DATA AND NOTES. 227 



Sesake, Nggela, Belaga, Pak: va, id. Aneityum: apart, han, id. 

 Pala: han, id. Duke of York, Matupit : wan, id. Kabakada, 

 Matupit, Baravon : wana, id. Lambell : han, id. King : ivuan, id. 

 Norbarbar: vana, a going. Vuras, Gog, Lakon: vanog, id. 

 Motlav: navnog, id. Tangoan Santo: thano, to go. 

 Hebrew : panah, to turn the back, to turn to go. 



See note 1 1 . 



With the readily adjustable exception of the metathetic form in Uvea the 

 Polynesian is a unit and in perfect identification with Efate bano. Other 

 Efate forms are banbmai, banamai, banimai, compounds with directive mat; 

 banbtu, binotl, bariats, bhiats, compounds with directive atu, with which our 

 author most cryptically includes bin'en, baina, notu, net; ba-ki, which he 

 says is "contracted for" ban, bano, while ba (n) is not. 



In considering the word in Melanesia we observe that there is a partition 

 into two markedly unequal areas according as the radical vowel is a or e, 

 to which is to be added the single instance of Volow which has o for a 

 (neutral vowel). 



The a-series has a wide range among the labials from b to v, but the forms 

 in v are far the most widely spread . The only exception to this vertical muta- 

 tion is the Tangoan Santo involving a f-th or horizontal mutation, that is 

 from labial to lingual, a change of which we have no other example, yet in 

 this case it seems quite correct. Aneityum, Pala, Lambell han, according to 

 my principle of the aspiration, is a vertical mutation. It maybe that in this 

 we see a transition form to Tangoan Santo thano; a labial h reached verti- 

 cally would in no way, save with the resources of comparative philology, 

 be distinguishable from a lingual h, and for the change from this h vertically 

 downward to th we have abundant evidence ( 1 7 Journal of the Polynesian 

 Society, 160). The mutation to w is abundantly supported and is all but 

 vertical; it is found only in the gate by which the Samoa track leaves 

 Indonesia, and, disregarding the extra prefix, which is probably formative, 

 I incline to regard the vu of King ivuan as but a reinforced w. In Kaba- 

 kada, Matupit, Baravon wana we shall best regard the final vowel as an 

 o-a change within the triangle of the neutral vowel ; the a-final in Norbarbar 

 vana is formative and establishes the verbal noun of van to go, as appears 

 above. In Vuras, Gog, Lakon, Motlav the final g is formative of the verbal 

 noun; the Motlav is na-vno-g, article-stem-suffix. There remain for con- 

 sideration a few irregular forms. Aneityum a pan we include because our 

 author includes it as an identification, which it is not. Apan means to go, 

 apam means to come ; the merest tyro should recognize the common stem 

 as apa, far removed from vano and further exhibited in apahai to go land- 

 ward and apaahni to go everywhere. He quite overlooked, or did not know 

 where to look for han, which is the true identification. The King vanwin 

 contains the vano stem together with an extra element which in the paucity 

 of our material from that center of New Ireland culture we may not 

 comprehend. 



The smaller e-series runs through its simple course of mutation and abra- 

 sion and calls for no more extended comment than to call attention to Tanna 

 (t)uven as a mixed form. 



