262 DIXON— A NEW THEORY OF POLYNESIAN ORIGINS. 



these differences as due to social and economic causes, others as the 

 result of different waves of immigrants. For the most part the 

 investigators confined their attention to portions of the field only, 

 studying the Hawaiians or the Maori for example, rather than in- 

 cluding all of the Polynesian peoples ; and largely relied, further- 

 more, on averages of measurements and indices for their results. 

 In consequence the situation remained obscure, correlation between 

 different portions of the Polynesian area, and between it and the 

 rest of Oceania was difficult, and while the whole problem was 

 seen to be growing in complexity, no satisfactory general theory was 

 possible. 



Having for many years been greatly interested in the whole 

 question of the origins and development of the Oceanic peoples, and 

 having on the basis of existing data, attempted most unsatisfactorily 

 to explain the whole matter to students, the author finally came to 

 the conclusion that only by a revision and reinvestigation of all 

 available data, on somewhat different lines, could the muddle be 

 cleared up. Accordingly, all accessible measurements, not only of 

 Polynesians but of all the people of Oceania together with those of 

 southeastern Asia, were gathered together and analyzed on what, 

 at least for this area, was a novel plan. The final results of the 

 whole study are not yet complete, but the analysis of the Polynesian 

 data has led to such unexpected and interesting, yet at the same time 

 logical and coherent results, that their brief presentation seemed 

 desirable. 



Before outlining the conclusions reached, a few words must be 

 said as tO' the method employed. For the most part in previous 

 studies, attention has been mainly directed to the cephalic index, or 

 if other indices and measurements were considered, little or no 

 attempt was made to study the correlation of these indices in 

 individual skulls. In the present investigation, a correlation was 

 made for the cephalic, length-height and nasal indices, with the addi- 

 tion of the facial index where it was available. Thus the accessible 

 series of Hawaiian crania was analyzed into groups, one compris- 

 ing all skulls that were for instance Dolichocephalic and at the same 

 time Hypsicephalic and Platyrrhine, another including all that were 



