140 



EASTER ISLAND. 



In each element of this material the Marquesan shows the closest 

 association with Mangareva and the position of the Paumotu is rela- 

 tively remote. 



As before, we set the three rearward elements of migration source in 

 Table 26. 



Table 26. 



In our study of the Tahiti material we took occasion to note at this 

 stage the identification of 105 items exclusively with Hawaii, and of 23 

 identifications shared with the Marquesas. In the collation of the 

 Marquesas we find, additional to the 23 items brought over from Tahiti, 

 the exclusive identification of 133 items in Hawaii and 23 others for 

 which we are able to identify a Proto-Samoan source. It would carry 

 us beyond the bounds of this province of Southeast Polynesia if we 

 were to venture here upon any discussion of this phenomenon. That 

 may properly be left to the workers in Hawaii. In the "Polynesian 

 Wanderings" the dissection of purely Proto-Samoan material showed 

 forth the trace of a migration of that earlier stock due north from Nuclear 

 Polynesia with no evidence of an intermediate halt. The items here 

 assembled show that a migration of indeterminate source passed from 

 Tahiti to Hawaii direct, from Tahiti to Hawaii by way of the Marquesas, 

 from the Marquesas to Hawaii, and in all this record of migration there 

 is clear evidence that at least one of the squadrons on the Marquesas 

 road to Hawaii was Proto-Samoan. 



The relations of the identifications 

 of earlier source to the mass of Mar- 

 quesas identifications are presented in 

 Table 27. 



Comparison with Table 20 in the 

 chapter on Tahiti shows that the 

 same principles are operative in the 

 Marquesas also. In the Rapanui 

 element we find the same high per- 

 centage appertaining to that indiscriminate source which we can class 

 in no other fashion than as general Polynesian, the low percentages of 

 the two distinctive migrations; in each language the Tongafiti shows 



Table 27. 



