38 EASTER ISLAND. 



In the lingual column the mutation l-n is characteristic of Nukuoro, 

 a speech that is best considered a somewhat recent Samoan derivative ; 

 it is sporadic in several languages. The extinction of / is strongly 

 marked in Niue, appears somewhat frequently in Tonga, and in South- 

 east Polynesia is abundant in the Marquesas. The consideration of the 

 retention of the Proto-Samoan aspiration and of its extinction, both 

 included in the foregoing tables, is postponed to later studies of the 

 aspiration in general. The extinction of the sibilant is common in 

 Mangareva, as we see in the chapter dealing with that tongue; it is 

 the rule in Rarotonga, it is sporadic in other Tongafiti languages; I 

 lack present record of its occurrence in Nuclear Polynesian. 



In the labial column the mutation f-v, an interesting variety since it 

 involves the unusual change from surd to sonant, is characteristic of 

 Viti, sporadic in several languages in each household. The extinction of 

 / is characteristic of Mangaia, Rarotonga, Bukabuka, strongly marked 

 in Mangareva, and sporadic in Nukuoro and Rotuma. 



We next pass to a similar tabulation of mutation of the Tongafiti ele- 

 ment registered upon Maori as the most readily available base. These 

 variants are very few ; they occur only in the linguals and labials : 



h 841 842 hw 839 844 



hw-ft 952 w-o 842 843 864 871 912 955 956 957 



hw-K 922 w-h 850 855 856 858 



The Maori hw being a mutation of Proto-Samoan /, the three entries 

 are reducible to f-h, f-v, /-, which have been established in the larger 

 table with Samoan comparatives. Similarly as Maori w is a mutation 

 of Proto-Samoan v, the foregoing entries reduce to zv-v, which therefore 

 cancels itself, and v-h. The former of these exhibits Rapanui as closer 

 to the Proto-Samoan than the Maori. The latter mutation is very 

 rare. In "The Polynesian Wanderings" I noted it once each in three 

 languages; three of the instances here noted show great irregularities 

 in the comparative histories of the several vocables upon which they are 

 based. 



Up to this point we have concerned ourselves with the investigation 

 and record of phonetic mutations, a point at which, undoubtedly from 

 motives of convenience, philological comparison most commonly re- 

 gards its labors as complete. But this is an exaltation of form over 

 substance. It does not call for deep insight into speech as the utter- 

 ance of the inward thought of sentient man to recognize that form 

 may be a grace, but it is the sense that is the life of the word. 



We next shall pass to the examination of these words of Rapanui, 

 wherever comparable, in the effort to discover what information they 

 may be made to give us of the position of this distant folk among greater 

 families of its race. It may be that we shall not find much ; it may well 

 be that rules for the government of such inquiry may not distinctly 

 establish themselves, for the field is new. A new acre in a field so little 



