondary forms of the guttural unit is s-m, s-b. Taking the Chino-Himalaic systems in their existing forms, it is clear that this has been the most prolific type of the unit, It wus the current unit of the era when the prescut 3 became concreted, 1: is preserved in several dialects in 10, 8 and 5 and several common forms of the current 1, 8, 10 and the unit 7 (6, 1) are referable to it. It is the integral form of the latest Chinese unit, being preserved in 3 and 10, and in a modified form in 1 and7. If any direct connection between the historical Chinese system and other systems of the Old World can be established, it must be through this form. In | the Scythi¢ systems have either the more archaic labial unit of Draviro-Australian—preserved in Chinese in 8 and 100, but obsolete as a unit when the system spread over the Himalaic province—or other forms of the guttural, dental and sibilant. The most com. Scythie 3 has a primary guttural form or a modification of it, kol, gur, kuj, chud &c.; tony, dong &c. The labial and the liquid unit are also found in the 8 of some groups. Beyond the proper Sey- thic limits—or in those of the earlier or proto—Scythic movements—the Chinese form is found in Caucasian, sa.mi, su.mi, ju.mi &c., and also in Semito-African, but preserved in Egyptian only sho.m, sham &c,, in which it is referable to a native Semito—African kho.m &c. The Chinese 5, as we have seen, is a remnant of 3, 2, and the term when the present form of the system retained its full integrity, was same nyo &c. The 6 of the Mon-Anam dialects retains the 3 in the forms san,. chang, ram (for sam). The Samoide sam, sum, sab, saba, sobo &c. 6 has the same form with the postfix —rigo, -fak &c.; but from the Tungusian tong, sun &e. and the Mongolian ta-bun, tha-ba &e. in which -bun, -ba is the common labial possessive postf., it is probable that sobo, saba &e. is identical with the Manatee form and that is radically so-ba, sa-/a. Cau- casian has the same form of the root chu-thi and in Abkhassian the same postfix chu-ba. The s-m, s—n form is very commonin the Afriean pro- vince, but as it is k-m in Semitic and there are examples of the passage of k— into s— forms, the latter appear to be historically connected, not with the secondary Chinese form, but with the primary Chino-Seythic k-m, k-n. The Ugrian and Turkish families have the archaie labial anit in 5. The Scythic forms of 6 appear to be all, or nearly all, quinary (5, 1 or 1) and not trinal as some philologers maintain. In this respect they accord with the Chino-Himalaic. With some of the Himaluic tongues, the Semitic, African and Ludv-European they have, the older sibilant and dental torm of the unit, and not the later liquid of Chinese and most of the Himalaic systems. Jt is not here intended to contest the triual character of the Indo-European and Semitic 6, although the fact of 3 heing itself the anit must render it uncertatn—when other evidence does nut exist—whether a unit form of 6 be aremnaut of 5, 1 or of J, 3. Ju African systems both forms occur. The Usrian and Turkish 7 has the later form of the sibilant unit as jn Chinese, site, sis, sat &e. Some Ugrian dialects that appear to have the older form shi-m, ta-b, sa-b are contractions of siz-ww &e. This torui is uso indo-Huropean sa-p aud Semuitio-Atiican sa ba, and in boty ETHNOLOGY OF THE INDO~PACIFIC ISLANDS.- 145 families the labial sper to be radically postfixual as in Scythic. The Seythic 8 and 9 are mostly denary like the Chino-Himalaic, but some quinary names are also current,—Koriak, Kamschatkan, some of the Yeuniseian, A form of 10 similar to the Chino-Himalaic sam &c. is found in a Samoiede 9 tu-ma, thun, Zen has the archaic labial unit; the primary guttural k-m as in some of the Himalaic names ; and later second- ary forms similar to the Chino-Himilaic t-s, l-k &e. But the connection indicated is through the prrmary forms, The Japanese, Koriak and Yuka~ te a for 9 preserve a guttural unit like the Chino-Himalaic (kup, chun, On the whole we may conclude that the Seythic and other Aso- African numeral systems (excluding the Draviro—Australian) are more elosely connected with each other than with the Chino-Himalaic; and are ouly connected with it through the older primary forms,—the Himalaic branch preserving examples of these similar to the Seythic &c. although lost in China, The general conclusions at which we have arrived are these. The first Chinese dialect that gave numerals to Tibet preserved the guttural unit and abroad dual. The Tibetan system spread to the South while it re- tained this archaic form. The next Chinese dialect, or phase of the nu— merals, that influenced the Tibetan had the secondary unit sam, sum, sap &e. in 1, 3,7 and 10. The Tibetan system which originated in it was the parent of the prevalent Mon-Anam and Tibcto-Burman systems, Two archaic movements to the south at different periods after this form was acquired, are traceable. The first had the labial prefix,—the unit had a decided tendency to pass into the liquid form,—and the dual appears to have had broad forms. When the older Sifan-Gangetie tribes followed the Mon-—Anam, if indeed they cun be separated, siiilay forms of the unit and dual were current in East 'Tibet. | the later and greatest Tibetan movement the guttural prefix prevailed,—the unit had broad dental and sibilant forms,—and the dual had broad, passing into slender, nasal forms. Historically and more exactly stated, these inferences stand thus, When the Chiuese system was received in Tibet seeondary or contracted primary numerals had ulready beco.ne fixed and concreted, in their present forms nearly, in 4 and 9 ofthe umt series and in 5 ofthe dual, The enquiry into the phonetic changes and dialectic intermixtures which resulted in the establishment of these names, belongs to Chinese and not to Himalaie phonology and ethnology. But the current unit in 1, 3, 6, 7 and 10 (as well as in8 and 100) and the dual in 2 and 8 were not so tixed. The numeral dialect that first took root in Tibet preserved a guttural unit, of which undoubted primary remnants are found in 1, 6, 7, 8, 10, 20 and 100, Some of the secondary Himalnic forms are also referable toit, The Himalaic forms gaum, khung (for khum), gyeb, kip, kep (whence kyok, gyud) are the originals, with modified vowels, of the ‘hinese chap, shap &c. 10, and the first Chinese dialect of Tibet must have had similar forms not only ini, 7 and 10, butin 6 and 8 alse, This dialact had probably the secondary form sum, sam in 3, as it is equal- ly universal with the formsin 4,9 and 5. It may have heen fixed in this dialect but that from which it was derived must have preserved it as the current unit in 1, 3,7, 8, 10 &e. and simultaneously or subsequently com- Maunicated it to Libet, Luthis dialect it appears to have usswued the form 146 ETHNOLOGY OF THE INDO~PACIFIC ISLANDS, suk, chuk, duk, luk &e. in 6 before it was carried to Tibet, as there is ne fT mauant of the s-m, s-p form in that numeral in any of the Himalaie dia- lects To the era when this phase of the unit prevailed a large number of the current Him«laic forms closely adhere, from which it is certain that during this era a Tibetan system was carrid south. Among the best ex- amples extant are the Bhoto-Changlo zum, dunof7, Kukisum, Suk su, Bhotian chu, Thochu du of 10. In several dislects, Sifan and Southern, the sibilant changed to r. A dialect which had this form was very influential at one period, In the south the typical forms are lum, lam. ‘The latter is widely current in various forms, one ef the most common being rai, The antiquity of this contracted form appears from its being found in 1 of Play Karen, 8 of Kasia, 5 (for 3) of Bongju (rai) and Mijhu (lei), 8 of Milchanang, and Bongju, 10 of Chong, and 100 of Kambojan (roe), Siamese (roi, hoi), and Boneju, The frequent occurrence of the r form in 8, or ia 8 and 100, where 10 and 1 have the sibilant form, must arise either (1st) from the same dialect having at one time used both torms in LO, or in 1 and 10, or (2nd) from having replaced the liquid of 10 by the sibilant ‘derived either from its own 1 or from the 10 of another dialect. Different forma of the unit have been used both synchronously and successively in several of the dialects for 1 or 10. Several express 10 by one ten, and in such u compound an archaic unit will be preserved in ten and the ‘eurrent unit in one, Thus in the Taying ha-long 10, lone corresponds with the 10 of 8 and 100, while ha or sa is a com, current unit in LO and 1 of Yuma-Gangetie dialects. The liquid appears to have been the 1 of the archaic Abor—Yuma and Mon-Anam groups. In muny dialects it has been replaced in 1 and 10 by the sibilant that characterised the later diffusive system of the South. In others again the sibilant—whether from the ‘Tibetan sum, or fromthe Tibetan gyum, khum, kap éc.—has Deen replaced by the liquid or by the Dravirian labial. The numerous southern systems are reducible, in their Himalaie element, and both in root and prefix, to two great and one or two minor migrations from Tibet, not differing very widely in the forme brought by them,—to a few special lines of movement in the South, —and to some phonetic changes that have taken place since the migra- tions, and of which the centres and lines of dispersion are more or less dis- tinctly marked. Precisely the same movements and changes are indicated by the distribution of the miscellaneous wovables that have been analysed and compared in the preceding pages; but be- fore their evidence ean be fully understood and historically marshalled, we must specially examine the Mon—Anam family, and the effects ou all the southern Himalaie vocabularies of the early contact of the Mon- Anam dialects with the Dravirian of the Gangetic basin. Tho preceding examination of the Himalaic numerals is far from suf Gcient to exphun their history in the separate groups and languages. in each of these the names have been more or less affected by successive phonetic changes, iuternal and external, and the consideration of each y itself willthrow further lighton the ethaie movements of the province. In some of the dialects alinost every numeral belongs to a difierent age _or phonetic phase, so thut the series basa simuar character Ww that of a ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 147 aseyrs section exhibiting a succession of unconformable strata, some ived from the waste of subjacent ones and others from distinct sources. The annexed table of all the published numerala from 1 to 10 of the ancient Chino—indian province, will greatly facilitate my readers in their comparisons.* I regret that I did not commence, instead of finishing, my own by compiling it, as the means whivh it gives of noting ata lance all the variations of each number, and all those of each root, would Save suved me much labour and not a few mistakes. I have included the Dravirian, as the earlier Himalaic—the Mon- Anam—combine Chinese with Dravirian roots. Savara has several pe- euliar numerals and the analysis of the table requires some explanations. 1. The root bo ovcurs also in bo-kodi 20 (one score). 2. The guttural sttix shows this dialect to be very archaic. 3. Ya-gi, this appears to be 2 (for 2, 1), the other dialects, 8. Drav. and Vind., having 1 (tor 2, 1), 6 ku-d-ru. This evidently compound term appears to be an archaic f name, and as 6 is 1 in the 8. Dravirian dialects, it is probably 5,1; d-re corresponds with the Kol tu-ru, tu-r which I have hitherto rewarded aa Himalaic. It is to be hoped that there are still some unpublished Vin— dyan systems which will throw further light on its origin. Meantime ‘there are some indications of Dravirian aflinity which have induced me, although with much hesitation, to analyse it in the Table as a Dravirian compound. ku is the com, Drav. postf., but tu is not the a, oof the 8. Dray. root (for ya, vo 1). It may however be the to for 1 in the ‘Telugu and Chentsu 9. Ku must be from an element in a name for 5, Dr. Stevenson in the Toda vocabulary furnished to him by Mr. Greiner gives khu 5. This is the guttural unit preserved in the Kol and Savara 10, and in S may either have represented 3 (3,2) or 1 (4, 1), 7 gu-/-ji, ‘The gu must here also represent 5 and not 2, 8 ta-mu-jt; mu-ji is the unit for 10 in the form occurring in the Tuluva mui 3 (Brahui mu-sé); ta is not a Drav. root for 2 and is probably a contraction of a form like e-ta (Toda, Mal.) ; the Telugu e.ni-mi-di isa similar compound, Norte. In lately issuing separately ch. v and the first5 sections of ch. vi, I pre- faced it by some explanations, which I subjoin here also, The remarks on the Dravirian numerals will be found to elucidate the analysis of the able, . “The readers who have accompanied me from the first do not need to be reminded that the publication of this work has exteuded over several years; * In writing out the names for the Tuble I overlooked the Rakhoin kh-rauk 6, and the correct avalysis, which I had given in the text, of tue Barman khyauk, khyok and the derivative Sak khyouk as kh-yauk, kh-vok, kh-youk. | The elliptic aud vary composite Deoria Chutia numerals having also been omitted [ give them here. 1 du-g-shu (Garo). 4 du-gi-ehi ~(Ahom T du-gu-ehi (Ahom 2 du-hu-ni (Garo). 5 di-gi-mu-a (Vindyan —). 8 du-gu-che Garo _ 3 du-g-da Garo), Gda-gu-chu (arch, Iim.). 9 du-qu-chu-ba tied jup . 10 du-gu-chu-ba du-g-she 148 ETHNOLOGY OF THE INDO—PACIFIC ISLANDS, and that the inereasing lizht thrown on the comparative characters of each family of language, during the revision of the section relating to it, has disclosed defects in the preceding ones. The continuity of publica. tion and equality of treatment, originally intended, have been prevented by frequent and, at times, prolonged breaks in the attention I have been able to vive to the subject, and, in some degree, by absence from the place of printing. The consistency which the work had when first rapidly written, as a statement of the opinions to which [ had been led by a review of the other linguistic groups with reference to the Oveanic, has been lost by the lapse of six years, during which ethnology has not stood still, while I have been endeavouring to bring these opinions to the test of a more searching Se into the peculiarities of the different groups. A final revision, on the completion of the work, can alone restore its uniformity, by bringing all its facts and inferences into harmony with the knowledge of the time at which it publication may be concluded, It seems necessary, however, on the separate issue of the present portion, nearly two years after its earlier pages appeared, to warn the reader that some of its glos- aarial details are at variance with the more accurate acquaintance with the Himalaic and Dravirian roots which I haye obtained from the minute comparisons in chap. vi. These errors will be best understood by a re- ference to that chapter, and especially to the comparative table of Dravirian an Himalaie roots which will be found in it. Some of the mostimpor- tant will be here noted, in addition to errors of the press and of haste. “In some places I have used the word Himalaic in a large sense, and as the paragraph explanatory of it was omitted in the proper place, it is necessary to mention here that, for want of a better term, I have applied it to that large group of cognate languayes and tribes which have imme- morially clustered in and around the Himalaya and the ranges subordimate to it, and the preservation of the native character of which must be chiefly gassed to the protection afforded by these mountains against the more powerful aud civilised races of Kastern Asia,—Cninese, Seythic, Dravirian and Arian, An extract from a letter to Mr. Hodgson (July 15 1856) will illustrate the application of the name. “That my Mon-Anam group was the Bengali of the pre-Tibetan era (using Tibetan for the present Scythoid branch) and conterminous with the Vindyan Dravirian dialevts is demonstrated ; but [ am not prepared to admit that Dravirian has not a distinct archnic ingredient, not derived either from the Mon- Anam or the Tibeto-Burman branch of what I have termed “ Himalaic” till you can supply us with a more appropriate name. I conceive the Driviro-Australian branch of Seythic or rather of Chino—Scythic, to be of rast antiquity, and to have long preceded the descent of the Chino« Tibetan race trom their trans-Himalayan abodes. Its strong Seytho- Caucasian element appears to me to show that it came round the western extemity of the great dividing barrier between middle and southern Asia, The Mon-Anam or “ast Himalaic stem was more Chinese and less Scythic than the later West Himalaie or Tibeto- Burman. All the earlier dispersed languages—that is, their mixed aud sometimes hybrid descendants—have a core of primary rovts, retaining a close resemblance to each other, and to those of the vo- cabularies that haye remainedin and near the primary abode of the Mid- Asiutie tribes, In this way I would explain the peculiar Chinese element of Himalaic, Caucasian ( preserved by the mountains), aud Draviro-Auev ~ ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 149 tralian, and the secondary Himalaic element of Caucasian, Draviro-Aus- tralian and other languages. The East Himalaic tribes probably occupi- ed much of what is now eastern Tibet and western China; and thouvh the precise line of their first southern migrations can hardly be traced with certainty, it is most consistent with the general character of the Mon-Anam ide to infer that they first descended into the Brahma- utra basin by the routes afterwards followed by the cognate Tibeto- Seah tribes, and thence spread over the Gangetic valley, mixing with the prior Drayirians, and, in the course of ages, eliminating the Dravi- rian physical element, though retaining Dravirian pronouns, numerals &e. Ofcourse there may have been other more eastern migrations, but the Mon-Anam branch, which predominated and spread everywhere in Ultraindia prior to the Tibeto-Burman, had its primary southern home and nursery in Bengal or the Bhramaputra-Gangetic valley, for its basis of Dravirian, and of a secondary or eorrupt dialect of Dravirian, could have been obtained nowhere else. ” The name is convenient in distinguishing the various elements of Asone-= sian ethnology, The latest of the three formations of the Indian province has appropriated its only general name, which is radically Himalaic.* This ies rendered it necessary to adopt a second name for that formation which would otherwise have had the first claim to the designation of In- dian,—the Dravirian. A third is required for the intermediate great for- mation of northern India and Ultraindia. Tibetan might be made to include the Indian and Transgangetic languages of the proper Tibetan type; but Mon-Anam has native characters which cannot be confounded with those of the more Scythoid Tibetan, and it is most convenient to use a distinet name for the formation as a whole. December 1856, Apprrions and CORRECTIONS, ch. v. sec. I Pronouns and Generic Particulars, Page 1. The calculation of chances here ascribed to Bopp, is Bun- sen’s. Alluding to the hypothesis that families of language had many dis- tinct origins, he says that “the very roots, full or empty, and all their words, whether monosyllabic or polysyllabic, must needs be entirely differ- ent.” “There may besides be some casual coincidences in real words; but the law of combination applied to the elements of sound gives a mathemati- cal proof, that, with all allowances, that chanye is Jess than one in a milli- on tor the same combination of sounds signifying the same precise ebject.” My objections to this position have been greatly confirmed by my subse- quent comparisons of Chinese, Himalaic and Seythic roots. The number of the elementary sounds taat entered into primitive language appears to have been exceedingly small. The same monosyllabic roots (phonetic) are repeated again and again, and meet us in every class uf words Lile all other arts, language was, in its earlier stages, rude and narrow, Only mateaial things were named ; and to the undeveloped family of sa- yaves, few even of these required names, The same uname suliced for muny objects having common properties, The growth of the danulytie faculty must have been very slow. Most new names were bat old ones * Sind (whence Mind, Ind) is a Himalaie root for river, 150 ETHNOLOGY OF THE INDO-PACIFIC ISLANDS. in new shapes. Distinct sounds were not in general invented or imitated for new conceptions. The conceptions and the names grew together from the old stock. The separation of families must have been the grand source of deaelopment, intellectual and linguistic. By this means chiefiy the primary roots acquired variety in phonology and application. Kach new family or tribe became the nursery of a new dialect; and the intercom- munication of these dialects gradually enriched ench with ideas and vo- cables. It was only by the aul of hundreds of sister-dialects that it be- came possible for any one dialect, after ages of growth, to make an ap- proach to a language in our sense of the word. In every period of time and in every group of languages the same mutual action 3 On, Hence, as thegenealogy of every existing dialect ascends to the begin- ning of human speech in the world or in the race, and passes through Jong periods of barbarism and of a minute subdivision of tribes, its voca- bulary has had innumerable proximate sources. Its discoverable homoge- neity is in proportion to the narrowness or exclusiveness of the circle of dialectic development and interaction. It may be ata maximum ina group that has always remained secluded, so far os the geography of any any province admits of this, and although the seclusion has lasted so long that archaically distinct dialects have now few vocables in commons. Page, 6. The new series of Vindyan vocabularies compiled for Mr. Hodgson by Mr. Nevill (J, B. As. Soc for 1856. p. 46) have the common form of the first pron, A-nuw Kondh, nga Savara, nui-se Gadaba, (noi-nyo oss., na-nu Yerukala, (na-mu, na-inbu-ru, pl. the secod form combimn- ing the absoyute and the relative pl, particles as in the Telugn mi-ru). The second pron. has the com. forin in Gadaba no, Yeruk. ni-nw (pl. ni- nga-lu=ni-ng—la Badaga, a-va-ru), Kondh yi-nu. Tn the Nilgiri series furnished by Mr, Metz, a du-m is viven as a sing. form of tlie 3rd pron. in Toda, along with a-du, »—-van, the pl. being a-va-r a-du-m. Page 15, Savara supplies a new proof that the labial 2d pron. of Kol is a plural form used in the singular, It uses the same pl. form, in both numbers, a-ma-n ¢how, a-ma-n ye (pe-n Gad., a-va ru Ye uk). it. Wumeralis. The new Nilgiri and Vindyan series have led me to adopt some modi- fication of the analysis in Chap. v. ‘The reasons will be found in the sec. on the Mon-Anam numerals in Chap. vi. ’ ; 1, The S. Dray. on-rw 1, I now read o-nru, and identify the root with vo of the 1 of Toda vo-dda, Telug. vo-/u-ti, Nilgiri vo-ndu, yo-dde, and the com. pa, ba of 10, pa-ku-du &e. (p. 56). Yerukala has vo-ndit =o-ndu Karn. The Male pa-ndo-ng, o-rds-ny 1 are similar forms. Comp. also the Telugu va-nd« 100. The Kol and Mon-Anam mo-é &c. 1 is the same root, with a different Dray. poss. postf. | 2. era-du and the variations in 2 and higher numbers I now read era- du, e.d, i.ru, er &e.e, i being the sole remnant cf the ultimate root, which in its oldest form had ra, re only as a postf, or second element, but afterwards superadded -du &c., probably on the earlier postf. con- ereting with the root remnant. Uraon has e-no 2, ma-no 3. From the facile and frequent elison of the initial labial it 1s probable that the full form of the initial root was be, bi. (p. 60.) The form be-ra &c. agrees with the Kol bay; and that ba is the initial root andr a second element or apostfix in-bar appears from its occurring with the guttural postfix in ETHNOLOGY OF THE INDO-PACIFIC ISLANDS, 15i Savara, ba-gu, a form peer also in7 of Yerukala vo-gu, Kiranti bha- g-ya and Mon bo-& (2 for 5), and identical with the Telugu vo-ka of vo-ka-ti 1 and the com. S. Drav. 10. It is probable that in be.ra &c, 2, ra was a second archaic definitive or unit. 4, If2 be e.ra-du &c,, 4 must be na-/n, i.e. the secondary element without the initial one. The Kol po.n retains the root of 1 with the conereted consonant of the second element or primary postf. The form opun-ia is probably from op-pu.n, i. e. op 3, pu.n tl. The Savara con- tracted form of 4, vo.n-ji (1_for 3, 1) is evidently the full form of the Tuluva o.n-ji 1. 5. The Gadaba mo-l/a-yi confirms the analysis of the Kol mu-na, mo-r &c. as 3 (S. Drav. mu-da &c.) for 3, 2. 8, e-ntu (not en-tw) as in 2. The Kol irl appears to be radically i.r-J, a contraction of the 8. Dray. Lra-du =i.ra-lu. The exceptional Gond, Telugu and Tuluva forms must, in conformity ‘with the amended analysis of 2, Be read a.na-mu-r, e.ni-mi-di, e.na-me i. e, ana, e.ni, e.na 2, and mu.r, mi-di, me 10 in the form of the unit found in the Kol 5, 8S. Dray. 3 &c. The Telugu tommidi must be to-mi-di,i. e, mi-di for 10 as in e.nt mi-di 8, and to, a distinct root for 1, preserved also in the Chentsu to-ta, 9, and corresponding with the Drav. dental 3d pron. and def. (p. 56), The Kol a-r of 9 Bes a for 1, as in the §. Dray. 6. The identification of the S. Drav. on 1 with vo.n, and e.ra 2 with be.ra &c. complete the proof of the agreement between the basis systems of 8. Dravirian and of Vindyan, and between both and the primitive labial system preserved in Australian. In its first form the system was simply the labial definitive, or the labial and liquid, repeated or compounded. It is clear that the liquid in ba.ra, bi.ra, &c. 2, is very archaic, and that it early possessed a distinct n.meral and plural force. In the Semito- African and Seythie systems, in which the same compound was the prin- cipal archaic numeral name, the liquid, changing to the sibilant &c., has been more stable than the labial initial, and there are strong reasons for regarding it as the essential element in 2. See App. te Sec. 6 of ch. v., The Semitic and African Numerals, pp. 18, 19, 43, See alsoa Note on the same subject inserted in the Journal with Sec. 1. of chp. v. 128. The note belongs to p. 129, and note f of p. 129 should be note * of p. 128. ig9, line 3 from the bottom. fer pishtk read ¢o-pisa, and in the follow- ing line, for to-pisa, read musa, mus. 137. The substantive root in the name for the Buffalve is the labial. The liquid is the root for water, e-ru-ma = water-cow. 138. In tango cow of Jili {not Singpho) the root is nga (a-nga). 140. The statement that in Chinese the root alone signifies buffaloe and the inference from it are incorrect. In Chinese, as in Dravirian, the name for the buffaloe is water-cow (or ow), and itis only by contraction that zu &e. alone is applied to it. 141. The Deer god is identical with the Bhotian god mare. 145. The sibilant name of Tiberkhad &c. is Tibeto-Ultraindian, 155. 7th line from foot, for moon, read silver. 157. del Sth line from ivot. ae | ria Sturn Scot il pa at —. a iiss! Pa i ies intvy A ee |: FL AT th ii bana 1 pit rn i ny Rind “ey tf Baw ae aos lee Ans * ys pele Z 2" a hi, +e: sate) ots lanlitey ; = ‘nesta Soar Ate a at ier ry Pa a | : . oe eo i. re Pt ee Pies c TO he ees % + ry in LJ : & £ * ; pon is Bp 7 pip ae ae 1 s vi, nis va Poke ars a ' J a a: hea ’ sey ‘ats eee : 2 i i! 7 im i wl ae ee Me he aos in a itp: ‘a ae a. wie Na : eal aa | i ee hae 4 ore 4 4 ain “s + tee at ee ca me 7 ie , * pg jf * oe Ga a leer = — Beh hi oie + atid ‘ te J J . - ' 1 = : e bs 3: a " , i 7 ie a i= om : Pa, pe. dy | 7 ~ ~.! " a = r . , * = = ¥ re : i = ss a 1 7 \ {7 s ] i) F > a - 3 rc . Fe ‘ 7 tw ¢ w 2 ‘ a 5 h % ul - Y * Ls a - “= ..* , a I é Pet ets ; > ¢ * v = ; a ~ % i’ — ae: os “hn F ta pe ‘ a ~*: ‘ J L é " * Cate ARRON NE RES > SN = - aie, * Riek ale, ae Sat al 1 re) a. a * 7 ek 4 dx . —_ we - es ye, Se Z r I eal Laie ‘ cal we wy Cie AS ee! fy LO: jt) .