82 EASTER ISLAND. 



matter. In navigation it becomes a rather simple problem of sailing close- 

 hauled, and to the solver, as to the captain of the canoe, the only serious 

 difficulty is to get across the equatorial doldrums. 



We shall now pass to the systematic examination of the speech of 

 Mangareva and its general and particular relations as may be deduced 

 from the affiliations which we are able to establish. 



We note at the outset Meinecke's very positive statement :* 



Die Bewohner von Mangareva sind Rarotonganer, die von ihnen gespro- 

 chene Spraehe ist bis auf unbedeutende Verschiedenheiten die von Rarotonga. 



His subsisting authority is not recorded, an unusual neglect to be 

 charged against this very painstaking and exact historian; but we may 

 infer it from his precisely similar statementf concerning the Paumotu : 



Diese in den ostliehen Inseln jetzt noch gebrauehte Spraehe ist nach CailletJ 

 ein rarotongischer Dialekt, und wenn gleich nieht wenige Worter ganz von den 

 in anderen polynesisehen Sprachen verbreiteten abzuweichen scheinen, so ist 

 doch eine andere Zahl wieder entschieden rarotongiseh, und aueh in der Gram- 

 matik ergeben sich keine erhebliehen Verschiedenheiten. 



We should observe that the Hervey Group is singular among the scenes 

 of the activity of the London Missionary Society in that no dictionary has 

 yet been published. Such a work was undertaken by the Rev. William 

 Wyatt Gill, but late in his career of great usefulness he heard the call to 

 the apostolate at Port Moresby and soon died of the bitter hardships of 

 pioneering in New Guinea. It was therefore impossible forCaillet to have 

 made such a determination, either for the eastern Paumotu or for Ma- 

 ngareva, in the absence of material upon which to erect a comparison. It 

 appears to me that what he did observe was that in the speech was an ele- 

 ment which he could recognize as non-Tahitian, and that he leaped to the 

 conclusion that if it were non-Tahitian it must yet have some source and 

 that therefore it must be Rarotongan as being next rearward on the track 

 of migration. How significant is the marked difference between Manga- 

 revan and Rarotongan is seen in the comparison of the alphabetic 

 scheme. The aspirate is entirely absent from Rarotonga; it is, indeed, so 

 objectionable a sound that f , which is generally mutable to the aspirate 

 proximate to the labial series, is for that reason frequently carried thereby 

 to extinction. On the other hand Mangareva retains the aspirates with 

 considerable persistence and the labial aspirate as a mutation product of f 

 is very commonly observed. We may disregard this pronouncement as 

 to the Paumotu and Mangarevan, since at the time of Caillet's investiga- 

 tion it had not yet come into the mind of any student to examine speech 

 sources through the division into the Proto-Samoan and the Tongafiti 

 migrations. 



*Die Inseln des stillen Oceans, ii, 222. 



tOp. cit., 215. 



JAnnales hydrographiques, xxxiii, 392. 



