NEW RADIOCARBON AGES FOR TWO KEY SITES of East Polynesian prehistoric colonization are presented and discussed in relation to additional results from recent research. Before coming to these it is necessary to specify the methodological stance adopted here concerning the definition of colonization and its visibility in radiocarbon chronologies. Colonization is a concept grasped more easily in the abstract than it is disclosed empirically by archaeological or related data. In the context of East Polynesian prehistory it expresses the general idea of earliest human settlement, but whether that means the first human contact of any kind, the first occupation, or the first permanent settlement is open to question on several counts discussed elsewhere (Anderson 1995; Graves and Addison 1995). Here we can deal only with sites that have been identified as belonging to the colonization era on the grounds that they contain examples of the earliest artifactual types known in East Polynesia and generally disclose a characteristic faunal signature of remains from extinct birds, marine mammals, and turtles (e.g., Leach et al. 1984; Steadman and Rolett 1996).
We accept that colonization need not have been an instantaneous event and that the probability of the earliest actual sites of colonization being represented amongst currently known sites is relatively low in general. Nevertheless, in detailed analysis of early radiocarbon ages, we prefer to emphasize provenance by using indubitably archaeological samples and by taking a site-by-site approach ("chronometric hygiene," e.g., Anderson 1991; Spriggs and Anderson 1993). There is an acknowledged risk of eliminating the presumed few and perhaps poorly provenanced results that may represent the earliest stage of colonization. However, this seems less problematic than the potential difficulties inherent in methods that rely upon inspection of trends in large and indiscriminate assemblages of potentially relevant radiocarbon determinations (e.g., Anderson 1989; Graves and Addison 1995; Hunt and Holsen 1991). These characteristically exhibit a "tail" of data extending earlier than it is possible to date any particular archaeological site. Dates with poor provenance, or inadequate laboratory pretreatment, and samples with high inbuilt age or of nonarchaeological origin, such as old soil charcoals from natural fires, may account anonymously for the oldest results. Our conservative preference enjoins the pragmatic consequence that the earliest archaeological sites identified as belonging to the colonization phase might not represent the beginning of it.
CENTRAL EAST POLYNESIAN ARCHAEOLOGICAL CHRONOLOGIES
The consensus in East Polynesian colonization chronology since the 1960s is that, whatever the actual ages, the settlement of central East Polynesia preceded habitation at the margins, and that the temperate region was colonized last. Only suggestions of Amerindian landfalls in the eastern margins have offered any challenge, and that never satisfactorily realized. How early East Polynesia was colonized is a question thus narrowed to the chronology of the central region and focused upon a handful of archaeological sites which, though by no means the only sites in the region to produce chronological, faunal, and artefactual evidence indicative of early occupation, have stood for 30 years or more as key sites in the ongoing debate. They are: Motu Paeao cemetery (Maupiti) and Vaito`otia-Fa`ahia (Huahine) in the Society Islands, and Hane (Ua Huka) and Ha`atuatua (Nuku Hiva) in the Marquesas Islands.
The early chronological research on these sites began with Suggs (1961) who dated Ha`atuatua to as early as 150 13.c. (in discussing earlier radiocarbon results we give the calendrical ages estimated at the time--see Kirch  for a summary of the early research, Spriggs and Anderson  for recent calibrated ages of older determinations). …