DATA AND NOTES. 255 



The change in the initial syllable from a to e, which we find in the remote 

 Marquesas, is here observed in Santo Wulua, Mosin, Leon, Norbarbar, and 

 Retan. Metathesis is observed in Leon from Mosin, and in Re tan from 

 Norbarbar. Retan vege might be fika, after the manner of the loss in 

 Samoan ; but as Leon vegig must be Mosin gcvig, which is kafik, we prefer 

 to consider Retan in accord with its neighbor. Motlav na-gveg I have 

 classed in the foregoing with the dissyllables of closed stem. The initial n 

 is the shadow of the article, just as ni in Tanna, but Motlav is a very diffi- 

 cult language to write because the article attracts the nearest vowel of the 

 stem, in this case gaveg; therefore na is article for this word and in attract- 

 ing has also subtracted. 



Pak, Sasar, and Alo Teqel are a separate and unassociable group. 



The Semitic proposed by Dr. Macdonald is clearly irrelevant. 



194. 

 kori, kuri, oria, kuria, a dog, a brave, a warrior. 



Samoa: 'nil, a dog. Tonga, Futuna, Niue, Uvea, Fotuna, Aniwa : 



kuli, id. Maori, Rarotonga, Manga reva, Paumotu, Sikayana: 



kuri, id. Tahiti: un, id. 

 Viti: kali, a dog. 

 Baki, Epi, Ambrym, Santa Cruz, Deni: kuli, a dog. Iai, Tanna, 



Eromanga, Malekula, Aneityum : kuri, id. Epi : koria, kuliu, id. 



Sesake: koriia, id. Malo: vuria, id. Tangoan Santo : vuriu, id. 



Bierian: kuliu, id. Santo: wurin, id. 

 Arabic: gorw', a young dog; gariyy' , brave. 



The identifications are satisfactory and the stem shows no variations 

 until we reach Malo and Santo with the k-v mutation, which is found only 

 in these two languages and in this word alone. The closely similar k-w 

 mutation spreads over a wider area of speech, yet also is found in but a 

 single word. The Santo word wurin derived from Dr. Macdonald's gram- 

 mar of that language is so close to Tangoan Santo vuriu that I must regard 

 the final n, otherwise incomprehensible, as a fault of the press. 



The Arabic does combine in similar words the Efate significations, yet 

 not even on that account is it convincing in the absence of confirmatory 

 evidence to bridge the gap. 



195- 

 koto bolo, a basket (bolo, basket). 



Samoa: 'ato, a basket. Tonga, Fotuna, Uvea: kato, a basket. 



Niue : kato, a basket, a bag. 

 Viti: kato, a basket. 



Lo: gat, a bag. Malekula: na-cat, a basket. Malo: cete, id. 

 Tanna: katum, id. Aneityum: in-cat, id.; nefehcat, a large 



basket. 

 Arabic: ka'tat, a basket for carrying dates. 

 As koto means a basket and bolo means a basket, the use of synonyms 

 in composition is an interesting example in Melanesia of the principle of 

 determinant compounds, the existence of which in Samoan I have else- 

 where established (14 Journal Polynesian Society, 40). Koto is readily 

 identifiable with kato in Nuclear Polynesia and Melanesia and no forms call 



