60 



EASTER ISLAND. 



In like manner we tabulate the three exterior elements by the Pau- 

 motu identifications in the same neighbor islands: 



Table 6. 



Finally, in computing the relation of these three external identifica- 

 tions to the mass of Paumotu identifications we obtain this table: 



Table 7. 



The last table shows at the merest glance that the two elements of 

 Paumotu speech vary widely in relation to the rearward past of their 

 race. That Paumotu which is recognizable as affiliated with Rapanui 

 preserves 80 per cent of its history ; the other Paumotu shows little more 

 than half as much. That the two represent different movements of 

 population is highly probable, but beyond the expression of the opinion 

 that such is the case I hesitate to venture. A certain antecedent prob- 

 ability fosters the view that this second Paumotu is the remnant of an 

 older population, longer seated in the Paumotu, upon which the Rapanui- 

 affiliate Paumotu descended in the course of their voyaging and there 

 deposited colonies of the younger stock, while the more venturesome 

 pushed bravely out into the enticing east. The prime bases of such an 

 opinion are, that this element is numerically almost two-thirds of Pau- 

 motu speech as known to us, and that it is fair to consider that the people 

 longer separated from the central home of the race must have under- 

 gone the greater loss of speech material. This is a justifiable reading of 

 the two percentages 80 and 45. But when we inspect the details of this 

 westward affiliation we encounter difficulties which we may not venture 

 to adjust to such a theory. It might be possible to pursue this farther 

 if it were not for the character of that element which I have for con- 



