THE PAUMOTU IN THE POLYNESIAN SCHEME. 61 



venience designated Polynesian; being common to the two migration 

 swarms, its presence in any given language might be due to a Proto- 

 Samoan migration or to a Tongafiti voyage. That is a point which we 

 may not determine, the material is incapable of revealing its source; 

 being neutral, it removes itself from the computations. The evidence 

 of the Proto-Samoan and the Tongafiti elements is also negative. On 

 the hypothesis of a secondary population receiving accession from a later 

 swarm, and taking into consideration our knowledge that a migration of 

 the Tongafiti left Samoa for new lands, we should expect to find in the 

 sedentary people a preponderance of the Proto-Samoan, in the newer 

 swarm a preponderance of the Tongafiti. But this table has made it 

 plain that the Tongafiti preponderates in each, and that, though the 

 figures vary, the ratio is practically the same ; the Tongafiti outbalances 

 the Proto-Samoan just about four to one. In the extra-Rapanui Pau- 

 motu there must be some significance in the extreme paucity of the gen- 

 eral Polynesian, about half of the Tongafiti. Tentatively I suggest that 

 this may signify that the 13 per cent Polynesian was brought in the 

 Tongafiti swarm. This contravenes the hypothesis that this Paumotu 

 element represents an earlier and sedentary population. 



When we pass backward to the next preceding table, wherein to the 

 triple earlier identification of the sources of the material is added the 

 record of the affiliation within the province of Southeast Polynesia, we 

 find a most interesting and really illuminating condition of affairs. In 

 the Paumotu common to Rapanui we find the Proto-Samoan and the 

 Tongafiti elements at practically the figure which we established in the 

 bulk computation, but the general Polynesian has increased by 10 per 

 cent, and that equally in the three other languages. We observe also 

 the great evenness of the distribution to Tahiti, Mangareva, and the 

 Marquesas; the percentages vary only in the slightest degree. This is 

 evidence that the Paumotu element which has reached Rapanui on the 

 long eastward voyage has made itself felt most evenly through the more 

 compactly placed archipelagoes of the province. But when we turn to 

 the other half of the table, that which deals with the Paumotu not shared 

 with Rapanui, we find great irregularities. In the elements shared 

 with Mangareva and with Marquesas the general Polynesian material 

 has increased by about a third over the table discussed in the preceding 

 paragraph, and the Proto-Samoan and the Tongafiti material remain at 

 practically the same figure. But when we examine the Tahiti affilia- 

 tions we find a striking change in all except the Proto-Samoan material, 

 which has undergone a small and negligible increase. The general Poly- 

 nesian has almost doubled, and the Tongafiti falls little short of the 

 same increase. This is the first instance in these considerations where 

 our attention has been directed to a close alliance between the Paumotu 

 and Tahiti. We shall have to concern ourselves again and yet again 

 with this alliance. 



