AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 121:319 –331 (2003) Cranial Variation in the Marquesas Islands Vincent H. Stefan1* and Patrick M. Chapman2 1 2 Department of Anthropology, Lehman College, CUNY, Bronx, New York 10468 Department of Anthropology, South Puget Sound Community College, Olympia, Washington 98512 KEY WORDS Marquesan; Marquesas Islands; craniometrics; cranial discrete traits ABSTRACT The Marquesas Islands have traditionally been divided into a northwestern and a southeastern group, a division which reﬂects language dialect differences. Additionally, archaeological studies have also suggested that differences in material culture existed between the northwestern and southeastern islands. This study examines Marquesan cranial discrete and metric traits to evaluate the level of intra-archipelago heterogeneity, and to determine if a northwest/southeast division is evident cranially. The data consist of 28 cranial discrete traits and 49 craniofacial measurements of prehistoric Marquesans. Male and female data are pooled for discrete trait and metric data, following a Z-score standardization technique adjustment. The data represent three island samples: Nuku Hiva (northwest), Fatuiva (southeast), and a combined Tahuata/Hiva ‘Oa (southeast). Of the 28 discrete traits, 16 are utilized in a mean measure of divergence analysis that provides scores of 0.259 for FatuivaTahuata/Hiva ‘Oa, 1.850 for Nuku Hiva-Fatuiva, and 1.491 for Nuku Hiva-Tahuata/Hiva ‘Oa. Of the 49 craniofacial measurements, 46 are utilized in RMET/NORM analyses, providing unbiased D2 values of 0.0433 for Fa- tuiva-Tahuata/Hiva ‘Oa, 0.1328 for Nuku Hiva-Fatuiva, and 0.0813 for Nuku Hiva-Tahuata/Hiva ‘Oa. The islands of the southeastern group are closer to each other than either was to the island of the northwestern group. When a sample from ‘Ua Huka is included in the craniometric analysis, the unbiased D2 values of 0.0829, 0.1175, and 0.0431 are calculated for ‘Ua Huka and Nuku Hiva, and Fatuiva and Tahuata/Hiva ‘Oa pairings, respectively, indicating a close similarity of ‘Ua Huka to the southeastern islands. Mean measure of divergence analysis of cranial discrete traits as well as RMET/NORM analyses of craniometric variables reveal that differences exist between the islands of the northwestern and southeastern Marquesas Islands. These results support previous research that documented linguistic and cultural differences between these regions of the archipelago. However, the results indicate that ‘Ua Huka, an island traditionally included in the northwestern Marquesas Islands, has an afﬁnity to the southeastern Marquesas Islands, possibly due to its pivotal position as a waypoint in the Marquesas Island interaction sphere. Am J Phys Anthropol 121:319 –331, 2003. Osteological studies of the prehistoric Polynesians have focused primarily on the location of their ancestral home and relationship to other Oceanic populations (e.g., Houghton, 1989; Howells, 1970; Pietrusewsky, 1976, 1977, 1983, 1984, 1990b,c, 1994). Although large collections of Paciﬁc osteological material are available in museums throughout the world, questions concerning the variation of local populations have largely been ignored, with a few notable exceptions (Easter Island: Chapman, 1993; Chapman and Gill, 1997; Gill and Owsley, 1993; Stefan, 1999; Tigner and Gill, 1986; Zimple and Gill, 1986; Hawai‘i: Pietrusewsky, n.d.; New Zealand and the Chatham Islands: Buranarugsa and Leach, 1993; Harlow, 1979, 1994; Houghton, 1980; Scott, 1893; Shima and Suzuki, 1967). Giles (1973, p. 400) raised this point over two decades ago, suggesting: “Osteological studies all seem to look through the wrong end of the . . . telescope: none has been concerned with what one might call local populations.” However, his comment has not yielded the desired effect, and the search for the ancestral homeland of the Polynesians continues in lieu of more detailed analyses of local populations. This is surprising, be- cause one of the observed trends in biological anthropology is the analysis of populations at the local level (Buikstra et al., 1990). Eastern Polynesia is an ideal region for the study of population history because of the relative ease in deﬁning speciﬁc populations and their general isolation from non-Polynesian populations. Indeed, Houghton (1980, 1989, 1990, 1991a, b) and Howells (1979) described the region as biologically homogeneous, with minimal external gene ﬂow limited primarily to Western Polynesia. In addition, the populations of the eastern Paciﬁc share a common and © 2003 WILEY-LISS, INC. © 2003 Wiley-Liss, Inc. Grant sponsor: Kon-Tiki Museum, Oslo. *Correspondence to: Vincent H. Stefan, Department of Anthropology, Lehman College, CUNY, 250 Bedford Park Blvd. West, Bronx, NY 10468. E-mail: firstname.lastname@example.org Received 30 April 2001; accepted 29 January 2003. DOI 10.1002/ajpa.10287 320 V.H. STEFAN AND P.M. CHAPMAN recent ancestry (Bellwood, 1989; Green, 1993; Hagelberg and Clegg, 1993; Jennings, 1979; Marck, 1996; Rolett, 1993), providing another advantage for the study of population history within the region. There are also relatively well-deﬁned environmental differences. For example, the islands of the Tuamotu Archipelago are primarily atolls, while the Society Islands are high, volcanic islands, representing the extremes of a broad spectrum of temperatures and microclimates. Given that the Polynesian islands have differing climates and resources, it may be possible to discern speciﬁc environmental factors inﬂuencing human biology. The assessment of population history focuses on relative similarities and differences, either through the use of anthropological genetics which examines the factors affecting genetic similarities between populations, or the use of osteological studies examining the morphological similarities between populations. Because osteological morphology is caused or inﬂuenced by either genetic and/or environmental factors (Harpending and Jenkins, 1973; Relethford, 1996), it is an appropriate means for studying Paciﬁc populations and their history. This paper investigates the cranial morphology and population history of one Eastern Polynesian island group, the Marquesas Islands. The possible division of the Marquesas population into northwestern and southeastern subpopulations is examined from a biological perspective. This study will utilize both cranial discrete and metric data to evaluate the similarities and dissimilarities of prehistoric populations from various islands within the Marquesas Islands. Given the similar climates and environments of the islands within this group (Freeman, 1951), any differences observed between populations could be due to other evolutionary forces such as isolation, genetic drift, and/or nonrandom mating. The purpose of this study is to document any interisland/intraarchipelago cranial morphological variation, in order to correlate this variation with other aspects of Marquesas Islands culture and population history. BACKGROUND Marquesas islands Located approximately 1,400 km from Tahiti, the Marquesas Islands are comprised of two distinct groups: the northwest Marquesas and the southeast Marquesas. The archipelago’s 10 main islands are Ha Tutu, ‘Ei ‘A‘o, Nuku Hiva, ‘Ua Huka, and ‘Ua Pou (northwest Marquesas); Fatu Huku, Hiva ‘Oa, Tahuata, Moho Tani, and Fatuiva (southeast Marquesas) (Freeman, 1951; Hughes and Fischer, 1998) (Fig. 1). Linguistic studies (Green, 1966; Lavondès and Randall, 1978) suggest a dialect differentiation between the northwestern and southeastern islands of the Marquesan archipelago. Lavondès and Randall (1978) noted regional differences in the names of common ﬁsh utilized in these two regions. Fig. 1. Marquesas Islands. Dashed line demarcates traditional division of archipelago into northwestern and southeastern island groups. Lavondès and Randall (1978) examined the Marquesan names for 83 types of ﬁsh from Nuku Hiva, ‘Ua Pou , ‘Ua Huka, Hiva ‘Oa, and Fatuiva. Some general observations were made that were consistent with a previous identiﬁcation of two regional dialects (Biggs, 1971). Some observations included phoneme changes, where the northwestern Marquesan (MQN) dialect used an /h/ instead of the /f/ used in southeastern Marquesan (MQS), and the use of /k/ in MQN in place of /n/ in MQS. Their conclusions were that Nuku Hiva and ‘Ua Pou, in addition to ‘Ua Huka, are in the northwestern group, and Hiva ‘Oa (with Tahuata) and Fatuiva are in the southeastern group. However, ‘Ua Huka also demonstrates a degree of afﬁnity with the southeastern islands that the other northwestern islands do not have (Lavondès and Randall, 1978). Archaeological studies also suggest that some differences in material culture existed between the northwestern and southeastern islands (Handy, 1923; Linton, 1923, 1925; Sinoto, 1979), though as a whole the material culture was rather uniform. In an exhaustive study of the Marquesas material culture, Linton (1923, p. 445) concluded that “most of these differences seem to have consisted in a greater or less stressing of features common to the whole culture rather than in a clear-cut absence or presence of traits.” Sullivan (1923c) further believed that in view of the trade and intercourse that was known to exist in the Marquesas during prehistoric times, the fact that regional differences were still evident at the time of his survey indicated to him that in MARQUESAN CRANIAL VARIATION prehistoric times, the local distinctions would have been more pronounced. Regional differences were noted in the design and construction of houses, the style of dress and ornamentation, methods of stone construction, design of ceremonial structures and sculptures, sociopolitical structure, and in the handling and disposal of the dead (Linton, 1923). Handy (1923, p. 21) generalized the cultural differences between the northwestern and southeastern Marquesas Islands in the following major features: 1) the northwestern Marquesans built with large stones, and erected large house platforms, ceremonial structures, and dance areas; and 2) the southeastern Marquesans had a more highly developed skill in the carving and sculpturing of stone and the cutting of stone blocks, and the art of wood carving and tattooing was centered in the south, which then spread northward. Recent archaeological and cultural investigations of the Marquesas Islands indicate that many of these differences arose predominantly after the end of the Developmental Period (AD 1300) (Rolett, 1989). A more detailed summary of the local differences in the material cultures and practices of the Marquesas Islands can be found in Linton (1923, p. 445– 446) Anthropometry studies by early nineteenth century physical anthropologists also identiﬁed physical differences between Polynesian populations. These studies identiﬁed what they believed to be three racial elements: a dolichocephalic Negroid race, a dolichocephalic or mesocephalic race which showed Caucasic afﬁnities, and a brachycephalic race with Mongoloid afﬁnities which Sullivan (1923b) called Indonesian. With regards to northwestern and southeastern Marquesans, Sullivan (1923a) identiﬁed an “Indonesian physical type” in the northwest and a “Polynesian physical type” in the southeast. Handy (1923, p. 21) characterized the southeastern Marquesans as having “longer heads, curlier hair, shorter stature, and lighter skin.” Other than these early observations, there is an overall paucity of information on the interisland morphological and anthropometric variation of the Marquesas Islanders, due to a lack of research and investigation by modern researchers. As with the other islands of French Polynesia, very little osteological information exists for the Marquesas Islands. Pietrusewsky (1976) examined archaeologically excavated skeletons from the Hane Dune site on ‘Ua Huka. His report included cranial and postcranial measurements and observations, dental observations, and determination of paleopathologies. His analysis of 42 individuals yielded one of the best archaeologically provenienced collections of an East Polynesian skeletal series. Multivariate analysis demonstrated that the Hane sample associates loosely with other East Polynesian samples. Until this study, Pietrusewsky (1976) was the only study of a Polynesian population that focused solely on the Marquesas Islands and that eval- 321 uated the morphology of individuals inhabiting those islands. Use of cranial metric and discrete traits in population studies A major difﬁculty in the analysis of prehistoric populations or subpopulations is that they are often poorly deﬁned or represented in the archaeological record. This obstacle is problematic to confront and resolve. Ethnic groups, cemetery samples, and skeletal samples have often been, correctly or incorrectly, equated with biological populations (Cadien et al., 1974). Because it is assumed that the skeletal sample represents a sequence of related individuals, these researchers suggest that skeletal samples can be considered skeletal “lineages” reﬂecting the action of microevolutionary forces over time. However, the time depth of the study sample could determine the validity of this “sequence of related individuals” assumption. If the sample encompasses only several generations, the assumption may hold true, but if the sample encompasses dozens of generations, the existence of long-term gene ﬂow could have signiﬁcantly altered the pattern of relatedness (Konigsberg, 1990; Relethford, 1999). With the existence of long-term gene ﬂow, a population could display greater afﬁnity to the populations contributing genes than to earlier generations of the same population prior to the onset of gene ﬂow. When analyzing past population structure, the element of time within cemetery skeletal samples makes analyses more difﬁcult. Konigsberg (1987) demonstrated that the genetic variance (measured with Wright’s FST) estimated for several combined generations is equivalent to the average of the genetic variance of each discrete generation, and noted that even though a certain amount of phenotypic variation within a skeletal sample is due to noncontemporaneous individuals, between-site analyses should not be affected by within-site variation. Therefore, skeletal samples or lineages can be considered equivalent to a biological population statistically. Due to the fact that many studies investigating the population biology of past peoples utilize skeletal samples of those prehistoric populations, the genetic variance of those populations cannot be directly measured. It was previously discussed that phenotypic variation is a reﬂection of genotypic variation within groups, and this relationship is considered consistent among groups. Because genotypic distances between populations cannot be calculated due to the unavailability of their underlying genotypic variance structure, a viable solution is to use phenotypic distances, calculated from population phenotypic variances, as estimates of genetic relatedness among populations. Therefore, phenotypic distances can be interpreted within a population genetics framework because they are directly proportional to the actual genetic distances among populations in a consistent way, assuming that herita- 322 V.H. STEFAN AND P.M. CHAPMAN bility is constant across populations (WilliamsBlangero and Blangero, 1989). The literature is replete with the discussion of the utility of cranial phenotypic variation (metric and discrete) to represent genotypic variation (e.g., Berry and Berry, 1967; Berry, 1963, 1964; Cheverud, 1988; Grünenberg, 1963; Konigsberg and Ousley, 1995; Molto, 1983; Relethford and Blangero, 1990; Saunders, 1989; van Vark and Schaafsma, 1992; Williams-Blangero and Blangero, 1989), the heritabilities of various traits (e.g., Bocquet-Appel, 1984; Byard et al., 1984; Cheverud, 1988; Cheverud and Buikstra, 1982; Cheverud et al., 1979; Corruccini, 1974, 1976; Devor et al., 1986a,b; Donnelly et al., 1998; Droessler, 1981; Howells, 1953; Kohn, 1991; Konigsberg and Ousley, 1995; Molto, 1983; Najem, 1997; Ossenberg, 1977; Relethford, 1994; Relethford and Harpending, 1994; Self and Leamy, 1978; Sjøvold, 1984,1986; Susanne, 1977; Susanne et al., 1983; Vandenberg, 1962), and the population genetics methodology developed to evaluate relatedness among populations. Therefore, a detailed discussion will not be provided here. The effective use of both craniometric and discrete traits in the investigation of prehistoric Polynesian relationships, both intraisland and interisland, has been repeatedly demonstrated. During the last 30 years, several researchers have been at the forefront of Paciﬁc bioanthropological research, including Brace and Hunt (1990), Brace et al. (1989, 1990, 1991), Hanihara (1992, 1996, 1997), Chapman (1993, 1998), Chapman and Gill (1997, 1998), Howells (1970, 1973, 1979, 1989, 1990, 1995), Katayama (1987, 1990), Tagaya and Katayama (1988), Pietrusewsky (1971, 1976, 1977, 1983, 1984, 1988, 1990a,b, 1992, 1994, 1996, 1997), Pietrusewsky and Ikehara-Quebral (2000), Pietrusewsky et al. (1992), Stefan (1999, 2000), and Stefan et al. (1998). Each of these researchers employed statistical techniques designed to identify phylogenetic relationships, with several having designed their analyses within a population genetics framework. This study further demonstrates the utility of cranial metric and discrete data in the assessment of intraisland relationships. MATERIALS AND METHODS Three island samples were utilized in this study to represent the Marquesas Islands regions: Fatuiva, a combined Tahuata/Hiva ‘Oa representing the southeastern region, and Nuku Hiva representing the northwestern region. Cranial nonmetric and metric data were collected only from adult individuals, as evidenced by a fused sphenooccipital sychondrosis (Krogman and Iscan, 1986), and only the most complete crania were selected for data collection. Age determination beyond that of “adult” was not deemed necessary for the analyses to be conducted. The sex of each cranium was determined utilizing standard anthropological techniques (Bass, 1995; Buikstra and Ubelaker, 1994). The Marquesas Islands samples utilized in this study are curated at the Bernice P. Bishop Museum (BPBM, State Museum of Natural and Cultural History, Honolulu, HI); the American Museum of Natural History (AMNH, New York, NY); the Natural History Museum (NHM, London, UK); and the Laboratoire d’Anthropologie Biologique, Musée de l’Homme (MH, Paris, France). The crania curated at the BPBM were collected by the Bayard Dominick Expedition in 1920 –1921, and by Y.H. Sinoto in 1964 –1965 from the Hane dune Site on ‘Ua Huka. The level of the Hane Dune site from which human remains were recovered, Level IV, has been dated to AD 1110 ⫾ 110 –1635 ⫾ 90 (Sinoto, 1970). The crania at AMNM were collected by H.L. Shapiro during the Templeton Crocker Paciﬁc Expedition in 1934 and possibly during his participation in the B.P. Bishop Museum Tuamotu Expedition in 1929. The crania at the MH were collected by C.L. Clavel while serving as medical ofﬁcer on the French sloop Hugon in 1881–1882. Exact provenience and dates for these specimens are not available. However, it is believed that these crania represent pre-European contact individuals. It should be noted that the authors collected their data independent of one another and at different times. Additionally, each author examined the same crania curated at NHM and MH, but only one of us (V.H.S.) examined the crania curated at BPBM and AMNH. This explains the discrepancies that will be observed in sample sizes from the various Marquesas Islands. Cranial discrete traits Cranial discrete trait data were collected for 28 traits, from a total of 101 adult crania, obtained from various islands within the Marquesas Islands. This study incorporated discrete traits of the cranial vault and face as deﬁned by Berry and Berry (1967), Olivier (1969), Ossenberg (1969, 1970), and Pardoe (1984). Detailed assessment and scoring methodologies for these traits are discussed in these references as well. To avoid or minimize interobserver error, all the cranial discrete data used in this study were collected by one of us (P.M.C.). Sixteen of 28 traits examined were analyzed using the mean measure of divergence (MMD). The traits used are listed in Table 1 (for 10 of the 16 traits, data were collected for both the right and left sides, which accounts for the 26 traits listed in Table 1). The remaining 12 traits were eliminated due to phenotypic homogeneity (absolutely no variation), lack of replicability, or the possibility of signiﬁcant environmental inﬂuence upon the attribute (for more information, see Chapman, 1998). The traits analyzed are similar to those in other studies of a similar nature (Berry, 1974, 1975; Konigsberg et al., 1993; Molto, 1983; Ossenberg, 1969, 1970, 1976; Pardoe, 1984; Pietrusewsky, 1977, 1983, 1984; Prowse and Lovell, 1996; Sjøvold, 1984,1986). 323 MARQUESAN CRANIAL VARIATION TABLE 1. Discrete trait frequencies for Marquesas Islands samples Fatuiva Nuku Hiva Tahuata/Hiva ‘Oa Trait F M Total F M Total F M Total Lambdoidal wormians (right) Lambdoidal wormians (left) Epactal ossicle Parietal notch bone (right) Parietal notch bone (left) Asterion ossicle (right) Asterion ossicle (left) Mastoid ossicle (right) Mastoid ossicle (left) Epipteric ossicle (right) Epipteric ossicle (left) Sagittal ossicle Divided hypoglossal canal (right) Divided hypoglossal canal (left) Accessory lesser palatine foramina (right) Accessory lesser palatine foramina (left) Infraorbital suture (right) Infraorbital suture (left) Accessory malar foramina (right) Accessory malar foramina (left) Palatine torus Sagittal sulcus (right-branching) Parietal foramen (right) Parietal foramen (left) Elliptic palate Pharyngeal fossa 4/6 4/6 2/6 0/7 3/7 0/7 1/7 1/7 1/6 0/7 2/7 3/6 1/6 2/6 6/7 5/7 3/7 2/7 4/7 5/7 1/7 6/7 6/7 4/7 2/7 0/6 12/19 12/19 0/19 1/19 5/18 5/18 4/19 3/18 3/18 2/19 2/19 7/15 1/19 1/19 13/18 12/18 7/19 3/19 17/18 15/18 1/19 12/19 13/19 8/19 0/18 2/19 16/25 16/25 2/25 1/26 8/25 5/25 5/26 4/25 4/24 2/26 4/26 10/21 2/25 3/25 19/25 17/25 10/26 5/26 21/25 20/25 2/26 18/26 19/26 12/26 2/25 2/25 9/10 7/10 0/10 2/10 1/10 0/10 1/10 0/10 0/10 1/10 0/10 2/9 1/10 1/10 6/9 4/9 4/10 1/10 7/9 7/10 0/10 8/10 9/10 7/10 1/10 2/10 13/16 14/17 0/18 2/20 1/20 2/20 1/20 1/20 2/20 1/20 1/20 7/14 1/19 2/19 14/19 14/20 3/20 2/19 14/20 17/20 4/19 17/20 15/19 14/19 1/20 6/20 22/26 21/27 0/28 4/30 2/30 2/30 2/30 1/30 2/30 2/30 1/30 9/23 2/29 3/29 20/28 18/29 7/30 3/29 21/29 24/30 4/29 25/30 24/29 21/29 2/30 8/30 9/10 9/10 0/10 4/10 3/10 4/10 2/10 2/10 2/10 2/10 1/10 3/7 0/10 0/10 4/10 5/8 6/10 6/10 8/9 6/10 2/9 9/10 8/10 6/10 0/9 3/10 7/9 6/9 1/9 2/10 1/10 3/9 1/10 3/9 3/10 0/10 1/10 4/9 0/9 0/9 7/9 6/10 4/10 3/10 7/8 9/10 0/10 7/10 7/10 8/10 1/10 3/9 16/19 15/19 1/19 6/20 4/20 7/19 3/20 5/19 5/20 2/20 2/20 7/16 0/19 0/19 11/19 11/18 10/20 9/20 15/17 15/20 2/19 16/20 15/20 14/20 1/19 6/19 Most of the traits are bilateral, i.e., they can be present on either or both the left and right sides. There are a number of ways of dealing with bilateral variables (Green et al., 1979; Korey, 1980; Saunders, 1989), each of which has its advantages. The three most common ways of treating these data are: 1) recording the trait as present if it is present on at least one side of the individual (sampling by individual); 2) treating each side as a separate variable; and 3) randomizing the side used for each individual. The ﬁrst and third methods give equal weight to unilateral and bilateral variables, whereas the second gives double weight to the bilateral variables, with each side treated separately. Saunders (1989) indicateD that there is a strong, but not perfect, positive correlation between side interdependence, suggesting that the second method artiﬁcially biases the results. However, sampling by individual (method 1) increases the attribute frequency within each population. This can create problems with statistical analyses using angular transformations of the frequencies if the frequencies are close to 95% present or absent. Therefore, this study uses method 3, randomizing the side used for each individual. This method ensures that bilateral attributes are given a weight equal to that of unilateral attributes without inﬂating frequencies. For fragmentary remains when data from both sides are not available, the side for which information is available is used. The mean measure of divergence (MMD) examines levels of similarity and difference between different samples and produces a distance measure. The MMD statistic is a summed-difference-of-means test of variable frequencies between two samples. The frequencies undergo angular transformation to obtain independence of variance (Pardoe, 1991, p. 5). The resulting proportional difference is then divided by a standard error with correction for sampling variance, and then averaged for all variables analyzed. The MMD scores are then divided by the standard deviation to produce standardized MMD scores (de Souza and Houghton, 1977). In order to ensure consistency in sample size, random samples were selected from the locations with the largest number of crania (Nuku Hiva, Fatuiva, and Tahuata/Hiva ‘Oa), thereby limiting the actual number of crania included in this study (Tables 1 and 3). In order to create consistent sample sizes, ranging from 19.4 –20.2, we combined neighboring Tahuata and Hiva ‘Oa. Geographically these islands are extremely close (approximately 5 km apart), indicating no major geographical barrier to gene ﬂow. Due to the demonstrable lack of signiﬁcant differences between the sexes, male and female crania were combined into one sample. Additional information concerning sample acquisition and composition may be found in Chapman (1998). Chapman (1998) illustrated that the MMD results are signiﬁcantly affected by sample size. Previous nonmetric analyses of Polynesian populations (including Chapman and Gill, 1998; Katayama, 1987; Pietrusewsky, 1977) using MMD may have produced biased results, given that the comparative samples all had signiﬁcantly varying sample sizes, reﬂecting the inﬂuence of sample sizes in addition to actual biological relationships. Therefore, to ensure unbiased results, the sample sizes for each location included in this study are consistent with one another, with the crania included chosen at random. 324 V.H. STEFAN AND P.M. CHAPMAN TABLE 2. Cranial sample sizes for Marquesas Islands craniometric analyses Location Female Male Total Fatuiva Nuku Hiva Tahuata/Hiva ‘Oa ’Ua Huka 9 39 14 12 20 60 16 10 29 99 30 22 Craniometric traits Cranial metric data were collected from a total of 210 adult crania from various islands within the Marquesas Islands. This study incorporated measures of the cranial vault, face, and interorbital region, as deﬁned by Bass (1995), Gill et al. (1988), Howells (1973), and Martin and Saller (1957). The list of 49 standardized measurements collected (with their abbreviations) is provided in the Appendix. To avoid or minimize interobserver error, all craniometric data used in this study were collected by one of us (V.H.S.). Due to insufﬁcient provenience information, only 185 individuals were included in this initial data analysis. As a result of the state of preservation of specimens examined, the initial dataset possessed 10.8% missing data, with 8,086 of 9,065 possible data points present. To reduce the percentage of missing data, ﬁve specimens were removed from the dataset, and three variables (STB, MXB, and MXS) were removed from the dataset due to excessive missing data. This resulted in the reduction of missing data to 6.1%, with 7,774 of 8,280 possible data points present. The remaining 180 individuals (Table 2) and 46 variables were then used in further multivariate analyses. Multivariate analysis procedures require that there be no missing data. Every observation must have a value for each variable entered into the analysis, or else the observation is eliminated. Deleting observations results in large amounts of information being lost, and the remaining completely observed cases would be unrepresentative of the population that they were intended to reﬂect. The problem of missing data needs to be addressed in order to optimize the number of observations utilized from each sample. Considerable attention has been given to the problem of missing data estimation, along with the most appropriate procedures for estimating missing data (e.g., Droessler, 1981). Traditionally, three options have been available: substitution of group means, substitution of grand means, or prediction of missing measurements by means of multiple regression. These alternatives have their associated advantages and disadvantages (Droessler, 1981, p. 80 – 84; Schafer, 1997). This research utilizes the NORM 2.01 statistical program (Schafer, 1999) for multiple generation of incomplete multivariate datasets and techniques, as described by Schafer (1997). The resulting datasets are used to assess the validity in combining point estimates and covariance matrices to be utilized in the intragroup variability analyses discussed below. These methods will minimize the problems inherent in missing value estimations discussed above. Detailed information on the computational methodologies utilized by NORM and the underlying statistical/mathematical foundations can be found in Schafer (1997, 1999) and Rubin (1987). For this study, each individual variable was corrected for size and sex, using the Z-score standardization techniques discussed by Howells (1973, 1989), following the estimation of missing values. This method is desirable for removing the effects of sexual dimorphism while retaining intrapopulation variation. Samples could then be considered without reference to their sex, thereby allowing combined sex samples and effectively increasing comparative sample sizes. This and similar procedures for the removal of size and sex effects are now standard procedures in studies of human phenotypic variation (Key, 1983; Key and Jantz, 1990; Konigsberg and Blangero, 1993; Relethford and Harpending, 1994; Williams-Blangero and Blangero, 1989). A stepwise discriminant function analysis was conducted on the Z-score standardized data to determine which variables would provide the best discrimination between samples (SAS Institute, 1990). The analysis revealed seven variables that best discriminated the samples (WCB, ZOS, EKB, BBH, NLB, ALB, and ASB). The RMET analyses will be conducted using both the complete dataset of 46 variables and the seven variables identiﬁed in the stepwise discriminant function analysis, to assess any difference that may be present due to the variables utilized. This research utilized three imputed datasets (justiﬁcation discussed below), which were ﬁrst corrected for size/sex effects (discussed above). Each imputed size/sex-adjusted dataset was then analyzed with the RMET 4.0 program to produce unbiased, estimated genetic distance (D2) values and their standard errors (Relethford and Blangero, 1990; Relethford et al., 1997). Though not exactly equivalent to Mahalanobis distances, the estimated genetic D2 values are proportional if all populations are weighted equally. An equal population size weighting scheme was utilized in this analysis. The RMET program allows for the assignment of trait heritabilities which produce conservative estimations of population distance, based on phenotypic traits that are not under 100% genetic control. An average heritability value of h2 ⫽ 0.55 was utilized in this analysis (Relethford, 1994; Relethford and Harpending, 1994). The D2 values were corrected for bias, using a standard bias correction provided in the program (Relethford et al., 1997). The D2 values and associated standard errors calculated from the RMET analyses from the three imputed datasets were then analyzed via the “MI (Multiple Imputation) Inference: Scalar Estimands” method in NORM 2.01. This method combines the results (estimates and standard errors) from individual analyses into a single set of results (Rubin, MARQUESAN CRANIAL VARIATION TABLE 3. Standardized MMD scores for Marquesas Islands1 Location n Nuku Hiva Fatuiva Tahuata/Hiva ‘Oa Nuku Hiva Fatuiva Tahuata/Hiva ‘Oa 20.2 19.7 19.4 0.000 1.850 1.491 0.000 0.259 0.000 n ⫽ total sample size. Sample size for each attribute is counted separately, and then averaged to determine overall sample size for each location. 1 1987; Schafer, 1997). When performing a multipleimputed analysis, the variation in results across the imputed datasets reﬂects statistical uncertainty due to missing data (Rubin, 1987). Rubin (1987) shows that the efﬁciency of an estimate based on m imputations is approximately 冉 1 ⫹ ␥ m 冊 ⫺1 where ␥ is the estimated rate of missing information, a value calculated by NORM (for derivation of ␥, see Rubin, 1987). The estimated rate of missing information is the relative increase in variance due to nonresponse. The rate of missing information, along with the number of imputations (m), determines the relative efﬁciency of the MI inference. The percent efﬁciency achieved for the estimated results was in excess of 98% (␥ ⬵ 0.061; m ⫽ 3). RESULTS Cranial discrete traits The cranial nonmetric analysis focused on three groups within the Marquesas Islands: the Fatuiva and Tahuata/Hiva ‘Oa in the southeast, and the Nuku Hiva in the northwest. The discrete trait frequencies for the samples are presented in Table 1. The standardized MMD scores are displayed in Table 3, showing the general patterns of similarity. Examination of the MMD scores matrix indicates that Fatuiva is more similar to Tahuata/Hiva ‘Oa than either is to Nuku Hiva. Statistically signiﬁcant MMD scores are conventionally those with values greater than 2.0 (de Souza and Houghton, 1977). None of the MMD scores were statistically signiﬁcantly different from zero, yet the scores for the Nuku Hiva-Fatuiva (1.850) and Nuku HivaTahuata/Hiva ‘Oa (1.491) pairings were nearly signiﬁcant. Craniometric traits The cranial metric analysis focuses on the same three island populations used in the analysis of cranial nonmetric traits. The averaged unbiased, estimated genetic D2 values and standard errors are presented in Table 4. The values clearly indicate that the Nuku Hiva sample was nearly equally divergent from the Fatuiva and Tahuata/Hiva ‘Oa samples, while the Fatuiva and Tahuata/Hiva ‘Oa samples were relatively close to each other. These results support those obtained by cranial discrete 325 trait analyses. Each unbiased, estimated genetic D2 value was tested to determine if the value was signiﬁcant using the Z-distribution test method, in which the distance value is divided by its standard error. The unbiased, estimated genetic D2 values for the Nuku Hiva-Fatuiva (0.1328, 46-variable analysis; 0.2726, seven-variable analysis) and the Nuku Hiva-Tahuata/Hiva ‘Oa (0.0813, 46-variable analysis; 0.2225, seven-variable analysis) pairings were stastically signiﬁcant, while the Fatuiva-Tahuata/ Hiva ‘Oa (0.0433; 0.0427) pairings were not statistically signiﬁcantly different from zero, using the Zdistribution test method. The D2 value results clearly indicate a greater difference between the Nuku Hiva sample and the Fatuiva and Tahuata/ Hiva ‘Oa samples, than between the Fatuiva and Tahuata/Hiva ‘Oa samples. The results of the 46-variable analysis and the seven-variable analysis provide similar patterns in the similarity/dissimilarity of the Marquesas Islands samples analyzed. Though the unbiased, estimated genetic D2 values differ in terms of absolute magnitude, they are quite similar in their relative magnitudes, patterning, and signiﬁcance (Table 4). There does not appear to have been a loss of diagnostic information through the reduction of the dataset from 46 variables to seven variables. The relationship of ‘Ua Huka to northwest and southeast Marquesas Island groups Conventionally, ‘Ua Huka has been included with the northwest Marquesas Island group. Comparisons between the archaeological material from the Hane Site (‘Ua Huka) and the Ha‘atuatua (Nuku Hiva) demonstrated similarity and continuity in culture on those two islands (Rolett, 1993), and were the principal sources for establishing the northern Marquesas cultural sequence (Sinoto, 1970). The earliest colonization of the Marquesas Islands appears to have occurred in the northwest island group, with subsequent dispersals into the southeast island group, yet both island groups appear to have possessed a basically homogeneous material culture (Rolett, 1993; Sinoto, 1970). The observed cultural similarities between ‘Ua Huka and Nuku Hiva might presume a correlate in the physical similarity of the islands’ inhabitants. The cranial sample from ‘Ua Huka was not included in the discrete trait analysis due to insufﬁcient sample size (n ⫽ 9). As discussed previously, MMD analyses are susceptible to bias due to small and unequal sample sizes between groups (Chapman, 1998). In order to maintain consistency, only groups with sample sizes of 20 or more were included in the cranial discrete analyses. Additionally, the samples analyzed were maintained at n ⫽ 20. However, a sample from ‘Ua Huka was included in analyses utilizing craniometric data (n ⫽ 22). The methods used in the craniometric analyses take into account differing sample sizes, and make the necessary bias corrections (Relethford et al., 1997). 326 V.H. STEFAN AND P.M. CHAPMAN TABLE 4. Estimated genetic distance (D2) value estimates and standard errors for Marquesas Islands1 Location n Nuku Hiva Fatuiva Tahuata/Hiva ‘Oa 97 29 30 Nuku Hiva Fatuiva Tahuata/Hiva ‘Oa 2 0.2225 (0.0929)2 0.0427 (0.0762) 0.2726 (0.0718) 0.1328 (0.0270)3 0.0813 (0.0146)3 0.0433 (0.0185) 1 n ⫽ total sample size. Standard errors in parentheses. Values resulting from “MI (Multiple Imputation) Inference: Scalar Estimands” method that combines results (estimates and standard errors) from individual analyses into a single set of results. Upper right triangle, seven-variable analysis; lower left triangle, 46-variable analysis. 2 Signiﬁcant estimated genetic D2 value: Z(0.05, 7) ⫽ 2.365. 3 Signiﬁcant estimated genetic D2 value: Z(0.05, 46) ⫽ 2.015. TABLE 5. Estimated genetic distance (D2) value estimates and standard errors for Marquesas Islands1 Location Nuku Hiva Fatuiva Tahuata/Hiva ‘Oa ’Ua Huka n 97 29 30 22 Nuku Hiva Fatuiva Tahuata/Hiva ‘Oa 2 0.1223 (0.0168)3 0.0732 (0.0200)3 0.0829 (0.0182)3 0.2577 (0.0936) 0.0713 (0.0196)3 0.1175 (0.0840) 0.2055 (0.0975) 0.0443 (0.0831) 2 ‘Ua Huka 0.2251 (0.0834)2 0.1608 (0.2380) 0.0607 (0.0525) 0.0431 (0.0189) n ⫽ Total sample size. Standard errors in parenthesis. Values resulting from “MI (Multiple Imputation) Inference: Scalar Estimands” method that combines results (estimates and standard errors) from individual analyses into a single set of results. Upper right triangle, seven-variable analysis; lower left triangle, 46-variable analysis. 2 Signiﬁcant estimated genetic D2 value: Z(0.05, 7) ⫽ 2.365. 3 Signiﬁcant estimated genetic D2 value: Z(0.05, 46) ⫽ 2.015. 1 The results of the analyses, including the ‘Ua Huka sample, are presented in Table 5. Contrary to expectations, it is evident that the ‘Ua Huka sample has the closest similarity to the combined Tahuata/ Hiva ‘Oa sample from the southeast Marquesas Islands (0.0431; 0.0607), with the unbiased, estimated genetic D2 value estimates for the ‘Ua Huka-Nuku Hiva (0.0829; 0.2251) pairing being statistically signiﬁcantly different from zero. These results may indicate a level of intraisland interaction focused on ‘Ua Huka that is more complex then previously documented. Several linguistic and archaeological peculiarities have been noted on ‘Ua Huka that indicate a southeastern Marquesas Islands inﬂuence. It was noted that the term “marae” was used to describe an entire religious structure on Nuku Hiva, ‘Ua Pou, and other northwest Marquesas Islands, while the word “me’ae” was more commonly used in the southeast Marquesas and ‘Ua Huka (Green, 2000, p. 86; Linton, 1925, p. 31). In an archaeometric analysis of Marquesan lithic artifacts manufactured from phonolite and recovered from sites on ‘Ua Huka, Nuku Hiva, and Tahuata (Rolett et al., 1997), researchers concluded that there was a single source for the phonolite material for these artifacts that could not be positively located, but was likely from a quarry on ‘Ua Pou (northwest Marquesas) or Tahuata (southeast Marquesas). Whether the phonolite material originated from a northwest or southeast Marquesas quarry does not negate that fact that the artifacts manufactured from this material were found on islands in the northwest and southeast Marquesas, indicating material culture exchange in the island group. Handy (1923) earlier speculated on some sort of linkage between Nuku Hiva and ‘Ua Huka to Hiva ‘Oa to some degree. If that linkage existed, it is not unreasonable to conclude that there would have been some cultural and linguistic exchange between them. DISCUSSION AND CONCLUSIONS Green (1966) separates the Marquesan language into two dialects: Northwest (NW) Marquesan and Southeast (SE) Marquesan. Evidence of archaeological differences between these two regions is not as clear, although there were some cultural differences between the northern and southern islands at European contact (Handy, 1923; Linton, 1925; Rolett et al., 1997). Rolett (1989, p. 373) suggested a “widespread continuity in material culture throughout the Marquesas” until the end of the Developmental Period at about AD 1300. Rolett (1989, p. 374) further stated: The transformation of the subsistence economy of the Developmental and Expansion cultures coincides with widespread changes in technology that occurred at roughly the same time throughout the northern and southern Marquesas. Sinoto (1979, p. 131) suggested that there were also differences in “diagnostic material culture” between the northwestern and southeastern Marquesas Islands. The results of the cranial discrete trait and metric study suggest a close relationship between the two southern locations, i.e., Fatuiva and the combined Tahuata/Hiva ‘Oa sample. However, both of these groups have a statistically signiﬁcant estimated genetic D2 score with Nuku Hiva in the north. The close relationship between the southern locations, to the exclusion of Nuku Hiva, is in agreement with linguistic studies (Green, 1966; Lavondès and Randall, 1978) suggesting that a degree of differentia- 327 MARQUESAN CRANIAL VARIATION TABLE 6. Estimated geographic distance for Marquesas Islands1 Location Nuku Hiva Fatuiva Tahuata/Hiva ‘Oa Nuku Hiva Fatuiva Tahuata/Hiva ‘Oa ’Ua Huka 108 62 20 36 100 45 1 ‘Ua Huka Distances are in nautical miles. tion existed between the southeast and northwest islands in Marquesan prehistory. The biological differences discerned here between the northern and southern Marquesas Islands may reﬂect increasing isolation between the two locations, especially after the 14th century AD (Murdoch, 2000; Rolett et al., 1997). A Pearson’s correlation analysis of the estimated genetic distances (46variable analysis, Table 4) and the estimated geographic distances for the islands of Nuku Hiva, Fatuiva, and the combined Tahuata/Hiva ‘Oa (Table 6) show a strong correlation (r ⫽ 0.9974, P ⬍ 0.0001). This signiﬁcant correlation may reﬂect the effect of isolation by distance on the population afﬁnity analyses of the cranial samples from these islands. With regards to ‘Ua Huka, the linguistic, archaeological, and now biological evidence indicates an inﬂuence from the southeastern Marquesas Islands. Its geographical position between the northwestern and southeastern Marquesas Islands would have served well as a waypoint in the prehistoric system of networking in the Marquesas. Though traditionally associated with the northern Marquesas Islands group, ‘Ua Huka’s archaeology, linguistics, and physical anthropology clearly indicate that it was a player in the Marquesas Islands interaction sphere. A Pearson’s correlation analysis of estimated genetic distances (46-variable analysis, Table 5) and estimated geographic distances for the islands of Nuku Hiva, Fatuiva, the combined Tahuata/Hiva ‘Oa, and ‘Ua Huka (Table 6) also shows a strong correlation (r ⫽ 0.7692, P ⬍ 0.0001), though not as strong as the one for the previous three-group analysis. However, when just the estimated genetic and geographic distances between the island of ‘Ua Huka and the other islands are examined, the resulting correlation is not signiﬁcant (r ⫽ 0.5000, P ⫽ 0.4000). These results clearly indicate that some additional evolutionary force (gene ﬂow), in addition to isolation by geographic distance, inﬂuenced the population afﬁnities of the northwestern and southeastern Marquesas Islands, and produced a closer association between the island of ‘Ua Huka and the southeastern Marquesas Islands group, than with the northwestern Marquesas Islands group with which it is traditionally included. This paper documents the interisland cranial morphological variation following the general pattern of the traditional subdivision of the Marquesas Islands into northwestern and southeastern groups, a pattern which has its correlate in the variation of other aspects of Marquesas Islands culture. Yet the results presented also indicate that a simple model of isolation by geographic distance is insufﬁcient to explain all the population afﬁnity found in these analyses, as demonstrated by the ‘Ua Huka example. This study clearly demonstrates the utility of biological anthropology evidence combined with archaeological and linguistic evidence to clarify the prehistory of the Marquesas Islands and Eastern Polynesia and/or corroborate the results obtained from the archaeological and linguistic evidence. ACKNOWLEDGMENTS The authors acknowledge the following individuals for their assistance and access to the skeletal material examined in this study: Dr. Betty Tatar, Kevin R. Montgomery, and Valerie J. Free (Bernice P. Bishop Museum, State Museum of Natural and Cultural History, Honolulu, HI); Dr. Ian Tattersall, Dr. Kenneth M. Mowbray, and Joanne Grant (American Museum of Natural History, New York, NY); Dr. Robert Kruszynski and Dr. Chris Stringer (Human Origins Group, Natural History Museum, London, UK); Professeur André Langaney, M. Philippe Mennecier, Dr. Miya Awazu Pereira da Silva, Mme. Simone Jousse, and Mme. Anne-Marie Bacon (Laboratoire d’Anthropologie Biologique, Musée de l’Homme, Paris, France); and Mme. Maeva Navarro and Mr. Mark Eddowes (Département Archéologie du Centre Polynésien des Sciences Humaines, Tahiti). This research was supported in part by a grant from the Kon-Tiki Museum (Oslo, Norway) to P.M.C. We thank Dr. Steven R. Fischer (Director, Institute of Polynesian Languages and Literature, Auckland, New Zealand) for his reading and commentary on the manuscript, as well as those of the anonymous reviewers whose comments and suggestions resulted in a much-improved article. APPENDIX A Standardized craniofacial measurements1 Maximum cranial length Nasion-occipital length Maximum cranial breadth Maximum frontal breadth Minimum frontal breadth Bizygomatic breadth Basion-bregma height Basion-nasion length Nasion-bregma chord Bregma-lambda chord Nasion-prosthion height Nasion-alveolare2 H-GOL* H-NOL* H-XCB* H-XFB* B-WFB H-ZYB* H-BBH* H-BNL* H-FRC* H-PAC* H-NPH* B-NAL 1 B, Bass (1995); GH, Gill et al. (1988); H, Howells (1973); M, Martin and Saller (1957). Asterisk indicates measurement abbreviations from Howells (1989). Other abbreviations developed by present authors for ease of data handling and analysis. 328 V.H. STEFAN AND P.M. CHAPMAN Biasterionic breadth Basion-prosthion length Bistephanic breadth Bijugal breadth Foramen magnum length Left orbital height Left orbital breadth, dacrion Left orbital breadth, max-f Biorbital breadth Nasal height Nasal breadth Bifrontal breadth Biauricular breadth Minimum cranial breadth Auricular height Porion-bregma height Porion-nasion3 Porion-subnasale4 Porion-prosthion5 Basion-porion height Maxillofrontal breadth Maxillofrontal subtense Zygoorbital breadth Zygoorbital subtense Alpha chord Alpha subtense Simotic chord Bimaxillary breadth Bimaxillary subtense Mastoid length Mastoid width Cheek height Malar length, inferior Malar length, maximum Palatal depth Maxilloalveolar breadth6 Maxilloalveolar length H-ASB* H-BPL* H-STB* H-JUB* H-FOL* H-OBH* B-OBD H-OBB* H-EKB* H-NLH* H-NLB* H-FMB* H-AUB* H-WCB* B-AUR B-PBH H-NAR* H-SSR* H-PRR* B-BPH GH-MXB GH-MXS GH-ZOB GH-ZOS GH-ALB GH-ALS H-WNB* H-ZMB* H-SSS* H-MDH* H-MDB* H-WMH* H-IML* H-XML* M-PAD H-MAB* B-MAL 2 Upper facial height (Bass, 1995). Nasion radius. 4 Subspinale radius. 5 Prosthion radius. 6 Palate breadth, external (Howells, 1973). 3 LITERATURE CITED Bass WM. 1995. Human osteology: a laboratory and ﬁeld manual of the human skeleton, 4th ed. Columbia, MO: Missouri Archaeology Society. Bellwood PS. 1989. The colonization of the Paciﬁc: some current hypotheses. In: Hill AVS, Serjeantson SW, editors. The colonization of the Paciﬁc. The genetic trail. Research monographs on human population biology no. 7. Oxford: Clarendon Press. p 1–59. Berry AC. 1974. The use of non-metrical variants of the cranium in the study of Scandinavian population movements. Am J Phys Anthropol 40:345–358. Berry AC. 1975. Factors affecting the incidence of non-metrical skeletal variants. J Anat 120:519 –535. Berry AC, Berry RJ. 1967. Epigenetic variation in the human cranium. J Anat 101:361–379. Berry RJ. 1963. Epigenetic polymorphism in wild populations of Mus musculus. Genet Res 4:193–220. Berry RJ. 1964. The evolution of an island population of the house mouse. Evolution 18:468 – 483. Biggs BG. 1971. The languages of Polynesia. In: Sebeok TA, editor. Current trends in linguistics, volume 8: linguistics in Oceania. The Hague: Mouton. p 466 –505. Bocquet-Appel AP. 1984. Biological evolution and history in 19th century Portugal. In: van Vark GN, Howells WW, editors. Multivariate statistical methods in physical anthropology. Drodrecht: D. Reidel Publishing Co. p 289 –322. Brace CL, Hunt KD. 1990. A nonracial craniofacial perspective on human variation: A(ustralia) to Z(uni). Am J Phys Anthropol 82:341–360. Brace CL, Brace ML, Leonard WR. 1989. Reﬂections on the face of Japan: a multivariate craniofacial and odontometric perspective. Am J Phys Anthropol 78:93–113. Brace CL, Brace ML, Dodo Y, Hunt KD, Leonard WR, Li Y, Sangvichien S, Shao X, Zhang Z. 1990. Micronesians, Asians, Thais and relations: a craniofacial and odontometric perspective. Micronesica [Suppl] 2:323–348. Brace CL, Tracer DP, Hunt KD. 1991. Human craniofacial form and the evidence for the peopling of the Paciﬁc. In: Bellwood PS, editor. Indo-Paciﬁc prehistory 1990. Proceedings of the 14th Congress of the Indo-Paciﬁc Prehistory Association. IndoPaciﬁc Prehistory Association bulletin 11. Canberra: Indo-Paciﬁc Prehistory Association. p 247–269. Buikstra JE, Ubelaker DH. 1994. Standards for data collection from human skeletal remains. Arkansas Archaeological Survey research series no. 44. Fayetteville: Arkansas Archaeological Survey. Buikstra JE, Frankenberg SR, Konigsberg LW. 1990. Skeletal biological distance studies in American physical anthropology: recent trends. Am J Phys Anthropol 82:1–7. Buranarugsa M, Leach BF. 1993. Coordinate geometry of Moriori crania and comparisons with Maori. Man Culture Oceania 9:1– 43. Byard PJ, Sharma K, Russell JM, Roa DC. 1984. A family study of anthropometric traits in a Punjabi community: II. An investigation of familial transmission. Am J Phys Anthropol 64:97– 104. Cadien JD, Harris EF, Jones WP, Mandarino LJ. 1974. Biological lineages, skeletal populations, and microevolution. Yrbk Phys Anthropol 18:194 –201. Chapman PM. 1993. Analysis of non-metric cranial traits from prehistoric Easter Island with comparisons to Peru. Master’s thesis, University of Wyoming, Laramie. Chapman PM. 1998. An examination of East Polynesian population history. Ph.D. dissertation, University of Otago, Dunedin. Chapman PM, Gill GW. 1997. Easter Islander origins: non-metric cranial trait comparison between Easter Island and Peru. Rapa Nui J 11:58 – 63. Chapman PM, Gill GW. 1998. An analysis of Easter Island population history. In: Stevenson CM, Lee G, Morin FJ, editors. Easter Island in Paciﬁc Context South Seas Symposium. Proceedings of the Fourth International Conference on Easter Island and East Polynesia. The Easter Island Foundation occasional paper 4. Los Osos: Easter Island Foundation. p 143– 150. Cheverud JM. 1988. A comparison of genetic and phenotypic correlations. Evolution 42:958 –968. Cheverud JM, Buikstra JE. 1982. Quantitative genetics of skeletal nonmetric traits in the rhesus macaques of Cayo Santiago. III. Am J Phys Anthropol 59:151–155. Cheverud JM, Buikstra JE, Twichell E. 1979. Relationships between non-metric skeletal traits and cranial size and shape. Am J Phys Anthropol 50:191–198. Corruccini RS. 1974. An examination of the meaning of cranial discrete traits for human skeletal biological studies. Am J Phys Anthropol 40:425– 446. Corruccini RS. 1976. The interaction between nonmetric and metric cranial variation. Am J Phys Anthropol 44:285–294. de Souza P, Houghton P. 1977. The mean measure of divergence and the use of non-metric data in the estimation of biological distance. J Archaeol Sci 4:163–169. MARQUESAN CRANIAL VARIATION Devor EJ, McGue M, Crawford MH, Lin PM. 1986a. Transmissible and nontransmissible components of anthropometric variation in the Alexanderwohl Mennonites: I. Description and familial correlations. Am J Phys Anthropol 69:71– 82. Devor EJ, McGue M, Crawford MH, Lin PM. 1986b. Transmissible and nontransmissible components of anthropometric variation in the Alexanderwohl Mennonites: II. Resolution by path analysis. Am J Phys Anthropol 69:83–92. Donnelly SM, Konigsberg LW, Stringer CB. 1998. Interpretation of population structure when group structure is unknown. Am J Phys Anthropol 26:106 [abstract]. Droessler JB. 1981. Craniometry and biological distance: biocultural continuity and change at the Late-Woodland-Mississippian interface. Center for American Archeology research series, volume 1. Evanston: Center for American Archeology. Freeman OW, editor. 1951. Geography of the Paciﬁc. New York: John Wiley and Sons. Giles E. 1973. Population analyses in Oceania. In: Crawford MH, Workman PL, editors. Methods and theories of anthropological genetics. Albuquerque: University of New Mexico Press. p 389 – 402. Gill GW, Owsley DW. 1993. Human osteology of Rapanui. In: Fischer SR, editor. Easter Island studies, contributions to the history of Rapanui in memory of William T. Mulloy. Oxbow monograph 32. Oxford: Oxbow Books. p 56 – 62. Gill GW, Hughes SS, Bennett SM, Gilbert BM. 1988. racial identiﬁcation from the midfacial skeleton with special reference to American Indians and whites. J Forensic Sci 13:92–99. Green RC. 1966. Linguistic subgrouping within Polynesia: the implications for prehistoric settlement. J Polynes Soc 75:6 –38. Green RC. 1993. Tropical Polynesian prehistory: where are we now? In: Spriggs M, Yen DE, Ambrose W, Jones R, Thorne A, Andrews A, editors. A community of culture: the people and prehistory of the Paciﬁc. The Australian National University occasional papers in prehistory 21. Canberra: Department of Prehistory, Research School of Paciﬁc Studies, Australian National University. p 219 –237. Green RC. 2000. Religious structures of southeastern Polynesia: even more Marae later. In: Wallin P, Martinsson-Wallin H, editors. Essays in honour of Arne Skjølsvold 75 years. The Kon-Tiki Museum occasional papers, volume 5. Oslo: Kon-Tiki Museum, Institute for Paciﬁc Archaeology and Cultural History. p 83–100. Green RC, Suchey JM, Gokhale DV. 1979. The statistical treatment of correlated bilateral traits in the analysis of cranial material. Am J Phys Anthropol 50:629 – 634. Grünenberg H. 1963. The pathology of development. Oxford: Blackwell. Hagelberg E, Clegg JB. 1993. Genetic polymorphisms in prehistoric Paciﬁc Islanders determined by analysis of ancient bone DNA. Proc R Soc Lond [Biol] 252:163–170. Handy ESC. 1923. The native culture in the Marquesas. Bernice P. Bishop Museum bulletin 9. Honolulu: Bernice P. Bishop Museum. Hanihara T. 1992. Dental and cranial afﬁnities among populations of East Asia and the Paciﬁc: the basic populations in East Asia, IV. Am J Phys Anthropol 88:163–182. Hanihara T. 1996. Comparison of craniofacial features of major human groups. Am J Phys Anthropol 99:389 – 412. Hanihara T. 1997. Craniofacial afﬁnities of Mariana Islanders and Circum-Paciﬁc peoples. Am J Phys Anthropol 104:411– 425. Harlow RB. 1979. Regional variation in Maori. N Z J Archaeol 1:123–138. Harlow RB. 1994. Maori dialectology and the settlement of New Zealand. In: Sutton DG, editor. The origins of the ﬁrst New Zealanders. Auckland: Auckland University Press. p 106 –122. Harpending HC, Jenkins T. 1973. Genetic distance among Southern African populations. In: Crawford MH, Workman PL, editors. Methods and theories of anthropological genetics. Albuquerque: University of New Mexico Press. p 177–200. Houghton P. 1980. The ﬁrst New Zealanders. Auckland: Hodder & Stoughton. Houghton P. 1989. Watom: the people. Rec Aust Mus 41:223–233. 329 Houghton P. 1990. The adaptive signiﬁcance of Polynesian body form. Ann Hum Biol 17:19 –23. Houghton P. 1991a. The early human biology of the Paciﬁc: some considerations. J Polynes Soc 100:167–196. Houghton P. 1991b. Selective inﬂuences and morphological variation amongst Paciﬁc Homo sapiens. J Hum Evol 21:49 –59. Howells WW. 1953. Correlations of brothers in factor scores. Am J Phys Anthropol 7:121–140. Howells WW. 1970. Anthropometric grouping analysis of Paciﬁc peoples. Archaeol Phys Anthropol Oceania 5:192–217. Howells WW. 1973. Cranial variation in man. A study by multivariate analysis of patterns of difference among recent human populations. Papers of the Peabody Museum of Archaeology and Ethnology, volume 67. Cambridge, MA: Peabody Museum, Harvard University. Howells WW. 1979. Physical anthropology. In: Jennings JD, editor. The prehistory of Polynesia. Canberra: Australian National University Press. p 271–285. Howells WW. 1989. Skull shapes and the map. Craniometric analyses in the dispersion of modern Homo. Papers of the Peabody Museum of Archaeology and Ethnology, volume 79. Cambridge, MA: Peabody Museum, Harvard University. Howells WW. 1990. Micronesia to Macromongolia: Micro-Polynesian craniometrics and the Mongoloid population complex. Micronesica [Suppl] 2:363–372. Howells WW. 1995. Who’s who in skulls. Ethnic identiﬁcation of crania from measurements. Papers of the Peabody Museum of Archaeology and Ethnology, volume 82. Cambridge, MA: Peabody Museum, Harvard University. Hughes HGA, Fischer SR, editors. 1998. William Pascoe Crook, Samuel Greatheed, and Tima’u Te’ite’i. An essay toward a dictionary and grammar of the Lesser-Australian language, according to the dialect used at the Marquesas (1799). Monograph series of the Institute of Polynesian Languages and Literature. Auckland: Institute of Polynesian Languages and Literature. Jennings JD, editor. 1979. The prehistory of Polynesia. Canberra: Australian National University Press. Katayama K. 1987. Physical anthropology in Polynesia. Man Culture Oceania 3:1–18. Katayama K. 1990. A scenario on prehistoric Mongoloid dispersals into the South Paciﬁc, with special reference to hypothetic Proto-Oceanic connection. Man Culture Oceania 6:151–159. Key PJ. 1983. Craniometric relationships among Plains Indians. Report of investigation no. 34. Knoxville: Department of Anthropology, University of Tennessee. Key PJ, Jantz RL. 1990. Statistical measures of intrasample variability. Hum Evol 5:457– 469. Kohn LAP. 1991. The role of genetics in craniofacial morphology and growth. Annu Rev Anthropol 20:261–278. Konigsberg LW. 1987. Population genetics models interpreting prehistoric intracemetery biological variation. Ph.D. dissertation, Northwestern University, Evanston, IL. Konigsberg LW. 1990. Analysis of prehistoric biological variation under a model of isolation by geographic and temporal distance. Hum Biol 62:49 –70. Konigsberg LW, Blangero J. 1993. Multivariate quantitative genetic simulations in anthropology with an example from the South Paciﬁc. Hum Biol 65:897–915. Konigsberg LW, Ousley SD. 1995. Multivariate quantitative genetics of anthropometric traits from the Boas data. Hum Biol 67:481– 498. Konigsberg LW, Kohn LAP, Cheverud JM. 1993. Cranial deformation and nonmetric trait variation. Am J Phys Anthropol 90:35– 48. Korey KA. 1980. The incidence of bilateral nonmetric skeletal traits: a reanalysis of sampling procedures. Am J Phys Anthropol 53:19 –23. Krogman WM, Iscan MY. 1986. The human skeleton in forensic medicine, 3rd ed. Springﬁeld, IL: C.C. Thomas. Lavondès H, Randall J. 1978. Les noms de poissons Marquisiens. J Soc Oceanistes 34:79 –111. Linton R. 1923. The material culture of the Marquesas Islands. Memoirs of the Bernice P. Bishop Museum, volume 8, no. 5. 330 V.H. STEFAN AND P.M. CHAPMAN Bayard Dominick Expedition publication number 5. Honolulu: Bernice P. Bishop Museum. Linton R. 1925. Archaeology of the Marquesas Islands. Bernice P. Bishop Museum bulletin 23. Bayard Dominick Expedition publication number 10. Honolulu: Bernice P. Bishop Museum. Marck J. 1996. Eastern Polynesian subgrouping today. In: Davidson JM, Irwin GJ, Leach BF, Pawley A, Brown D, editors. Oceanic culture history: essays in honour of Roger Green. New Zealand Journal of Archaeology special publication. Dunedin: New Zealand Journal of Archaeology. p 491–511. Martin R, Saller K. 1957. Lehrbuch der Anthropologie, revised 3rd ed. Stuttgart: Gustav Fischer Verlag. Molto JE. 1983. Biological relationships of Southern Ontario Woodland peoples: the evidence of discontinuous cranial morphology. National Museum of Man Mercury series. Ottawa: National Museum of Man. Murdoch K. 2000. An ethnohistory of Polynesian voyaging. Master’s thesis, University of Otago, Dunedin. Najem WJ. 1997. The correlation of craniofacial form between siblings. Master’s thesis, Case Western Reserve University, Cleveland, OH. Olivier G. 1969. Practique anthropologie. Paris: Vigot Frères. Ossenberg NS. 1969. Discontinuous morphological variation in the human cranium. Ph.D. dissertation, University of Toronto, Toronto. Ossenberg NS. 1970. The inﬂuence of artiﬁcial cranial deformation on discontinuous morphological traits. Am J Phys Anthropol 33:357–372. Ossenberg NS. 1976. Within and between race distances in population studies based on discrete traits of the human skull. Am J Phys Anthropol 45:701–716. Ossenberg NS. 1977. Congruence of distance matrices based on cranial discrete traits, cranial measurements, and linguisticgeographic criteria in ﬁve Alaskan populations. Am J Phys Anthropol 47:93–98. Pardoe C. 1984. Prehistoric human mophological variation in Australia. Ph.D. dissertation, Australian National University, Canberra. Pardoe C. 1991. Isolation and evolution in Tasmania. Curr Anthropol 32:1–21. Pietrusewsky M. 1971. Application of distance statistics to anthroposcopic data and a comparison of results with those obtained by using discrete traits of the skull. Archaeol Phys Anthropol Oceania 6:21–33. Pietrusewsky M. 1976. Prehistoric human skeletal remains from Papua New Guinea and the Marquesas. Asian and Paciﬁc archaeology series, number 7. Manoa: Social Science and Linguistic Institute, University of Hawaii at Manoa. Pietrusewsky M. 1977. Étude des relations entre les populations du Paciﬁque par les méthodes d’analyse multivariée appliquées aux variations crâniennes. Anthropologie 81:67–97. Pietrusewsky M. 1983. Multivariate analysis of New Guinea and Melanesian skulls: a review. J Hum Evol 12:61–76. Pietrusewsky M. 1984. Metric and non-metric cranial variation in Australian Aboriginal populations compared with populations from the Paciﬁc and Asia. Occasional papers in human biology 3. Canberra: Australian Institute of Aboriginal Studies. Pietrusewsky M. 1988. Multivariate comparisons of recently excavated Neolithic human crania from the Socialist Republic of Vietnam. Int J Anthropol 3:267–283. Pietrusewsky M. 1990a. Craniometric variation in Micronesia and the Paciﬁc: a multivariate study. Micronesica [Suppl] 2:373– 402. Pietrusewsky M. 1990b. Craniofacial variation in Australasian and Paciﬁc populations. Am J Phys Anthropol 82:319 –340. Pietrusewsky M. 1990c. Lapita-associated skeletons from Waton Island, Papua New Guinea, and the origins of the Polynesians. Asian Perspect 28:83–90. Pietrusewsky M. 1992. Japan, Asia and the Paciﬁc: a multivariate craniometric investigation. In: Hanihara K, editor. Japanese as a member of the Asian and Paciﬁc populations. International symposium no. 4. Kyoto: International Research Center for Japanese Studies. p 9 –52. Pietrusewsky M. 1994. Paciﬁc-Asian relationships: a physical anthropological perspective. Ocean Ling 33:407– 429. Pietrusewsky M. 1996. The physical anthropology of Polynesia: a review of some cranial and skeletal studies. In: Davidson JM, Irwin GJ, Leach BF, Pawley A, Brown D, editors. Oceanic culture history: essays in honour of Roger Green. New Zealand Journal of Archaeology special publication. Dunedin: New Zealand Journal of Archaeology. p 343–353. Pietrusewsky M. 1997. Biological origins of Hawaiians: evidence from skulls. Man Culture Oceania 13:1–37. Pietrusewsky M. n.d. Hawaiian ancestors: multivariate studies of crania. Anthropology Library, University of Otago. Dunedin. Pietrusewsky M, Ikehara-Quebral RM. 2000. Multivariate comparisons of Rapa Nui (Easter Island), Polynesian, and CircumPolynesian crania. Rapa Nui J 14:63 [abstract]. Pietrusewsky M, Li Y, Shao X, Nguyen QQ. 1992. Modern and near modern populations of Asia and the Paciﬁc: a multivariate craniometric interpretation. In: Akazawa T, Aoki K, Kimura T, editors. The evolution and dispersal of modern humans in Asia. Tokyo: Hokusen-sha. p 531–558. Prowse TL, Lovell NC. 1996. Concordance of cranial and dental morphological traits and evidence for endogamy in ancient Egypt. Am J Phys Anthropol 101:237–246. Relethford JH. 1994. Craniometric variation among modern human populations. Am J Phys Anthropol 95:53– 62. Relethford JH. 1996. Genetic drift can obscure population history: problem and solution. Hum Biol 68:29 — 44. Relethford JH. 1999. Models, predictions, and the fossil record of modern human origins. Evol Anthropol 8:7–10. Relethford JH, Blangero J. 1990. Detection of differential gene ﬂow from patterns of quantitative variation. Hum Biol 62:5–25. Relethford JH, Harpending HC. 1994. Craniometric variation, genetic theory, and modern human origins. Am J Phys Anthropol 95:249 –270. Relethford JH, Crawford MH, Blangero J. 1997. Genetic drift and gene ﬂow in post-Famine Ireland. Hum Biol 69:443– 465. Rolett BV. 1989. Hanamiai: changing subsistence and ecology in the prehistory of Tahuata (Marquesas Islands, French Polynesia). Ph.D. dissertation, Yale University, New Haven. Rolett BV. 1993. Marquesan prehistory and the origins of East Polynesian culture. J Soc Oceanistes 96:29 – 47. Rolett BV, Conte E, Pearthree E, Sinton JM. 1997. Marquesan voyaging: archaeometric evidence for inter-island contact. In: Weisler MI, editor. Prehistoric long-distance interaction in Oceania: an interdisciplinary approach. New Zealand Archaeological Association monograph volume 21. Auckland: New Zealand Archaeological Association. p 134 –148. Rubin DB. 1987. Multiple imputation for nonresponse in surveys. New York: J. Wiley & Sons. SAS Institute. 1990. SAS/STAT user’s guide, version 6, 4th ed. Cary, NC: SAS Institute, Inc. Saunders SR. 1989. Nonmetric skeletal variation. In: Iscan MY, Kennedy KAR, editors. Reconstruction of life from the skeleton. New York: Alan R. Liss, Inc. p 95–108. Schafer JL. 1997. Analysis of incomplete multivariate data. Monographs on statistics and applied probability 72. London: Chapman & Hall. Schafer JL. 1999. NORM: multiple imputation of incomplete multivariate data under a normal model, version 2. Software for Windows 95/98/NT. University Park: Pennsylvania State University. Scott JH. 1893. Contribution to the osteology of the Aborigines of New Zealand and of the Chatham Islands. Trans N Z Inst 26:1– 64. Self SG, Leamy L. 1978. Heritability of quasi-continuous skeletal traits in a random-bred population of house mice. Genetics 88:109 –120. Shima G, Suzuki M. 1967. Problems of race formation of the Maori and Moriori in terms of skulls. Osaka City Med J 13:9 – 54. Sinoto YH. 1970. An archaeological based assessment of the Marquesas as a dispersal center in East Polynesia. In: Green RC, Kelly M, editors. Studies in Oceanic culture history. Volume 1. MARQUESAN CRANIAL VARIATION Paciﬁc anthropological record no. 11. Honolulu: Bernice P. Bishop Museum. p 105–132. Sinoto YH. 1979. The Marquesas. In: Jennings JD, editor. The prehistory of Polynesia. Canberra: Australian National Museum Press. p 110 –134. Sjøvold T. 1984. A report on the heritability of some cranial measurements and non-metric traits. In: van Vark GN, Howells WW, editors. Multivariate statistical methods in physical anthropology. Dordrecht: D. Reidel Publishing Co. p 223–246. Sjøvold T. 1986. Infrapopulation distances and genetics of nonmetrical traits. In: Herrmann B, editor. Innovative trends in der prehistoric anthropology, Band 7. Mitteilungen der Berliner Gesellschaft für Anthropologie, Ethnologie und Urgeschichte. Berlin: Berliner Gesellschaft für Anthropologie, Ethnologie und Urgeschichte. Stefan VH. 1999. Craniometric variation and homogeneity in prehistoric/protohistoric Rapa Nui (Easter Island) regional populations. Am J Phys Anthropol 110:407– 419. Stefan VH. 2000. Craniometric variation and biological afﬁnity of the prehistoric Rapanui (Easter Islanders): their origin, evolution, and place in Polynesian prehistory. Ph.D. dissertation, University of New Mexico, Albuquerque. Stefan VH, Gill GW, Owsley DW. 1998. Craniometric variability among prehistoric Easter Island regional populations and their discrimination. In: Stevenson CM, Lee G, Morin FJ, editors. Easter Island in Paciﬁc Context South Seas Symposium. Proceedings of the Fourth International Conference on Easter Island and East Polynesia. The Easter Island Foundation occasional paper 4. Los Osos: Easter Island Foundation. p 151– 155. Sullivan LR. 1923a. The racial diversity of the Polynesian peoples. J Polynes Soc 32:79 – 84. 331 Sullivan LR. 1923b. New light on the races of Polynesia. Asia 23:17–20. Sullivan LR. 1923c. Marquesan somatology with comparative notes on Samoa and Tonga. Memoirs of the Bernice P. Bishop Museum, volume 9, no. 2. Bayard Dominick Expedition publication number 6. Honolulu: Bernice P. Bishop Museum. Susanne C. 1977. Heritability of anthropological characters. Hum Biol 49:573–580. Susanne C, Defrise-Gussenhoven E, Van Wanseele P, Tassin A. 1983. Genetic and environmental factors in head and face measurements in Belgian twins. Acta Genet Med Gemellol (Roma) 32:229 –238. Tagaya A, Katayama K. 1988. Craniometric variation in Polynesians: multivariate positioning of male crania from Mangia, Cook Islands. Man Culture Oceania 4:75– 89. Tigner LF, Gill GW. 1986. Postcranial discrete traits of prehistoric Easter Island populations. Am J Phys Anthropol 69:272 [abstract]. van Vark GN, Schaafsma W. 1992. Advances in the quantitative analysis of skeletal morphology. In: Saunders SR, Katzenberg MA, editors. Skeletal biology of past peoples. New York: WileyLiss, Inc. p 225–257. Vandenberg SG. 1962. How “stable” are heritability estimates? A comparison of heritability estimates from six anthropometric studies. Am J Phys Anthropol 20:331–338. Williams-Blangero S, Blangero J. 1989. Anthropometric variation and the genetic structure of the Jirels of Nepal. Hum Biol 61:1–12. Zimple C, Gill GW. 1986. A comparison of discrete cranial traits among late prehistoric Easter Islanders. Am J Phys Anthropol 69:283 [abstract].