A Racial Study of the Fijians

Illustration: Simplified map of Fiji showing four regional divisions of population made by the author.

Simplified map of Fiji showing four regional divisions of population made by the author ... frontispiece

The data used in this survey were secured in 1954 during a stay of seven months in Fiji. My plan was to obtain anthropometric samples from several parts of the archipelago; this plan was only slightly altered as time and transportation facilities directed. Each of the three main administrative districts into which the islands are divided were visited and within each district samples were secured from most of the constituent provinces. The original sample consisted of 880 subjects. Later, 65 subjects were excluded for various reasons: some were part Samoan or Tongan, a few were Rotumans, and others were immature. The number finally used stands at 815.
A limited amount of comparative material has been included in order to help locate the Fijians in the overall Pacific picture. These data were drawn from W. W. Howells, Anthropometry and Blood Types in Fiji and the Solomon Islands in The American Museum of Natural History, Anthropological Papers, volume 33, part 4, 1933, and from L. R. Sullivan, A Contribution to Tongan Somatology based on the field studies of E. W. Gifford and W. C. McKern, in Memoires of the Bernice P. Bishop Museum, volume 8, number 4, 1922. The latter report provides comparison with what may be termed western Polynesians who are also the nearest Polynesians to the Fijians. The Fijian data in Howell's paper make it possible for me to check some of my own Fijian material, and the Solomon Island data in the same report provide a Melanesian measuring stick.
Since an over-all description of the Fijians is the initial concern of this paper, each physical trait measured or derived from measurement is tabulated according to range, average, and deviation. Traits observed but not measured are presented according to degree of development, e.g., absent, medium, and pronounced, and according to percentage of occurrence. Further statistical manipulation is not deemed necessary for the writer's purposes.