44 EASTER ISLAND. 



lack the antecedent norm for a standard of deviation, yet we must 

 employ the Maori because of its superior dictionary equipment. 



We are well aware that, so long as the comparison must rest upon 

 some modern Tongafiti language, we should obtain better results by 

 employing some one of the languages spoken in the Hervey or neighbor- 

 ing groups. Every argument, every reading of Maori and other tradi- 

 tion, points clearly to that region as such a distributing center of the 

 later Tongafiti migration as Samoa has been for the Proto-Samoan 

 wanderers. But we lack dictionary provision and must content our- 

 selves with the Maori. 



Since the subject has arisen for consideration we may, before passing, 

 note the geography of that mid region. To the west lies Nuclear Poly- 

 nesia, which I set apart as a linguistic province in an earlier work upon 

 this theme. To the east lies this province of Southeast Polynesia whose 

 essential unity is established in the course of the studies recorded in 

 this volume. Spread in the intervening sea lie the Cook and Austral 

 groups, together with lesser islets, in which the Tongafiti character is 

 well marked, and in which such research as I have been able to conduct 

 has revealed very scanty stock of distinctively Proto-Samoan material. 

 These islands undoubtedly became the principal home of the Tongafiti 

 after the Matamatame onfall drove them from Samoa. It is from 

 them that the voyages of discovery and voyages of rediscovery carried 

 them to Aotearoa, which lies upon our charts as New Zealand, where 

 they found some population of Proto-Samoans who had voyaged thither 

 direct from Samoa and whom in time they reduced to subjection but 

 not to linguistic extinction. From the same central oceanic base the 

 Tongafiti passed to the nearer archipelagoes of Southeast Polynesia in 

 whose five languages we are now examining their condition. From the 

 same base, either directly or proximately through Tahiti and the Mar- 

 quesas, we find that they reached Hawaii, and there, far in the north, 

 they subjected a prior population of Proto-Samoans, but not to lin- 

 guistic extinction. Here in Southeast Polynesia we find the two stages 

 of the language existing without mixture in equal streams. We shall 

 find pleasure and profit in studying out, so far as we may, the evidences 

 of original colonization and secondary distribution by movements of 

 convection within the province. But we must confess that in this 

 branch of the investigation we are hampered through the failure to 

 establish the Tongafiti mother speech to serve as the standard of com- 

 parison in deviation of sense, for with form as phonetics we are to deal 

 very lightly. 



The Rapanui variants in the Tongafiti class exhibit a slightly smaller 

 percentage than those in the Proto-Samoan class. This might be a 

 matter of greater value if the two comparisons were more equal. We 

 should expect, moreover, to find such a deviation founded upon the 

 marked difference in the age of each migration source within the Pacific. 



