1865.] 21 [Chase. 



9 ^:=one from the second ;* Ip. fakoq, Yu. sakoq, Chr. scabi, C. sap, 

 Ma. sablas, ten. 



In some of the Algonquin dialects, the word palin is used for five 

 or ten. E. g. Ck. palin, ten; Le. palina;;;', or palinu^c, Pn. pali- 

 nascu, = one pnb'n, Jive. The resemblance to Gr. ndXtv is signifi- 

 cant. We find in Chinese pa, to turn the back vpon, to put asunder, 

 pa', numeral of thinijs grasped with the hand (compare S. pantfan, 

 Co. manucht, five, L. manus, Hm. mn, fore-arm, shin), lin, five 

 families, neighbors, the appearance of numbers. 



Some successive probable derivations can be traced to an extent 

 that is wonderful, considering that the languages have no literature 

 or other means of checking dialectic variations. Thus a satisfactory 

 connection, in my opinion, can be traced between Wa. tau, three, and 

 Dk. loga (in shageloga, eight = second three), through the interme- 

 diate forms Dk. dogag, Le. na;fa, Nd. dahau, Wa. tau. Le. af, 

 second, and Wa. tau, three, give S. aftau= Gr. ozrcu, eight. Po. faca, 

 Kn. fogca, Dk .naptCiogka, naptfigwagka, nine (nape, hand, napin, 

 botli, napin-tfir)-wai)ka= &o^7i ivanting one), are all probably afiiliated. 



Some instances of special resemblance, which, if unsupported, 

 might be regarded as accidental, are entitled to consideration in con- 

 nection with the broader analogies that have been pointed out. Such 

 are Co. semmus, one ; Ma. sembilam (one from ten), nine ; Cn. clip, 

 first, before, H. aleph; Ip. sin, C. sing, L. singulus, one; C. si, Pa. 

 sis, Cd. sise, Pr. &\,foHr ; Ak. catif, S. catur, four ; Wy. tsutore, Ir, 

 tsata, S. saptan (all of which may be formed by Vesperian roots sig- 

 nifying second two), seven; Ak. nukinivan (^=. one from ten), S. 



* Duponceau, in his Volney prize essay (M^moire sur le Systfeme 

 Grammatical, »&c., p. 59), makes the following reference to the Algon- 

 quin numerals. "Mais de cinq a clix ces langues suivent une autre 

 m^thode, et c'est la meme que celle qui a ete suivie par I'inventeur de 

 chiffres romains. Pour six, on dit cinq un, pour sept cinq deux, pour 

 huit cinq trois et pour neuf un dix, c'est-£l-dire dix moins un, ce qui 

 r^presente exactement les caractferes numeriques YI, YII, YIII, IX. 

 Get ordre d' iddes n' existe point dans la langue de I'ancienne Kome, et 

 cependant se trouve dans sa numeration ecrite et abr^ge^. D'ailleurs on 

 ne raper5oit point dans la formation des noms de nombre des autres 

 langues connues." 



Hager (Explanation of the Elementary Characters of the Chinese, &c.) 

 has pointed out the resemblance of the Chinese and Roman numerical 

 characters, and I subjoin some of my reasons for believing that the re- 

 semblance is traceable in many of the Aryan, Shemitic, Hamitic, and 

 Turanian languages, as well as in their "numeration ecrite et abr6ge6." 



