154 EASTER ISLAND. 



worthy only in the Tongafiti element, remembering that in our known 

 landmark of Samoan history a primitive race of savages groaned under 

 Tongafiti savagery; for there seem to be degrees in that sort of thing 

 recognizable when one is on the same plane. Yet in this record we 

 have a not inconsiderable Proto-Samoan contingent. The explanation 

 seems to me to be that at some spot to the westward a colony of Proto- 

 Samoans had settled ; that upon them came a swarm of Tongafiti ; that 

 the earlier settlers made their escape up the wind and settled in the 

 province ; that in time the Tongafiti marauders, whom we know to have 

 been expelled from Samoa, set out in pursuit and added to the record 

 this uneven curve. The smoothness of the general Polynesian curve 

 and of the Proto-Samoan curve is evidence that the colony from which 

 the escape was made must have been a place where the two streams had 

 been together sufficiently long to admit of a considerable mingling of 

 speech material. We know such to have been the case in Samoa. The 

 party which should follow over the same course should be more strongly 

 marked as to Tongafiti character. If this place of swarming were 

 Samoa, we know that the element expelled was Tongafiti. If it were 

 some intermediate spot upon which the Tongafiti sought to repeat their 

 Samoan conduct, the party which should set out after the Proto-Samoan 

 refugees would be of the dominant Tongafiti oppressors. So far, then, 

 as relates to this chapter of the history of the province, I feel convinced 

 that in these records we have the history of two migrations, one of a 

 somewhat mixed character followed by one strongly Tongafiti. 



There remains that chapter which is to deal with the speech material 

 more or less general to Southeast Polynesia, but which has passed 

 from use elsewhere in the family. Mangareva and the Marquesas are 

 equal at the lowest point; the one lies beyond the extreme weatherly 

 point to be reached closehauled on the trade wind when sailing out of 

 the west and in the westerly variables, therefore less likely to be reached ; 

 the other lies to windward on the other tack, and before it can be reached 

 the wandering fleets will have had the opportunity to find settlement 

 in Tahiti and the Paumotu. It is for this sea reason that we expect 

 to find, and do find, them occupying a higher position in our record, 

 Tahiti one and a half and the Paumotu three times the minimum. As 

 between the two, we may readily account for the discrepancy, the Pau- 

 motu stretches all of three times as far athwart the wind as Tahiti. 



Now we may derive some conclusions as to the relative age of these 

 two elements. In the Paumotu the restricted provincial speech is nearly 

 five times the general Polynesian, eight times the Proto-Samoan, two 

 and a half times the Tongafiti. In Mangareva it is twice the general 

 Polynesian, almost three times the Proto-Samoan and the Tongafiti. 

 In the Marquesas, here also in close association with Mangareva, it is 

 twice the general Polynesian and the Proto-Samoan, a little under twice 

 the Tongafiti. Finally in Tahiti it is three times the general Polynesian, 



