DATA AND NOTES. 303 



Matabello: nifoa, tooth. Saparua: nio, id. Malagasy: nify, id. 

 Guam: nifin, id. Manatolo, Sula : nihi, id. Bouton: racfti, id. 

 East Vaiqueno: nissy, id. West Brissi: nissin, id. Savo: 

 nuhsi, id. Kisa: nihan, id. Kayan: knipan, id. Magindano: 

 nipun, id. Tagalog: ngipin, id. 

 Arabic: 7ta6', nnbiib', tooth. 

 Dr. Macdonald does record that this is &a/i (39) with the article, and he 

 does set it down that Polynesian nifo is "another word for tooth, teeth," 

 and he includes his Semitic w r ithin the same brackets as the lower set of 

 these teeth. But why does he collate the nifo material with this nabati if, 

 despite his protestations, he did not think it added to his Semitic scheme? 

 The w-form which holds without exception through Polynesia is found in 

 the western Pacific only in Ulawa, Saa, Bululaha, and Buka, stages in the 

 northern Solomons on the Samoa track, and in Lifu and Iai at the extreme 

 south, the terminus of a migration concerning which Codrington notes 

 (Melanesian Languages, 17) "more archaic they well may be, belonging to an 

 earlier movement of population, carried forwards by an earlier wave of 

 speech passing onwards among the islands, but having somewhere a com- 

 mon origin with those which have since and successively passed among 

 them." Yet in the instance of this word we observe as a curious fact that 

 they most closely resemble Mangaia and Rarotonga of the Tongafiti swarm, 

 known to be the most modern of the successive waves of the speech. 



Buka with its niho and liho is the sure identification of the transition 

 phase of the passage to the /-forms which are so characteristic of Melanesia. 

 These forms differ in the second consonant from the primal / in Fagani lifo 

 and the immediately proximate h in Ugi and Bougainville liho and Wango 

 riho, all these in the Solomon Island crop colonies of the Samoa track ; up 

 the labial column to v in Alite, Vaturanga, Nggela, Bierian and Epi (if this 

 be not a reference to the Bierian) ; thence to the semivowel approximating 

 the labials, w, in Mota, Motlav, and Maewo; finally down the column to 

 b in Malekula ribo. So far, in all Polynesia and in all Melanesia, the two 

 vowels have remained unaltered in quality and fixed in position. 



When we examine the Indonesian retention of the nifo stem we shall find 

 but one of the four elements of the word which has been treated with care, 

 namely, the former vowel, which preserves its place unaltered in all except 

 Savo. Matabello alone has nifo and has added a decoration of its own; 

 Saparua has the form which we have seen in the Loyalties and in Mangaia 

 and Rarotonga. There are here no /-forms; the n is found intact in Mata- 

 bello, Saparua, Malagasy, Guam, Manatolo, Sula, Bouton, East Vaiqueno, 

 West Brissi, Savo, Kisa, Magindano ; in Kayan it is prefaced by a palatal, 

 and in Tagalog it has become the palatal of its own group. The second 

 consonant is kept as / in Matabello, Malagasy, Guam. It becomes h only 

 in Manatolo, Sula, and Kisa, and in Bouton nichi it follows the same change 

 as in Sikayana nitcho. It passes irregularly to s in East Vaiqueno and West 

 Brissi, and to a gruffer sibilant in Savo. This mutation we explain as 

 meaning that the first change is to an aspirate lying close to the labial posi- 

 tion of the buccal organs; then that the mouth can not hold in this position 

 so amorphous a sound as this aspirate; that the aspirates lying near all 

 three series tend to assume the position of the central, or lingual, aspirate; 



