138 



EASTER ISLAND. 



The sum of all these partial results is set forth in the following table, 

 still retaining the primal division on the Rapanui base and bringing 

 forward the corresponding sums from Table 14 on page 113: 



Table 22. 



In Bishop Dordillon's dictionary we count 12,000 entries. This is 

 gross exaggeration. It has already been noted that there are two dia- 

 lects; the dictionary is clearly a loose compaction of both. It is quite 

 safe to say that practically there is duplication of every item, that its 

 form in one dialect and its form in the other dialect double the entries. 

 It is inconceivable that the Marquesans possess twice the speech equip- 

 ment which the research of careful investigators for more than a century 

 has been able to discover for Tahiti. The fact that the count of the 

 Marquesas speech is almost exactly double that of Tahiti is evidential 

 that the explanation of the discrepancy lies in the agglomeration of two 

 dialects, each of about the same figure as that of Tahiti. Since the 



Table 23. 



author's designation of dialectic diversity is by no means consistently 

 or completely carried out in that dictionary, it is impossible to collate 

 the text for the enumeration of each dialect. We shall, however, not 

 go very far astray in assuming for the base figure in the computations 

 upon which we are next to engage that which we employed in Tahiti, or 

 the round number of 6,000. 



We have developed in the Marquesan the identification of i,886 affili- 

 ates, 31 per cent of the language; with this we compare identification in 



