46 EASTER ISLAND. 



lists in accordance therewith. I omit those occurrences, readily dis- 

 coverable in the vocabulary, of h in Rapanui where no comparable data 

 are available, and also the vocables which are entered in the vocabulary 

 both with and without the aspirations. The first table records the items 

 in which comparable data go no further than the province of Southeast 

 Polynesia, and where, accordingly, we may not establish a determinant 

 comparison : 



We note these few instances in which the Rapanui aspirate is clearly 

 labial : 



43 51 52 59 205 206 209 



In the next group of tables we have the advantage of data for com- 

 parison, for these items are drawn from the distinctively Proto-Samoan 

 element of the language. 



First we shall examine those items in which we have been able to 

 establish in the Proto-Samoan a pure aspirate. In Rapanui this is in 

 some cases dropped, in others retained, and in the following table the 

 instances of the preservation of that fickle sound are distinguished by 

 bold-faced type: 



302 305 324 345 351 359 4J8 657 692 695 



Our next table records the instances in which the Proto-Samoan 

 sibilant passes into the lingual aspirate: 



329 358 364 374 377 381 384 392 399 401 409 472 547 680 

 355 361 371 376 380 383 390 394 400 408 416 473 624 



In by far the greater number of cases the Rapanui aspirate is a 

 mutation product of the Proto-Samoan/, therefore a labial aspiration, 

 as shown in the items of this table : 



We next consider the element whose source is found in that insepar- 

 able mass of the Polynesian common to both migrations. Inasmuch 

 as this affords us Proto-Samoan comparatives it might have been incor- 

 porated with the foregoing tables, but since the division has been of 

 value in the consideration of other topics it is here maintained for the 

 sake of uniformity. 



There are but two items which bear upon the Proto-Samoan aspirate, 

 in 792 Rapanui retains it and in 833 it discards it. 



