240 ETHNOLOGY OF TIJli: 1NT)0-PACO-IC ISLANDS. 



root for 2, \t vnmtha.ve become ob&okte in that numUer when the system le* 

 came tJiat of Tibet J, niid no remnftnts of it are pre.st^rved in other 

 jTumlier^. The IfllU-r circtinistaiKe i.-^ hardly tionsislpjit with its beiiigra 

 jiniunry dual rtjot of the M stem, Oji the other hand if it he ti unit it iaidrn- 

 titiiJ with the oomnioii .sibilant sitKl liquid formi* of the Chino-lliinuhiicuuitj 

 aiid theChincr^e fi\>t('in mmt have liud a tiinal basis thrmigJjouf,— thiit is 

 4 must linvL' been iij 1, in likoumnner us 5 was Ij, 2, and 7 wiis ti, 1. 'J'Jiis 

 mode of fujTOiny 4 ia rare, btU ejtujtiideii of it otror hi the K. E. Anan and 

 Afn€?tn pwiiiLCfl ; and the Yiruh an 4 is also InJiftl. As the Giino-liiuui- 

 3ait: sibihint 1 and !i has forms precisely Bimilar to 4, lioth in 3 and in other 

 nninbera, I do not Jiesittite to ijanster 4 t'lam tlie dual w the unit fieiies. 



Vt'hen the Hvf-t Ilimalaic diaieut carded south ft broad o, « form of 

 tlie dual ivfli* conmion to '2, 6, 7, and, proljably, to 8; nml it aniears to 

 Imvehad the labial jmfix throujihoat, as it preserves it in fj, uiul nia few 

 dialects later forms hu\e it in 2, The uint aum'jutrd with thi.-i form 

 of the dual niusl hit ve been, une of the oldejit carried South , ami ua the 

 later form wns eontemiicirinieous with the Tibetan pvud, ryat A:c, 

 it IB probable that the njjo, nbu, dual was that of !he ^iy^^e^^l isbieh had 

 klmuii- ke, rs the unit, in the Abor group buth khung; and ngo [2) 

 are jireseived, 



"When the next *;reyt mifrmtion took place the predominant Tibetan dia- 

 lect had a inimeral series which had the form iias- &c. in '2 and in iIjm quina- 

 xy 7j — ni^soeiated with turn, ?i:nu, ram He. In 1, 3, lU, and in the denary 7 ; 

 and the hdjiid prefix had ^iven pbue to the puttuinl. Jjtiphla retains fnr- 

 ijoijr in 7, naii in B; and in 2 iSniiniin wv. has niirh, Ten^nn n-imt (for f^ff- 

 nat)j L"hani;io nfxik, AVith a few exception^ the t^iMitliei n Yiinia*Uangetic 

 lbrn]5 are moditieationti of this form. In Tibet tlje brond vuuul is returned 

 in Thochn and Manyak nfra, im uiid in the lh>r|iii 2 of 20 m\. hi the other 

 'Jibitan dialects nml in iiiotit of the Southern ones* slender and contracted 

 forms now prevail. Tln» eliinifre has j^enerally been kjcul. The Horjin j-nc 

 nml Gynrun^^ A/t-nes of Tare from ii form of2 similar to the ( iynrun^- At-nis 

 ill !iU and Dhotjan //-nyis 2. Jloth are probaljly fnan t uo diiilect, pcrhiip:^ 

 Hrtrpjij whieh liii:* a .^iitiilur I'urni of tlio jtreJix in U js-'^ii. It has ji^e in 'J 

 which may liave l»een /v-ii^es, whwu c j./i-Ju'S, The Gyarunir ^A-ne* 



o! 7 is jiot the turreiil Lu-nvis 2, but a concreted vocable, it lakes the 

 eurrent pn tix /, (f-jib nes. The Inter ssoutbem forius havtr niiiuy varietiei*, 

 ajjd rtonie piiuilur to the Tibetan, bat the< are all of haul f^rowth. Thou 

 the Bodo arid (i;iru s-m 7, tilthoijjtfh m closo ti> the fJorjia ;-ne, is from 

 s*-ni preserved in tlie ^ilikir Af-ni 2 (Hinjijdjo .<Ji-nit 7); iind A)-ni ii* from 

 //i-ni, -yj-ni d'aro. In the Viuasi 7 ii In-cniiie*? r, but the prefix identities 

 it witl) these (ttiiip.etie furmn. Cojup, Tenji^a in'uif Ssik (iui-wly 8hindu 

 ishn-Yi, Kami, Knki 6«-ri &c. The assodattd form nit \s;c. shuwri that ni, 

 ri bc]oii|;s la the iater ngiik, ngik, ug^it series in 2. 



The llimalaic form of 2 brouj^ht south hy the 3Ion-Anam family 

 nji]»ears to lune been the bruiid lurm with tfie Itibial prelix preserved 

 in the Tibeto-l Itraiiulian 5. In tlie Don^'-jii, Knki, Tiipkaaud Afor S and 

 in the iJophhi 8, hirni of - is eonjtjined w itli the liqnid 1 (for 3 and 

 10), whence it ia clear thai when that form of 1, u and lU wus urrrentj 

 tin: Chiuo-Jiintaluic hrrm of 3, njro, niiu, was current as 2. If the 

 Lau MHi^, i»anfr id from an 1 or n form it wuiS probably one of the 

 varieties of the AIou-Anaiu 2. Tlio ctureut Id in all the other dialects 



