EFATE AND VITI AND POLYNESIA. 



41 



a satisfactory basis. The Efate material is adjusted to the preceding 

 table of Viti-Proto-Samoan-Samoan. Corroboration is not listed 

 in the case of the more frequent concords or mutations. 



The following mutations, which have been observed in the com- 

 parison of Viti and Samoan, are not involved in any of the material 

 for which we possess Efate data : a-u, nd-h, dh-k, k-t, v-ng, ng-n, mb-v. 



The results of this collation are exhibited in the following table 

 showing to what extent there is an agreement persisting through 

 Efate- Viti-Samoa, in cases where Viti and Samoa differ to what 

 extent Efate agrees with Samoa and with Viti respectively, and to 

 what extent it differs from each. The cases of agreeing vowels have 

 been omitted for convenience. The spirants and mutes of the labial 

 series have been omitted because, owing to the absence of one or 

 other from each language, it has proved impracticable to establish 

 a comparison; the items involved in this omission amount to 52. 



We now pass to the consideration of the data involving Efate 

 and Polynesian which find no bridge in Viti. (Table C, page 40.) 

 These will be found under the items 48-63,73-120, 147-185, 239-271. 

 Here again, in the possession of 160 items, we have satisfactorily 

 sufficient data for a comparative study. 



The results of this collation are displayed in the following table. 

 As before, the spirants and mutes of the labial series have been 

 omitted because comparison is impracticable; /, however, being 

 comparable, is included. 



From this material we diagram the relation of Efate and Samoa 

 to their common parent. 



Efate -Parent-Samoa. 



r-r-1 



1-1-1 



tr-r-1 



ng-ng-ng n-n-n m-m-m 



