THE SOUTHERN GATEWAY. 487 



328. 

 We should remove the Mekeo and Roro words from this item; they are 

 variants of lango (309) and have already been discussed in their proper 

 place. So far as we have words with the n-m skeleton we may feel secure 

 in the identification. At the same time it should be noted that the vowel 

 treatment is the less usual one in each of these New Guinea languages, but 

 each finds confirmation in Melanesia, and the namo form is found in the 

 Polynesian of Fotuna. The more complicated forms call for examination in 

 detail. In the composites two alien elements are involved. The iodi of 

 Awalama is so clearly the eoti of Tavara that we are assured that the mo of 

 the latter is a wasted form of Awalama himo. As between the succeeding 

 members of the other group, kini of Taupota and Wedau, and the nika and 

 niku of Nada and Murua, there is an evident metathesis, but it is imprac- 

 ticable to determine from our material the direction in which it has acted. 

 We then encounter namu in the series giwiu-kimu-sitmo-simu-kimo-imo-mo. 

 The gumu form illuminates Tangoan Santo moke, and himo the Malo mohe, 



both by metathesis. 



329- 



It is interesting to note in this area the close proximity of udi and udu, 



the latter of which is so strongly prevalent in the eastern gateway. 



330. 

 As soon as we pass from the identical niu in New Guinea we encounter 

 some very interesting series of Indonesian resemblances. Misima niku, 

 with the inner assumption of the aspirate, is identical with the Malagasy. 

 The forms with final liquid in Mukawa, Awalama, Taupota, and Kwagila 

 are readily comparable with forms in Indonesia subject to the same modifi- 

 cation. The riira of Kubiri and Kiviri is a doubtful form, but the presence of 

 the final liquid syllable may serve as a link to the foregoing. The Kiriwina 

 forms seem associable with the Sulu and Ahtiago, metathesis of the 132 type 

 affecting the two vowels. 



The Motu identification is so distinct (and Galoma is clearly metathetic, 

 of the 3214 type) that it carries the nanu forms. Furthermore the muta- 

 tion \-n is well established, not only in New Guinea and in Melanesia but 

 in Polynesia as well. 



All these identifications are satisfactory until we encounter Mekeo nga- 

 ngau-nga. I have admitted it for inspection because it has the lau stem 

 vowels properly placed, but the consonant mutation (which, however, is 

 normal in Moiki and has a certain standing in Melanesia) and the dupli- 

 cation lack confirmation in the speech and in the life-history of the word. 



336. 

 After the excellent identifications in Mekeo and Tagula we shall have to 

 grope our way through local modifiers and degradation forms. Excising 

 the formative elements, we pick out nono in Mukawa, Tavara, Wedau, 

 Awalama, and Taupota ; and this is cut down to no in Oiun. The loss of 

 the initial liquid (a well-supported movement in the four languages) gives 

 ona in Roro, ano (321 metathesis) in Raqa, oono in Kabadi, which is yet 

 further reduced to oa in Uni. 



