POLYNESIAN GRAMMAR. 249 



In the Tahitian and Rarotongan they discard the ta'au and takau altogether, using only 

 dhuru or yauru for ten, and forming the higher numbers regularly (rua ahuru, torn 

 ahuru, &c.) up to haneri, hundred. In Hawaiian, they proceed from tanaha, forty, to 

 tancdima, fifty, tanaono, sixty, and so on to the same word hcmeri. 



In the New Zealand Grammar of Professor Lee, and in the missionary translations 

 into that dialect, tekau is used for ten in all the numbers above nineteen ; for twenty, 

 they give rua tekau, for thirty, torn tekau, &c. Yet it is certain that these terms mean 

 respectively twenty pairs, thirty pairs, and so on ; or, at least, this is their proper and 

 original signification, although some of the natives, under the instruction of the mission- 

 aries, may now have adopted their mode of computation. The origin of the mistake pro- 

 bably was the fact that the natives rarely have occasion to use the higher numbers, except 

 in counting fish and potatoes, and these are always counted in pairs. A person hearing 

 a native say for one pair, tahi, meaning simply one, for two pairs, rua, meaning two, 

 and so on, would naturally suppose that tekau, used for ten pairs, meant simply ten. 



In the Paumotuan we can observe the process by which the reduplication of the Tahi- 

 tian and Rarotongan was probably effected. In this language there is a double set of 

 numerals, one for counting single objects, and the other for pairs. They are respectively 

 as follows : 



rari, one 



ite, two tikaite, one pair 



yeti, three 



ope, four terueope, two pairs 



hene, five 



yeka, six miahene, three pairs 



hito, seven 



haiva, eight tueptka, four pairs 



nipa, nine 



horihori, ten horihori, five pairs 



For twenty the term given (as we understood it) was ite takau. We supposed then that 

 these natives followed the same mode of computation as that given in the New Zealand 

 grammar, of the incorrectness of which we were not then aware. It seems likely that 

 we made here precisely the same mistake as the compiler of that grammar, and that 

 takau does in fact belong to the second, or duplicate set of numerals, and means therefore 

 ten pairs. It would then be just the double of horihori, when the latter is used for five 

 pairs. It is easy to see how takau might, by a careless usage, be transferred by the 

 natives to the first set of numerals, and be taken for the double of horihori when the 

 latter is used for ten ; in which case takau would mean simply twenty. 



The Paumotuan word for hundred, penu, means also the head ; we are reminded of 

 the Samoanfulu, ten, which means likewise hair, and lima, which signifies both five 

 and hand. The notion of connecting the names of numbers with parts of the body would 

 naturally arise from the habit of counting on the fingers. 



It should be observed that the natives, in most of the groups, commonly prefix to the 

 numerals the particles ka or a, and e, which are probably the verbal particles of affirma- 

 tion and present time (vide 50 and 52). The first (ka or a) is commonly used when 

 the numeral precedes the noun, as ka toru ya waka, N. Z., there are three canoes ; a 



63 



