The 13th century polynesian colonization of Hawai’i Island more

Journal of Archaeological Science 38 (2011) 2740e2749 Contents lists available at ScienceDirect Journal of Archaeological Science journal homepage: http://www.elsevier.com/locate/jas The 13th century polynesian colonization of Hawai’i Island Timothy M. Rieth a, *, Terry L. Hunt b, Carl Lipo c, Janet M. Wilmshurst d a International Archaeological Research Institute, Inc., 2081 Young Street, Honolulu, HI 96826, USA Department of Anthropology, University of Hawai’i-Manoa, 2424 Maile Way, Honolulu, HI 96822, USA c Department of Anthropology and IIRMES, California State University Long Beach, 1250 Bellflower Blvd., Long Beach, CA 90840, USA d Landcare Research, PO Box 40, Lincoln 7640, New Zealand b a r t i c l e i n f o Article history: Received 29 March 2011 Received in revised form 16 June 2011 Accepted 18 June 2011 Keywords: Hawai’i East Polynesia Radiocarbon dating Colonization a b s t r a c t We assess 926 radiocarbon dates from Hawai’i Island, the largest assemblage of dates compiled from a single island in Oceania. Based on a classificatory approach that arranges the dates based on their reliability, accuracy, and precision, our results indicate that the most reliable estimate for the initial Polynesian colonization of Hawai’i Island is AD 1220e1261, w250 to 450 years later than the current consensus. This conclusion is strikingly convergent with recent estimates for the colonization of remote East Polynesia. Our analysis highlights the need for wood charcoal identification to insure selection of short-lived plants/plant parts for radiocarbon dating, and that a reliance on dating unidentified wood charcoal is a waste of resources that only serves to retard progress in refining the settlement chronology of Hawai’i Island and other locations. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction Since its advent in the 1950s, radiocarbon dating has played a central role in building cultural chronologies. In Polynesia, the method gained importance since many of the archipelagos lacked pottery e a class of artifacts that had proven critical to building chronologies elsewhere based on its facility in seriation analyses (Dunnell, 1978). Despite the strong reliance on radiocarbon dating, the resulting chronologies for Polynesia, and for the Hawaiian Islands in particular, have seen dramatic shifts in the assignment of dates for colonization. The findings of different investigators have varied by a thousand years or more, leading to enduring controversy that has defied easy resolution. As chronologies have changed, or as different investigators have solidified their own chronometric beliefs, so too have models for explaining colonization, population growth, human impacts on island ecosystems, changes in resource use, production systems, and the evolution of social complexity, among other dimensions of prehistory. A robust chronology thus forms a critical foundation to reliably resolving multiple research questions. It has been a quarter century since the radiocarbon chronology for the Hawaiian archipelago has been assessed (Hommon, 1986; Kirch, 1985, 1986; see also Dye, 1994a, and Dye and Komori, * Corresponding author. Tel.: þ1 808 946 2548x115; fax: þ1 808 943 0716. E-mail address: trieth@iarii.org (T.M. Rieth). 0305-4403/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jas.2011.06.017 1992). Over this interval there has been an exponential increase in the number of radiocarbon dates produced. As this radiocarbon corpus has grown, the general consensus on the age of initial settlement of the Hawaiian Islands has varied, but with a general trend of getting progressively shorter from presumed colonization in the first centuries AD, to a range from approximately AD 700e1000 (Athens et al., 2002; Dye and Pantaleo, 2010; Kirch, 2007; Masse and Tuggle, 1998). The more recent and shorter chronologies have been based largely on paleoenvironmental investigations of wetland and lacustrine sediment cores for indications for changes in fire history and plant communities, as well as, radiocarbon dating of commensal Polynesian rat bones (Athens, 1997; Athens et al., 1995, 1999, 2002; Burney, 2002; Burney and Burney, 2003; Burney et al., 2001; see also Masse and Tuggle, 1998 for a similar chronology derived from archaeoastronomy). This emphasis on reconstructing environmental history has resulted in changes in our overall chronologies. Previously, however, with few exceptions (e.g., Anderson, 1991; Hunt and Lipo, 2006; Wilmshurst et al., 2011a), revised chronologies for Hawai’i and other islands of East Polynesia have not relied on systematic, critical reviews of large suites of dates, but on the exclusion of unreliable early dates from particular sites previously identified as representative of early settlement (e.g., Kirch and McCoy, 2007). Barring a rigorous review of a large assemblage of dates, most settlement estimates have merely “crept” forward in time. One of the reasons for many of the dating inconsistencies and resulting controversies clearly derives from attempts to build chronologies T.M. Rieth et al. / Journal of Archaeological Science 38 (2011) 2740e2749 2741 with radiocarbon dates acquired from (1) unidentified wood charcoal with an unknown amount of potential in-built age, (2) dates on bone that may be subject to contamination or dietary effects, and which are often difficult to calibrate precisely, and (3) dates from marine shell or bone that require marine reservoir corrections (delta r) which exist for some locations but may be highly variable depending on various factors (Petchey, 2009). Wilmshurst et al. (2011a) have also shown that high-precision radiocarbon dating has resulted in the overall chronology for East Polynesia becoming much shorter. They report a meta-analysis of radiocarbon dates from across the archipelagos of East Polynesia and document the relatively recent and rapid colonization of this vast region. Remarkably, far-flung locations in East Polynesia provide nearly simultaneous ages for colonization when their assessment standards of reliability and precision are employed. They show that approximately AD 1200e1290 is a reliable measure of when people first arrived in these islands, including the Hawaiian archipelago. Wilmshurst et al. (2011a) demonstrate that this pattern of consistent ages for regional colonization is not a result of sampling effects, but more to do with archaeologists applying minimal standards for reliability. The conclusion of late and rapid colonization of the region can be further tested with additional sets of reliable radiocarbon dates from the largest island in the Hawaiian archipelago. To this end, as the first island-specific test of conformity to this pattern, we assess 926 radiocarbon dates from Hawai’i Island, the largest assemblage of dates yet compiled from a single island in Oceania. We include previously unassessed data sourced from the unpublished “gray literature” (c.f. Ford, 2010), i.e., work generated from commercial archaeology undertaken on Hawai’i Island, greatly expanding the total number of dates available for the island compared with Wilmshurst et al (2011a). Our new analysis addresses two fundamental substantive and methodological research issues: (1) the timing of initial Polynesian colonization of Hawai’i Island, and by extension the Hawaiian archipelago, and (2) the effects of radiocarbon sample selection and potential in-built age in radiocarbon samples. These issues are assessed in relation to identifying the timing and development of post-colonization events such as population increase, the development of agricultural complexes, and other chronologically based inferences for prehistory. We define colonization by empirical standards, that is, by radiometric dates that are clearly associated with human activity. Expectations about colonization assemblage characteristics, such as an abundance of extinct/extirpated avifauna, the remains of large marine taxa (c.f. Hunt and Lipo, 2006), the presence of exotic materials, and/or archaic artifact forms (c.f., Walter, 1996), can clearly inform on the potential age of a deposit, but cannot conclusively allow it to be classified as being early without radiometric support. Similarly, if the earliest reliable radiometric dates are not associated with an expected “colonization” assemblage, it is equally unsubstantiated to assume the dates do not represent colonization. This line of reasoning would allow one to continually argue that colonization occurred at some earlier, yet empirically undefined, date. By definition, the earliest reliable dates for a human presence record colonization, a conclusion that is falsifiable. Otherwise, archaeologists risk making circular and untestable arguments. 2. Previous archaeological models of Hawaiian chronology The Hawaiian archipelago consists of eight main islands configured in a northwest-southeast island arc created by movement of the Pacific Plate across a hotspot (Fig. 1). Hawai’i Island is the southern-most, largest, and geologically youngest island. The Fig. 1. Map of the Hawaiian Archipelago. 2742 T.M. Rieth et al. / Journal of Archaeological Science 38 (2011) 2740e2749 island is 16,636 km2 in area and includes diverse ecozones ranging from tropical moist broadleaf forest to alpine desert. Mauna Kea (4205 m) and Mauna Loa (4170 m) are the highest peaks of the five volcanoes that form the island. As a consequence of the prevailing northeast trade winds and the mountainous center of the island, distinctive wet windward and dry leeward regions divide the island. The quantity and frequency of rainfall varies markedly between these regions. In geologically older portions of the island, windward zones are characterized by narrow, incised valleys with permanent stream flow suitable for irrigated agriculture, while leeward zones typically have more gradual, arid slopes with intermittent drainages, leaving only options for rain-fed agriculture. By the time of European contact in 1778, Hawai’i Island was  divided into six political districts (moku): Kohala, Kona, Ka’u, Puna, Hilo, and Hmakua. Each district was subdivided into smaller a  geographical units termed ahupua’a, established as economically semi-autonomous communities. Detailed summaries of 19th and 20th century historical and archaeological considerations for the chronology of the Hawaiian Islands have been provided elsewhere (Cordy, 2000; Dye, 1989; Kirch, 1985, 1986), and we briefly summarize them here. Early 20th century archaeological research was focused on “the origins, migrations, and external contacts” of Hawaiians with research directed primarily at analysis of surface architecture (National Research Council, 1921: p. 117; see also Dye, 1989). The supposition at the time was that excavation would be unproductive given poor preservation in the tropics and limited change in material culture over time (Dye, 1989). Challenging these assumptions, excavations by the Bishop Museum and the University of Hawai’i during the 1950s and 1960s revealed relatively deep, stratified cultural deposits at a number of locations. Significantly, samples from several of these locations were submitted for what was then the new chronometric method of radiocarbon dating. Among the results were ages suggesting a significant time depth for settlement of over a millennium for the Hawaiian Islands (Emory and Sinoto, 1961, 1969; Libby, 1951: p. 295). These findings challenged earlier assumptions. Radiocarbon results coupled with changes documented in fishhook forms from cultural deposits on Hawai’i Island (Sites H1 and H8) led researchers to believe that the earliest settlement of the archipelago dated to the first centuries AD (Emory and Sinoto, 1969). Subsequent excavations at Bellows, O’ahu Island (Site O18; Pearson et al., 1971) and the Halawa Dune site on Moloka’i (Kirch and Kelly, 1975) seemed to corroborate these age estimates. Kirch (1985, 1986) and Hommon (1986) provided the first syntheses of larger radiocarbon suites for the Hawaiian Islands, reviewing overlapping assemblages of 254 dates and 163 radiocarbon dates, respectively. Based on contemporary ideas about the timing of the colonization of the Marquesas and Society Islands, perhaps as early as the last centuries BC (Kirch, 1986), an early settlement for Hawai’i seemed certain. Despite some early e and indeed relevant e concerns (e.g., Emory and Sinoto, 1969), researchers generally accepted the suites of dates from Site H1 on Hawai’i Island and from Site O18 on O’ahu. However, for each site dates were measured from samples of unidentified wood charcoal (as well as shell, bone, and sea urchin) and included internally inconsistent dates lacking stratigraphic integrity. Based on these results, and those that would follow from extensive field research at Halawa Valley (Moloka’i), Kahana and Kawainui (O’ahu), and from Wai’ahukini Rockshelter (Hawai’i Island), Kirch (1985) proposed that the colonization of the archipelago occurred around AD 300. Until the 1990s, this view received general consensus, with some (e.g., Hunt and Holsen, 1991) suggesting the possibility for even earlier settlement. Eventually, researchers began to question the archaeological basis for an approximate AD 300 chronology with additional dating and critical reexaminations of earlier results from O18, H1, and Halawa Valley (Dye, 1992; Dye and Pantaleo, 2010; Kirch and McCoy, 2007; Tuggle and Spriggs, 2001). Tuggle (1997: 48) questioned the Bellows dates and stated the “dates as a whole are considered completely unreliable: they involve stratigraphic inversion, a modern date, extremely large standard errors, dubious laboratory data, and questionable proveniences.” For Bellows, Dye and Pantaleo (2010) analyzed the results of nine new radiocarbon determinations from short-lived plant taxa as well as pearl shell using a Bayesian model. Their results indicated a chronology starting later than previously claimed, dating to around AD 1040e1219.1 Dye’s (1992) re-analysis of the Pu’u Ali’i radiocarbon dates using annual frequency distributions suggests initial occupation in the early 15th century. For a second well-known “early” site, Kirch and McCoy (2007) suggest an age no older than AD 1300 for the Halawa Dune deposit, based on six new Accelerator Mass Spectrometry (AMS) radiocarbon dates. At an island-scale, only Moloka’i (McCoy, 2007; Weisler, 1989) and Kaua’i islands (Carson 2005) have seen recent radiocarbon syntheses. For the Kaua’i review, Carson (2005) applied a general “chronometric hygiene” (e.g., Rieth and Hunt, 2008; Spriggs and Anderson, 1993) procedure to reject dates that were poorly provenienced or obtained from highly questionable samples. McCoy’s (2007) synthesis for Moloka’i did not entail such an approach, but both suites of dates nonetheless provide valuable datasets for comparison. Both island assemblages include remarkably small numbers of samples that date older than about AD 1200, none of which come from samples of short-lived plant taxa. In both island sets, the vast majority of samples are associated with dates that fall after AD 1400. These results are in general concordance with Athens et al.’s (2002) review of 194 radiocarbon dates from the ‘Ewa Plain of O‘ahu, which represents the largest published radiocarbon dataset for that island. Re-evaluations of allegedly early cultural deposits across the Hawaiian Islands and results from the two island-wide syntheses indicate the need to critically assess the burgeoning corpus of dates for the archipelago. The proliferation of cultural resource management (CRM) archaeology across the islands has produced the overwhelming majority of radiocarbon dates. As noted for other regions in the Pacific (Liston, 2005 for Palau; Rieth and Hunt, 2008 for Smoa), as well as the United Kingdom (Ford, 2010), synthesis of this a “gray literature” is essential for furthering our understanding of prehistory. Hawai’i Island, which has played a significant role in the acceptance of an early settlement date for the archipelago, has also seen a significant amount of academic and CRM-mandated archaeological research over the last several decades. As a consequence, chronological data from this island are essential for evaluating the timing of colonization and the chronology of Hawaiian settlement. 3. Methods: assessing radiocarbon dates using a classificatory approach We use a method of classification (following Wilmshurst et al., 2011a used for East Polynesia) to assess the reliability, accuracy, and precision of radiocarbon dates (Barry, 1978: 15). Just as specific research questions guide the development of artifact and feature classifications, we suggest a similar approach is warranted for radiocarbon age determinations. Our primary interest is the timing of initial Polynesian colonization of Hawai’i Island, and by extension, its implications for settlement of the entire archipelago. Answering this question requires a high degree of reliability with 1 Dye and Pantaleo (2010) bracket their Bayesian model with a postulated settlement date for the Hawaiian archipelago at AD 800 Æ 50. Introducing a different settlement age into the model will produce different results. T.M. Rieth et al. / Journal of Archaeological Science 38 (2011) 2740e2749 2743 radiocarbon-dated samples accurately representing archaeological target events (Dean, 1978; as distinguished from the radiocarbon event). Colonization events with relatively discrete timing require short probability age distributions to minimize the “noise” within the calibrated age that may obscure recognition of the best estimated age of the archaeological event. We divide the suite of Hawai’i Island radiocarbon dates into two primary groups: conventional radiocarbon ages (CRA) that are !400 BP, and CRA that are 399 BP. We chose !400 BP for the division as it was the earliest possible age allowing for a sufficient number of dates for statistical evaluation. Wilmshurst et al. (2011a) analyze dates reported to be !300 BP for the purposes of broad comparability across multiple East Polynesian island groups with sufficient radiocarbon dates. Since our analyses focused solely on Hawai’i Island, we were able to narrow our analyses to an earlier point in time given the large number of dates. Our choice to include only older dates reflects the relative abundance of dates for Hawai’i and thus the greater precision and statistical confidence of the analytic results. While a more restrictive CRA (e.g., !700 BP) for creating the pool of oldest dates is possible, this criterion would greatly decrease the overall sample size, thus introducing statistical error. Here, we did not include the 399 BP group in our analysis, although we list them in the online supplement (Table S1). 3.1. Classification for CRA !400 BP Dates are assumed to derive from samples with archaeological proveniences, unless otherwise noted by the original investigators. Following Wilmshurst et al. (2011a), we categorized all samples by sample material: (1) short-lived plant taxa or plant parts, (2) longlived plant taxa, (3) unidentified charcoal or wood, (4) fully terrestrial (ultrafiltered or XAD-2 resin) bone, (5) marine influenced bone, (6) marine shell and other marine invertebrates, (7) coral, (8) bulk soil/ ash, and (9) mixed short-lived plants and unidentified charcoal. We then organized the dates into three reliability classes (Fig. 2).  Class 1 dates are obtained from identified short-lived plant taxa or plant parts (Table 1) and the standard error for the CRA is 10% of the age determination. Dates from terrestrial bone subject to ultrafiltration or XAD-2 resin extraction that have a standard error which is 10% of the age determination are also classified as Class 1. Dates within this class are justified to be reliable ages based on good target material. 100%  Class 2 dates comprise dates obtained from unidentified wood charcoal (including those from mixed samples of identified plant taxa and unidentified charcoal) that have a standard error which is 10% of the age determination, identified long-lived plant taxa, or are from identified short-lived plant taxa or plant parts but have a standard error that is >10% of the age determination. Dates from marine shell, coral, and marine animal bone (including seabirds) are also included in this class because of the current uncertainty in the apparent localized variation in the delta r for Hawai’i (Petchey, 2009). These dates may be subject to an unknown degree of in-built age or lack a sufficiently short calibrated age distribution for identification of a colonization event.  Class 3 dates include dates from unidentified wood charcoal, shell, and bone that have a standard error which is >10% of the age determination and all ages from bulk soil/bulk ash and soil samples. Results that are reported as measured radiocarbon ages without the adjusted CRA are also within this class. Additionally, radiocarbon samples obtained from deposits in features identified as irrigation channels (‘auwai’) are included in Class 3, since given the dynamic depositional history, it is unclear what archaeological events these samples date. Class 3 samples are subject to an unknown degree of in-built age and/ or may not link the radiocarbon event with the target archaeological event. Requiring a standard error for a CRA that is 10% maintains a high degree of precision. Admittedly the standard error of a CRA does not always directly correspond to the span of the calibrated age, as “flat areas” of the calibration curves may result in longer, multi-modal age distributions for CRA errors that are relatively small (e.g., McFadgen et al., 1994). However, in most instances the correspondence between the CRA standard error and the length of the calibrated probability age distribution warrants use of the former as a proxy for the latter. Additionally, w10% error is a typical standard result for AMS dates processed since c. 1995, when it also became more common for wood charcoal to be identified to genus or species. Thus the two factors combined result in Class 1 designations for the majority of dates from identified short-lived specimens. Increasing the acceptable amount of error will mechanically result in dates with broader ranges, though these values will only reflect additional uncertainty and not longer chronologies, as some might claim (see Wilmshurst et al., 2011b). Short-lived plant 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Class 1 Class 2 Class 3 Long-lived plant Unidentified charcoal Terrestrial bone Marine bone Marine shell & invert. Coral Bulk soil/ash Mixed ident. & unid. charcoal Fig. 2. Percentage of sample materials making up each overall reliability class. 2744 T.M. Rieth et al. / Journal of Archaeological Science 38 (2011) 2740e2749 Table 1 General age categories for Hawaiian plant taxa commonly used as radiocarbon dating samples. Species Aleurites moluccana nutshell Bidens Bobea sp. Chamaesyce sp. Chenopodium oahuense Coprosma sp. Cordyline fruticosa fern caudex Ipomea batatas Lagenaria siceraria Nototrichium sp. Osteomeles anthyllidifolia Pipturus albidus Psychotria sp. Railliardia sp.¼ Dubautia Rauvolfia sandwicensis Senna sp. Sida fallax Styphelia tameiameiae Wikstroemia sp. Acacia koa Aleurites moluccana Artocarpus altilis Diospyros sandwicensis Dodonaea viscosa Metrosideros polymorpha Myoporum sandwicense Pritchardia sp. Santalum sp. Sophora chrysophylla 1 to w50 years x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x >50 years 3.2. Calibration and statistical methods Conversion of radiocarbon ages into calendrical dates requires calibration with respect to the temporally variable amount of 14C in the atmosphere. For the samples compiled in our study, we used INTCAL09 calibration (Reimer et al., 2009). We then evaluated the calibrated Class 1 dates for the earliest evidence of human occupation in terms of two kinds of hypotheses. The first hypothesis consists of determining the earliest date before which we can reliably state that the colonization occurred. This date is simply a statistical statement that identifies the point at which it is more likely that the event of interest occurs earlier rather than later, and is referred to as the Late Age Estimation Model (LAEM) (see Wilmshurst et al, 2011a). The second hypothesis centers on identifying the date after which we can confidently state that the colonization event must have occurred. Given the probabilities of individual calibrated radiocarbon dates, there is a non-zero, though usually small, probability that the actual event took place on the very earliest part of the distribution. The use of values from multiple independent dates establishes the point at which statistical confidence is sufficient to state that the colonization event occurs at this point or after, referred to here as the Early Age Estimation Model (EAEM). To evaluate the first hypothesis, we sum the individual probability distribution from each Class 1 calibrated radiocarbon date to form a single aggregate distribution. The premise behind this operation is that the greater the amount of overlap of probability from each point in time in a sample’s probability distribution, the greater the overall likelihood the event took place at that time. We can then calculate the cumulative probability and measure how confidence changes as a function of time and the distributions of the summed probabilities. To do this, we have to arbitrarily pick a point in time by which we consider the colonization event to have occurred with 100% certainty, which for Hawai’i Island, we assume to be AD 1300. An estimated age for the earliest probable colonization can be determined using the Early Age Estimation Model (EAEM). This point in time represents the age at which the aggregate probability distributions form a significant positive, non-zero slope due to the numbers of overlapping probability values from multiple Class 1 dates. In the case of Hawai’i Island, the relatively few dates precluded meaningful statistical evaluation, so we approximated the EAEM age. Additional dates will increase the precision of the EAEM determination. In addition to specifying precision of the laboratory results, sample choice must also be made to ensure a high degree of accuracy (i.e., a ‘true’ age of the sample). This requirement is the predicate for the exclusion of unidentified wood charcoal (see Dye, 2000). An unidentified charcoal sample from a long-lived tree with significant in-built age could produce a precise radiocarbon result (i.e., with a standard error 10% of the CRA) but it may not be an accurate measure of the archaeological (target) event of interest. Similarly, a recent review of delta r measurements for the Hawaiian archipelago identified a wide range in values (Petchey, 2009; see also Dye, 1994b). Until localized delta r values are determined for multiple areas of the archipelago, dates from marine invertebrate and vertebrate remains must be reviewed with caution as they add unknown degrees of uncertainty. Even when delta r values can be specified, problems persist in calibrating these dates given the necessary assumption that the materials dated originally came from the local environment measured for delta r values. Unlike chronometric hygiene approaches (e.g., Anderson, 1991; Fitzpatrick, 2006; Hunt and Lipo, 2006; Rieth and Hunt, 2008; Spriggs and Anderson, 1993), which are designed for the acceptance or rejection of dates based on specific criteria, the approach we use here presents a means of assessing and categorizing the relative reliability of radiocarbon dating results. The distinction between the protocols used in our classificatory approach versus chronometric hygiene is significant. The less subjective classificatory method raises specific challenges for the usefulness of certain types of sample material used for dating archaeological events, and future researchers will need to directly address these issues if they rely upon such dates for chronology building. Additionally, our approach is a complementary primary step for more detailed secondary analyses such as calibration using a model-based Bayesian framework (e.g., Dye and Pantaleo, 2010). For these reasons, we used only Class 1 dates in our analysis. 4. Classification results We tabulated 303 radiocarbon dates with CRA !400 BP, of which 16 are Class 1 (Table 2). Five additional dates were obtained from suitable samples but had error ranges slightly greater than 10%, and therefore are categorized as Class 2. Inclusion of these samples, which range from 450 to 550 Æ 60 BP, would not change our results. Eighty-six dates are included within Class 2, while the remaining 201 dates are categorized in Class 3. The reliable Class 1 dates produce a cumulative probability estimate for colonization from AD 1220 (EAEM) to AD 1261 (LAEM) (see Fig. 3). Additional Class 1 dates would allow us to refine the age estimations providing even greater precision for the age of colonization. These results are w500 to 550 calibrated years later than the earliest Class 2 date,2 with its inherent level of uncertainty. Accepting the less reliable Class 3 dates would suggest (with a low degree of 2 Beta-32084 is identified as the earliest Class 2 date since Beta-43670 (10,440 Æ 70) is well beyond any arguments with respect to a target of human colonization. T.M. Rieth et al. / Journal of Archaeological Science 38 (2011) 2740e2749 Table 2 Class 1 radiocarbon dates. District Ahupua‘a Site Sample category 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 14 2745 C Lab ID CRA CRA Initial Measurement Measured From Overall Citation for source years error quality error >10% age not CRA ‘Auwai reliability BP class class 640 440 440 400 400 568 463 781 696 450 430 440 420 580 760 430 40 40 40 40 40 38 31 38 35 40 40 40 40 40 60 40 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Williams (2002) Williams (2002) McCoy and Graves (2010) McCoy and Graves (2010) Field et al. (2010) Field and Graves (2008) Field and Graves (2008) Field and Graves (2008) Field and Graves (2008) Ladefoged and Graves (2008) Ladefoged and Graves (2008) Ladefoged and Graves (2008) Ladefoged and Graves (2008) Ladefoged and Graves (2008) Robins et al. (2000) O’Day and Rieth (2007) Hamakua Hamakua North Kohala North Kohala North Kohala North Kohala North Kohala North Kohala North Kohala North Kohala North Kohala North Kohala North Kohala North Kohala South Kohala South Kohala Ka‘ohe Ka‘ohe Halawa Halawa Kalala Pololu Pololu Pololu Pololu Unclear Unclear Unclear Unclear Unclear Ouli Waimea 50-10-31-21286 50-10-31-21283 50-10-02-26807 50-10-02-26061 KAL-1 4916 4908 4916 4916 T-12 T-21 T-7 T-7 T-50 50-10-05-14751 50-10-06-21873 Beta-135126 Beta-135125 Beta-233042 Beta-233043 Beta-256577 Wk-19312 Wk-19313 Wk-19311 Wk-19310 Beta-189737 Beta-189745 Beta-189730 Beta-189731 Beta-208143 Beta-135829 Beta-219668 confidence) that settlement could have begun during the early first millennium BC. The effects of in-built age are readily apparent when the dates are reviewed by sample category (Fig. 4). Dating a mixture of shortlived plant taxa and unidentified charcoal suggests settlement w50 to 100 calibrated years earlier, while the dating of long-lived plant taxa and bulk soil/ash results in the perception that colonization could have occurred w200 to 250 calibrated years earlier. An uncritical approach to a synthesis of results from dating unidentified wood charcoal, by far the most common sample material and the material often classified as Class 3, would raise the possibility of settlement nearly 2000 calibrated years earlier. Reviewing our data by date class and sample material reveals the systematic effect that using unreliable radiocarbon dating results has in pushing the age for colonization back in time. Geographically, all but two of the Class 1 dates are from the Kohala District (Fig. 5). The assemblage of radiocarbon dates is weighted towards the leeward portion of the island, with Kona  (n ¼ 86), Kohala (n ¼ 74), and Ka‘u (n ¼ 51) providing nearly 70% of the total collection. A large number of dates have been obtained   from the Pohakuloa region of Ka’ohe Ahupua’a within the Hamakua District (n ¼ 77). The Hilo (n ¼ 2) and Puna (n ¼ 13) Districts are not as well represented. The spatial distribution of dates does reveal a collection bias, with the preponderance of CRM archaeology that generated most of these dates occurring in leeward areas. 5. Discussion and implications for the Hawaiian archipelago and East Polynesia By assessing the reliability of radiocarbon dates based on the sample material and standard error, Wilmshurst et al. (2011a) have documented a rapid colonization phase for the distant archipelagos beyond the Society Islands in East Polynesia during the 12th to 13th centuries. Their results indicate settlement of this region several centuries later than longer chronologies had once proposed (e.g., Kirch, 1986). At the northern-most point of Polynesia, the settlement of Hawai’i represents one of the geographical extremes of Oceanic expansion. Our results for the time of initial colonization of Hawai’i Island between AD 1220 and AD 1261 corroborate the chronological pattern documented by Wilmshurst et al. (2011a) and are strikingly consistent with the well established 13th century settlement for New Zealand (Anderson, 1991; Higham et al., 1999; Wilmshurst et al., 2008) and Rapa Nui (Hunt and Lipo, 2006), and several other archipelagos of eastern Polynesia (Wilmshurst et al., 2011a) indicating an explosive period of settlement from central eastern Polynesia to the northern, southwestern, and southeastern limits of Polynesia. Our dataset of 926 radiocarbon dates from Hawai’i Island is the largest assemblage analyzed for the Hawaiian archipelago, and for a single island within Polynesia. The relative proximity of the other Hawaiian Islands presupposes that settlement of each of the major islands occurred within a relatively short time following initial colonization; a suggestion supported by the analysis of Class 1 dates from all islands in the Hawaiian archipelago (Wilmshurst et al., 2011a). Fig. 3. Estimate for the timing of colonization of Hawai’i Island based on the 16 Class 1 dates. The one-sigma ranges for the Class 1 calibrated radiocarbon dates are shown as black horizontal lines; circles represent median. The red dashed line represents the sum of the probability distribution for each calibrated Class 1 date. The solid blue line shows the cumulative probability relative to AD 1300 as measured on the right axis. The thin, dashed light blue line indicates the point on these curves where the area under the aggregate probability curve is 50%. This point represents the date at which it is more likely that the colonization event occurred before this point in time, a value based on the Late Age Estimation Model (LAEM). The range between the values for each model is shown by the thickness of the vertical yellow bar. This range represents the period of uncertainty where it is not possible to further specify a date of the colonization event. 2746 T.M. Rieth et al. / Journal of Archaeological Science 38 (2011) 2740e2749 Fig. 4. Minimum and maximum ages for each sample category. Based on the revised regional chronology, exploration, discovery, and colonization of Hawai’i appear to have occurred within decades at most. There is no evidence for a significantly earlier period of exploration followed by colonization (Graves and Addison, 1995). From a behavioral point of view, colonization might be taken to mean the landing of a canoe and setting foot on shore. This is problematic for several reasons including the elevation of a single event over a process, but practically because the current precision of radiocarbon dating does not allow us to finely differentiate the dating of events that occurred within less than about 100 years of each other. Thus arguments that our results document “settlement” (established communities with archaeological visibility) rather than “colonization” would be based on assumptions about the archaeological record that are not falsifiable. Our estimate for an early 13th century colonization date and the recent additional dates from Bellows and Halawa raises the possibility that these deposits are indeed among some of the earliest for the archipelago. Models for rapid population growth on newly colonized islands (e.g., Birdsell, 1957), suggests that expansion of settlements across Hawai’i Island, and likely the archipelago, occurred in a matter of a century or so. Directly related to such rapid population growth and settlement expansion is human-induced vegetation change and extinctions, including the impacts of invasive species (e.g., Athens et al., 2002; Athens, 2009; Hunt, 2007). Large-scale deforestation and extinctions may have occurred over decades, rather than centuries, and warrants further investigation. Taxonomic identification of avifaunal remains and direct dating of a large number of bones from introduced commensals such as the Polynesian rat (Rattus exulans) using ultrafiltration or XAD-2 resin extraction may help to refine this issue. Our assessment of the radiocarbon dataset from Hawai’i Island also reveals that much of the commonly accepted synthesis of Hawai’i’s post-colonization chronology now seems problematic. Along with the earliest Class 1 dates, only 97 younger dates (CRA 399 BP) on identified short-lived specimens can be used to accurately develop a chronology of events following initial colonization. This pool is decreased when any samples obtained from problematic contexts (e.g., ‘auwai’) or dates with large standard errors are removed. There are exceptions, such as the investigations of the Kohala Field System by Ladefoged and colleagues (e.g., Ladefoged and Graves (2007, 2008), Ladefoged et al. (2003, 2008) but questions about the development of social complexity, agricultural production strategies, resource use, warfare (defensive occupations), territoriality, and community interaction, among other inferences, remain poorly anchored in chronology. We have “end” points documented at the time of European contact in the late 18th century, but it is clear from our analysis that the prehistoric period still lacks significant chronological control. Resolving these broader research questions highlights methodological concerns. Unidentified wood charcoal remains the most common material submitted for radiocarbon dating by archaeologists working in Hawai’i. These dates are not only unreliable, they represent literally hundreds of thousands of dollars poorly spent. Care in the selection of sample provenience and sample material (Dye, 2000) must be a basic protocol if a reliable chronology for Hawai’i is to be established. An important issue that requires additional research is the discordance between the archaeological evidence and paleoenvironmental data relating to human colonization. Radiocarbon dates from sediment cores, and from bones of the introduced Polynesian rat from O’ahu and Kaua’i have been argued to suggest colonization between AD 800 and 1000, approximately 200e400 years earlier than our estimate (Athens et al., 2002; Burney, 2002; Burney and Burney, 2003). It is possible that these geologically older islands could have been settled earlier than Hawai’i Island. These islands have large valleys, permanent streams, and deep soils suitable for irrigated cultivation (Kirch, 2007; Ladefoged et al., 2009) along with well-developed reefs and near-shore marine ecosystems, offering a more abundant and reliable resource base than the younger islands. However, settlement of the older islands up to 400 years prior to archaeological evidence for Polynesians on Hawai’i Island or elsewhere in remote eastern Polynesia begs an extremely slow population growth rate and limited inter-island exploration. Conversely, it would demand a “cryptic” colonization (Anderson, 1995) of Hawai’i Island measured in centuries. T.M. Rieth et al. / Journal of Archaeological Science 38 (2011) 2740e2749 2747 Fig. 5. Geographical distribution of Class 1 dates; fine lines are ahupua’a boundaries, thick lines are district boundaries. In the sediment cores, the interpolation of ages based on a limited number of radiocarbon dates and rate of deposition models may be producing some of the inconsistencies (Athens et al., 2002: p. 61). Indeed, the initial occurrence of charcoal in the Ordy Pond core from ‘Ewa, O‘ahu, bracketed by two radiocarbon dates, could be interpreted as dating between approximately AD 1000 and 1100 (Athens et al., 1999: Table 8). Likewise, the potential for Polynesian rats up-taking old carbon from the limestone ‘Ewa Plain ecosystem or from a partial marine diet (see Caut et al., 2008a,b, 2009) may explain these earlier dates. As there are no isotope values reported for the rat bone samples, it remains unknown what influence a marine diet might have on these radiocarbon results. To further compound the issue, even if a marine influence is identified, the variability of the delta r estimates for O’ahu (ranging from À479 Æ 120 to 3842 Æ 100 for the island as a whole; 532 Æ 80 to 822 Æ 80 for Barbers Point, ‘Ewa; Petchey, 2009) still confounds a precise calibration. Clearly, additional investigations are needed. 6. Conclusions Our assessment of the reliability of 926 radiocarbon dates from Hawai’i Island provides the current best estimate for the Polynesian colonization of this island between AD 1220 and 1261. These results are approximately 250e450 years later than the current consensus, yet are convergent with Wilmshurst et al.’s (2011a) recent estimates for the colonization of remote East Polynesia beyond the Society Islands. Despite the enormous number of additional radiocarbon dates sourced for this study, it is discouraging that only about 5% of the additional dates !400 BP are Class 1, that is on sample materials capable of providing a reliable and precise dates. Using minimal 2748 T.M. Rieth et al. / Journal of Archaeological Science 38 (2011) 2740e2749 Dye, T., 2000. Effects of 14C sample selection in archaeology: an example from Hawai’i. Radiocarbon 42, 203e217. Dye, T.S., Komori, E., 1992. A pre-censal population history of Hawai’i. New Zealand Journal of Archaeology 14, 113e128. Dye, T.S., Pantaleo, J., 2010. Age of the O18 site, Hawai’i. Archaeology in Oceania 45, 113e119. Emory, K.P., Sinoto, Y.H., 1961. Hawaiian archaeology: Oahu excavations. In: Bishop Museum Special Publication, vol. 49. Bernice P. Emory, K.P., Sinoto, Y.H., 1969. Age of sites in the South Point area, Ka’u, Hawaii. In: Pacific Anthropological Records 8. Department of Anthropology, Bernice P. Bishop Museum, Honolulu. Field, J.S., Graves, M.W., 2008. A new chronology for Polulu valley, Hawai’i Island: occupational history and agricultural development. Radiocarbon 50, 205e222. Field, J.S., Kirch, P.V., Kawelu, K., Ladefoged, T.N., 2010. Households and hierarchy: domestic modes of production in leeward Kohala, Hawai’i Island. Journal of Island Coastal Archaeology 5, 52e85. Fitzpatrick, S., 2006. A critical approach to 14C dating in the Caribbean: using chronometric hygiene to evaluate chronological control and prehistoric settlement. Latin American Antiquity 17, 389e418. Graves, M.W., Addison, D.J., 1995. The Polynesian settlement of the Hawaiian Archipelago: integrating models and methods in archaeological interpretation. World Archaeology 26, 380e399. Hommon, R.J., 1986. Social evolution in ancient Hawai’i. In: Kirch, P.V. (Ed.), Island Societies: Archaeological Approaches to Evolution and Transformation. Cambridge University Press, Cambridge, pp. 55e68. Hunt, T.L., 2007. Rethinking Easter Islands’ ecological catastrophe. Journal of Archaeological Science 34, 485e502. Hunt, T.L., Holsen, R.M., 1991. An early radiocarbon chronology for the Hawaiian Islands: a preliminary analysis. Asian Perspectives 30, 147e161. Hunt, T.L., Lipo, C.P., 2006. Late colonization of Easter Island. Science 311, 1603e1606. Kirch, P.V., 1985. Feathered Gods and Fishhooks: an Introduction to Hawaiian Archaeology and Prehistory. University of Hawaii Press, Honolulu. Kirch, P.V., 1986. Rethinking East Polynesian prehistory. Journal of the Polynesian Society 95, 9e40. Kirch, P.V., 2007. Hawaii as a model system for human ecodynamics. American Anthropologist 109, 8e26. Kirch, P.V., Kelly, M. (Eds.), 1975. Prehistory and Ecology in a Windward Hawaiian Valley: Halawa Valley, Molokai. Pacific Anthropological Records 24. Department of Anthropology, Bernice P. Bishop Museum, Honolulu. Kirch, P.V., McCoy, M.D., 2007. Reconfiguring the Hawaiian cultural sequence:  results of re-dating the Halawa dune site (MO-A1-3), Moloka‘i Island. Journal of the Polynesian Society 116, 385e406. Ladefoged, T.N., Graves, M.W., 2007. Modeling agricultural development and demography in Kohala, Hawai’i. In: Kirch, P.V., Rallu, J. (Eds.), The Growth, Regulation, and Collapse of Island Societies. University of Hawai’i Press, Honolulu, pp. 70e89. Ladefoged, T.N., Graves, M.W., 2008. Variable development of dryland agriculture in Hawai’i: A fine-grained chronology from the Kohala Field System, Hawai’i Island. Current Anthropology 49, 771e802. Ladefoged, T.N., Graves, M.W., McCoy, M., 2003. Archaeological evidence for agricultural development in Kohala, Island of Hawai’i. Journal of Archaeological Science 30, 923e940. Ladefoged, T.N., Kirch, P.V., Gon III, S.M., Chadwick, O.A., Hartshorn, A.S., Vitousek, P.M., 2009. Opportunities and constraints for intensive agriculture in the Hawaiian archipelago prior to European contact. Journal of Archaeological Science 36, 2374e2383. Ladefoged, T.N., Lee, C., Graves, M.W., 2008. Modeling life expectancy and surplus production of dynamic pre-contact territories in leeward Kohala, Hawai’i. Journal of Anthropological Archaeology 27, 93e110. Liston, J., 2005. An assessment of radiocarbon dates from Palau, Western Micronesia. Radiocarbon 47, 295e334. Masse, W.B., Tuggle, H.D., 1998. The date of Hawaiian colonization. In: Stevenson, C.M., Lee, G., Morin, F.J. (Eds.), Easter Island in Pacific Context: Sout Seas Symposium: Proceedings of the Fourth International Conference on Easter Island and East Polynesia. Easter Island Foundation Occassional Paper 4. Bearsville and Cloud Mountain Presses, Los Osos, California, pp. 229e235. McCoy, M.D., 2007. A revised late Holocene culture history for Moloka’i Island, Hawai’i. Radiocarbon 49, 1273e1322. McCoy, M.D., Graves, M.W., 2010. The role of agricultural innovation on Pacific islands: a case study from Hawai’i Island. World Archaeology 42, 90e107. McFadgen, B.G., Knox, F.B., Cole, T.R.L., 1994. Radiocarbon calibration curve variations and their implications for the interpretation of New Zealand prehistory. Radiocarbon 36, 221e236. National Research Council, 1921. Recommendations for anthropological research in Polynesia. In: Committee on Publication (Ed.), Proceedings of the First PanPacific Scientific Conference Held Under the Auspices of the Pan-Pacific Union, Part 1. Bernice P. Bishop Museum Special Publication, vol. 7, pp. 103e123. Honolulu. O’Day, P.M., Rieth, T.M., 2007. Archaeological Monitoring, Phases II and III, Luala‘i Subdivision, Kamuela-Waimea, South Kohala, Hawai’i Island; TMK (3) 6e7-02. Prepared for D.R. Horton-Schuler Division. International Archaeological Research Institute, Inc. Pearson, R.J., Kirch, P.V., Pietrusewsky, M., 1971. An early prehistoric site at Bellows Beach, Waimanalo, Oahu, Hawaiian Islands. Archaeology and Physical Anthropology in Oceania 6, 204e234. criteria to evaluate the reliability of radiocarbon dates reveals that the majority of dates may be subject to a degree of uncertainty based on in-built age. The lack of wood charcoal identification for the selection of short-lived plant specimens for dating has resulted in a pool of dates that cannot reliably inform upon substantive questions in Hawaiian prehistory, and reveals that much of the accepted timing of chronological events remains speculative. Wood charcoal identification and selection of short-lived plants/plant parts for radiocarbon dating would greatly enhance our understanding and refine the settlement chronology of Hawai’i Island, and of other islands lacking Class 1 dates for colonization. Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version at doi:10.1016/j.jas.2011.06.017. References Anderson, A., 1991. The chronology of colonization in New Zealand. Antiquity 65, 767e795. Anderson, A., 1995. Current approaches in East Polynesian colonisation research. Journal of the Polynesian Society 104, 110e132. Athens, J.S., 1997. Hawaiian native lowland vegetation in prehistory. In: Kirch, P.V., Hunt, T.L. (Eds.), Historical Ecology in the Pacific Islands. Yale University Press, New Haven, pp. 248e270. Athens, J.S., 2009. Rattus exulans and the catastrophic disappearance of Hawai‘i’s native lowland forest. Biological Invasions 11, 1489e1501. Athens, J.S., Tuggle, H.D., Ward, J.V., Welch, D.J., 2002. Avifaunal extinctions, vegetation change, and Polynesian impacts in prehistoric Hawai‘i. Archaeology in Oceania 37, 57e78. Athens, J.S., Ward, J.V., Blinn, D.W., 1995. Paleoenvironmental Investigations at ‘Uko‘a Pond, Kawailoa ahupua’a, O’ahu, Hawaii. Prepared for Engineers Surveyors Hawaii, Inc. International Archaeological Research Institute, Inc. Athens, J.S., Ward, J.V., Tuggle, H.D., Welch, D.J., 1999. Environment, Vegetation Change, and Early Human Settlement on the ‘Ewa Plain: a Cultural Resource Inventory of Naval Air Station, Barbers Point, O’ahu, Hawai’i; Part III: Paleoenvironmental Investigations. Prepared for Department of the Navy, Pacific Division, Naval Facilities Engineering Command, Pearl Harbor, Under Contract with Belt Collins Hawaii. International Archaeological Research Institute, Inc. Barry, B.A., 1978. Errors in Practical Measurement in Science, Engineering and Technology. John Wiley & Sons, New York. Birdsell, J., 1957. Some population problems involving Pleistocene man. Cold Spring Harbor Symposia on Quantitative Biology 22, 47e69. Burney, D.A., 2002. Late Quaternary chronology and stratigraphy of twelve sites on Kaua’i. Radiocarbon 44, 13e44. Burney, D.A., James, H.F., Burney, L.P., Olson, S.L., Kikuchi, W., Wagner, W.L., Burney, M., McCloskey, D., Kikuchi, D., Grady, F.V., Gage, R., Nishek, R., 2001. Fossil evidence for a diverse biota from Kaua‘i and its transformation since human arrival. Ecological Monographs 71, 615e641. Burney, L.P., Burney, D.A., 2003. Charcoal stratigraphies for Kaua’i and the timing of human arrival. Pacific Science 57, 211e226. Carson, M.T., 2005. A radiocarbon dating synthesis for Kaua’i. In: Carson, M.T., Graves, M.W. (Eds.), Na Mea Kahiko o Kau’i: Archaeological Studies in Kaua’i. Society for Hawaiian Archaeology, Special Publication, No. 2, pp. 11e32. Honolulu. Caut, S., Angulo, E., Courchamp, F., 2008a. Dietary shift of an invasive predator: rats, seabirds, and sea turtles. Journal of Applied Ecology 45, 425e534. Caut, S., Angulo, E., Courchamp, F., 2008b. Discrimination factors (D15N and D13C) in an omnivorous consumer: effect of diet isotopic ration. Functional Ecology 22, 255e263. Caut, S., Angulo, E., Courchamp, F., 2009. Avoiding surprise effects on Surprise Island: alien species control in a multitrophic level perspective. Biological Invasions 11, 1689e1703. Cordy, R., 2000. Exalted Sits the Chief: The Ancient History of Hawai’i Island. Mutual Publishing, Honolulu. Dean, J.S., 1978. Independent dating in archaeological analysis. In: Schiffer, M.B. (Ed.), Advances in Archaeological Method and Theory, vol. 1. Academic Press, New York, pp. 223e258. Dunnell, R.C., 1978. Style and function: a fundamental dichotomy. American Antiquity 43, 192e202. Dye, T., 1989. Tales of two cultures: traditional historical and archaeological interpretations of Hawaiian prehistory. Bishop Museum Occasional Papers 29, 3e22. Dye, T., 1992. The South Point radiocarbon dates thirty years later. New Zealand Journal of Archaeology 14, 89e97. Dye, T., 1994a. Population trends in Hawai’i before 1778. The Hawaiian Journal of History 28, 1e20. Dye, T., 1994b. Apparent ages of marine shells: implications for archaeological dating in Hawai‘i. Radiocarbon 36, 51e57. T.M. Rieth et al. / Journal of Archaeological Science 38 (2011) 2740e2749 Petchey, F., 2009. Dating marine shell in Oceania: issues and prospects. In: O’Connor, S., Fairbairn, A. (Eds.), Proceedings of the 2005 Australasian Archaeometry Conference. Terra Australis. ANU E Press, Canberra, pp. 157e172. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey, C., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., Weyhenmeyer, C.E., 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0e50,000 years cal BP. Radiocarbon 51, 1111e1150.  Rieth, T.M., Hunt, T.L., 2008. A radiocarbon chronology for Samoan prehistory. Journal of Archaeological Science 35, 1901e1927. Spriggs, M., Anderson, A., 1993. Late colonization of East Polynesia. Antiquity 67, 200e217. Robins, J., Eble, F., Anderson, L., 2000. Data Recovery of Archaeological Sites in the Land of Ouli (TMK:3-6-01:18, 74 & 75), District of South Kohala, Hawai’i Island, Hawaii. Prepared for Hale Wailani, L.P. Ogden Environmental and Energy Services Co., Inc. Tuggle, H.D., 1997. Archaeological Research of Areas Proposed for Development of Military Family Housing and Expansion of Military Training at Bellows Air Force Station, O’ahu. Prepared for Pacific Division, Naval Facilities Engineering Command, Pearl Harbor, Hawai’i. International Archaeological Research Institute, Inc. 2749 Tuggle, H.D., Spriggs, M., 2001. Age of the Bellows Dune Site O18, O‘ahu, Hawai’i, and the antiquity of Hawaiian colonization. Asian Perspectives 39, 165e188. Walter, R.K., 1996. What is the East Polynesian ‘archaic’?: A view from the Cook Islands. In: Davidson, J.M., Irwin, G., Leach, F., Pawley, A., Brown, D. (Eds.), Oceanic Culture History: Essays in Honour of Roger Green. New Zealand Archaeological Association Special Publication, pp. 513e530. Wellington. Weisler, M., 1989. Chronometric dating and late Holocene prehistory in the Hawaiian Islands: a critical review of radiocarbon dates from Moloka‘i Island. Radiocarbon 31, 121e145. Williams, S.S., 2002. Ecosystem Management Program, Cultural Resources Inventory Survey of Previously Unsurveyd Areas, Redleg Trail Vicinity, U.S. Army Pohakuloa Training Area, Island of Hawai’i, Hawai’i. Prepared for U.S. Army Engineer District, Honolulu. Ogden Environmental and Energy Services Co., Inc. Wilmshurst, J.M., Anderson, A.J., Higham, T.F.G., Worthy, T.H., 2008. Dating the late prehistoric dispersal of Polynesians to New Zealand using the commensal Pacific rat. Proceedings of the National Academy of Sciences 105, 7676e7680. Wilmshurst, J.M., Hunt, T.L., Lipo, C.P., Anderson, A.J., 2011a. High-precision radiocarbon dating shows recent and rapid colonization of East Polynesia. Proceedings of the National Academy of Sciences 108, 1815e1820. Wilmshurst, J.M., Hunt, T.L., Lipo, C.P., Anderson, A.J., 2011b. Reply to Mulrooney et al: accepting lower precision radiocarbon dates results in longer colonization chronologies. Proceedings of the National Academy of Sciences 108, E195.