Characterization of population structure from the mitochondrial DNA vis--vis language and geography in Papua New Guinea.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 142:613–624 (2010) Characterization of Population Structure from the Mitochondrial DNA Vis-à-Vis Language and Geography in Papua New Guinea Esther J. Lee,1* George Koki,2 and D. Andrew Merriwether1 1 2 Department of Anthropology, Binghamton University, Binghamton, NY 13902-6000 Papua New Guinea Institute of Medical Research, Goroka EHP 441, Papua New Guinea KEY WORDS Papua New Guinea; mtDNA; language; geography ABSTRACT Situated along a corridor linking the Asian continent with the outer islands of the Paciﬁc, Papua New Guinea has long played a key role in understanding the initial peopling of Oceania. The vast diversity in languages and unique geographical environments in the region have been central to the debates on human migration and the degree of interaction between the Pleistocene settlers and newer migrants. To better understand the role of Papua New Guinea in shaping the region’s prehistory, we sequenced the mitochondrial DNA (mtDNA) control region of three populations, a total of 94 individuals, located in the East Sepik Province of Papua New Guinea. We analyzed these samples with a large data set of Oceania populations to examine the role of geography and language in shaping population structure within New Papua New Guinea constitutes the eastern half of island New Guinea and includes 600 surrounding islands, including the large volcanic islands of the Bismarck Archipelago and Bougainville. The western part of the island, Western New Guinea, has been referred to as Irian Jaya and is currently part of Indonesia. Our use of ‘‘New Guinea’’ is restricted to the eastern part of the island and excludes the surrounding islands in this work, whereas we use ‘‘Papua New Guinea’’ when referring to the politically deﬁned region including the surrounding areas. We use the term ‘‘Near Oceania’’ to denote the areas of New Guinea, Bismarck Archipelago and main islands of the Solomon Islands. Islands to the south and east of the Solomon Islands constitute Remote Oceania. Lying in the corridor to Remote Oceania, Papua New Guinea has been the focus of attention in understanding patterns of migration into the Paciﬁc especially with regards to its cultural, linguistic, and genetic diversity. The earliest archaeological evidence of human occupation in New Guinea comes from the Huon Peninsula, located on the east coast, dated to just under 40,000 years before present (YBP) (Groube et al., 1986; O’Connell and Allen, 2004). Human occupation in the interior mountainous regions of the island occurred by 30,000 YBP in sites such as Nombe and Kosipe (White et al., 1970; Gillieson and Mountain, 1983; Mountain, 1991; Chappell, 2000). Other sites in the Bismarck Archipelago, including New Britain and New Ireland, are dated as early if not earlier than sites discovered in the mainland [review in Summerhayes (2007)]. Interaction between mainland New Guinea and its surrounding C 2010 V WILEY-LISS, INC. Guinea and between the region and Island Melanesia. Our results from median-joining networks, star-cluster age estimates, and population genetic analyses show that while highland New Guinea populations seem to be the oldest settlers, there has been signiﬁcant gene ﬂow within New Guinea with little inﬂuence from geography or language. The highest genetic division is between Papuan speakers of New Guinea versus East Papuan speakers located outside of mainland New Guinea. Our study supports the weak language barriers to genetic structuring among populations in close contact and highlights the complexity of understanding the genetic histories of Papua New Guinea in association with language and geography. Am J Phys Anthropol 142:613–624, 2010. V 2010 C Wiley-Liss, Inc. islands is documented by the movement of animals and plants beginning around 20,000 YBP (Summerhayes, 2007). Exchange between highland and coastal areas of New Guinea included not only cultigens but also shell and stone (White, 1972). Natural sources of obsidian only occur in West New Britain (Admiralties and Fergusson Island), but obsidian is found in the eastern highlands dated to 4,500 YBP. This illustrates the evidence of interaction between New Guinea and its surrounding islands at different times in history (Summerhayes et al., 1998). It was not until around 3,300 YBP that populations are suggested to have started migrating into Remote Oceania. This migration is thought to have introduced a new culture and interacted with those already settled along their migration route (Green, 1991). Languages in Oceania are classiﬁed into Austronesian and Papuan (also referred as non-Austronesian) language families. The Austronesian language family contains a large number of languages and is widespread, Additional Supporting Information may be found in the online version of this article. *Correspondence to: Esther J. Lee, Department of Anthropology, Binghamton University, PO Box 6000, Binghamton, NY 13902-6000. E-mail: firstname.lastname@example.org Received 8 June 2009; accepted 11 January 2010 DOI 10.1002/ajpa.21284 Published online 3 May 2010 in Wiley InterScience (www.interscience.wiley.com). 614 E.J. LEE ET AL. found as far west as Madagascar in the Indian Ocean to Easter Island in Remote Oceania to the east (Pawley, 2002). It has been proposed that Austronesian speakers came from the island of Taiwan and rapidly migrated into Melanesia and outward toward Micronesia and Polynesia (Blust, 1985; Gray et al., 2009). Although Austronesian languages in Oceania are fairly well understood concerning their history in correlation with archaeological events, far less is understood about Papuan languages. Their geographical range, linguistic, and cultural diversity point to a longer history in situ than Austronesian languages, but little is known of their origins and expansions (Ross, 2005). Papuan languages are predominant in mainland New Guinea, with no relatives outside of the Melanesia-East Indonesia region, and their 40-odd various language groups are not believed to have close relationships to each other (Pawley, 2007). It has been suggested that speakers of both families ﬁrst encountered one another in the Timor area as evidenced by the Austronesian languages of east Nusantara showing signs of their speakers’ bilingualism in Papuan languages (Ross, 2005). Furthermore, evidence of contact between Austronesian speakers and Papuan speakers of New Guinea and Island Melanesia (IM) is thought to be reﬂected in the Austronesian languages spoken in Near Oceania as well as in Micronesia and Polynesia (Ross, 2005). Among Papuan speakers in New Guinea, some suggest that Torricelli speakers, located in the highlands of northern New Guinea, are likely to have been solely descended from the initial Pleistocene occupants in the region, based on the idea that their language appears to be the oldest in Papua New Guinea (Swadling, 1990). Genetic studies using mitochondrial DNA (mtDNA), Y-chromosomal, and autosomal data suggest substantial interaction between people in New Guinea with migrants coming from East Asia/Taiwan en route to populate Polynesia. The genetic diversity of populations in Near Oceania has been documented by many (Stoneking et al., 1990; Merriwether et al., 2005; Ohashi et al., 2006; Friedlaender et al., 2007; Friedlaender et al., 2008; Kayser et al., 2008). In particular, admixture between Polynesian ancestors and indigenous Melanesians is thought to have taken place in the coastal/lowland of New Guinea and in the Bismarck Archipelago (Scheinfeldt et al., 2006; Friedlaender et al., 2007). Scholars have often correlated patterns of cultural, linguistic, and biological diversity as explained by two major ‘‘waves,’’ which have been distinct temporal events separated by thousands of years. Some have characterized these two waves as representing Austronesian-speakers versus Papuan-speakers. It has been argued that the people characterized by a variant of mtDNA haplogroup B, also known as the ‘‘Polynesian-motif,’’ did not often penetrate into the interiors and especially not into the highlands of mainland New Guinea and the nearby larger islands of New Ireland, New Britain, and Bougainville (Stoneking et al., 1990; Friedlaender et al., 2005). Previous genetic studies of populations in Papua New Guinea have shown great diversity in genetic variation and considerable antiquity evidenced by deep roots in the major mtDNA haplogroups (Ingman and Gyllensten, 2003; Easteal et al., 2005; Friedlaender et al., 2005, 2007, 2008; Merriwether et al., 2005). It has been suggested that a small number of people founded the current populations in New Guinea consisting the major maternal lineages, haplogroups P, Q, and M (M27–29) (Easteal et al., 2005; American Journal of Physical Anthropology Merriwether et al., 2005). These groups are the most widespread in New Guinea and suggested to have been the founding population for Papuan speakers. Star-cluster age estimates of these haplogroups correlate with the earliest date of ﬁrst settlement in the region (Friedlaender et al., 2005). On the other hand, early studies suggested a signiﬁcant structuring of mtDNA variation between highland and coastal PNG populations (Stoneking et al., 1990). Based on these observations, it has been suggested that the highland New Guinea populations are remnants of the original, older population that ﬁrst settled in the region (Easteal et al., 2005). In this study, we examine a number of these assumptions about the relationship between language groups and geography using newly collected mtDNA sequence data for 94 individuals from three populations in the East Sepik Province of Papua New Guinea, Dreikikir, Jama-Sepik Plains, and Kubalia. The samples are from a region with great linguistic diversity and while previous studies have examined populations from the vicinity, they are unlikely to be representative of the region’s history. We examined the mtDNA, because it has the largest available dataset of populations already sequenced for comparison and is thought to mutate fast enough for mutations to correspond to the relatively recent linguistic changes over the last 45,000 years. In addition, because our samples are from an old archival plasma collection, mtDNA was most likely to be successfully recovered than carrying out Y-chromosome typing or whole mtDNA genome sequencing. We report the haplogroups present in the region and estimate the ages in each region based on the diversity and phylogeny of each haplogroup to determine the likely order of the peopling of Near Oceania. We tested for genetic structuring based on linguistic and geographical groupings using F-statistics and AMOVA analyses. Nucleotide diversity, in conjunction with tests of neutrality, for each population was used to identify the most variable (and oldest) and least bottlenecked populations in each region. METHODS Population samples Blood samples were collected by one of the authors (GK) and Kuldeep Bhatia from the Institute of Medical Research in Goroka, Papua New Guinea, in compliance with institutional review boards. Genealogy information was obtained, so that unrelated individuals for two to three generations were selected to be sampled. Samples are from the East Sepik Province of Papua New Guinea and include the following populations: Dreikikir (n 5 28), Jama-Sepik Plains (n 5 30), and Kubalia (n 5 36). Dreikikir is located in the Torricelli Mountains close to the north coast. Jama-Sepik Plains is located in the Sepik valley, and Kubalia is situated in the hinterland, south of the township of Wewak. The location of these populations is indicated in Figure 1. All three populations speak Papuan languages. Although the language spoken in Dreikikir is classiﬁed in the Torricelli phylum, people from Jama-Sepik Plains and Kubalia speak languages of the Sepik–Ramu phylum. We screened all samples for the entire mtDNA control region as explained below. Sequences are available in Genbank at the following accession numbers: GQ202742–GQ202835. mtDNA GENETIC STRUCTURE OF PAPUA NEW GUINEA 615 Fig. 1. Map of Papua New Guinea and location of the three populations, Dreikikir, Jama-Sepik Plains, and Kubalia, along with the major township in the region, Wewak. Map is modiﬁed from maps.yahoo.com. Laboratory and statistical analyses DNA was extracted from plasma using the columnbased Qiagen extraction kit following manufacturer’s protocol (Qiagen, Valencia, CA). Primers spanning nucleotide positions (np) 15938-00429 were used for PCR ampliﬁcation of the entire mtDNA control region following standard protocols. Ampliﬁcation was veriﬁed by agarose gel electrophoresis, and successful amplicons were puriﬁed and prepared for sequencing using BigDye Terminator v.3.1 reaction kits (Applied Biosystems, Foster City, CA). Direct sequencing was carried out on an ABI PRISMTM 377XL DNA Sequencer (Applied Biosystems). Sequence data were aligned with Sequencher: Forensic Version (GeneCodes). Samples were also screened for the intergenic COII/tRNALys 9-bp deletion. Additional sequences from Papua New Guinea and Oceania were compiled using data from the literature (Friedlaender et al., 2005; Merriwether et al., 2005; Vilar et al., 2008). We only used the hypervariable region 1 (HV1) sequences for this broader comparison (np 16,019– 16,351). Sequences were aligned using ClustalX (Jeanmougin et al., 1998) and edited in MacClade 4.03 (Maddison and Maddison, 2000). Median-joining networks were calculated for haplogroups P and Q in Network 4.502 (www.ﬂuxus-engineering.com; Bandelt et al. 1999). On the basis of the distribution of samples in haplogroups P and Q median-joining networks, we describe shared and derived haplotypes based on geography and language family. We deﬁne shared haplotypes as samples that have the same haplotype while derived haplotypes are deﬁned as being one mutation away from the node located closer to the central node. Thus, either derived haplotypes are derived from the population of the haplotype closer to the central node or both derived and preceding haplotypes have a shared parental population. This is another way to infer gene ﬂow between populations. Population statistics were calculated using ARLEQUIN (Schneider et al., 2000), which included analysis of AMOVA and F-statistic values as well as diversity indices and neutrality tests. In particular, F-statistic values were generated based on shared haplotype frequencies by permutating individuals for 10,000 simulated random migrations between populations. Furthermore, F-statistic estimates were used to generate a two-dimensional scaling to illustrate the inferred distances using ALSCAL in SPSS (Statistical Package for Social Sciences, Chicago, IL). RESULTS Haplogroup characterization Fifty-nine distinct haplotypes were identiﬁed based on the mtDNA control region. The majority of haplotypes are assigned to haplogroups P and Q (Table 1). Two individuals from Kubalia are assigned to haplogroup B4a1a1a (Table S4). The two haplotypes assigned to B4 were conﬁrmed by the 9-bp deletion at the COII/tRNALys intergenic region. One B4 haplotype (RW248) shows the same mutations in HV1 that are also found in the East Sepik of New Guinea (Vilar et al., 2008). One haplotype from Jama-Sepik and one from Dreikikir are unassigned. One haplotype shows mutations at 16,362; 16,519; 73; and 185 (RW160) and American Journal of Physical Anthropology 616 E.J. LEE ET AL. the other shows mutations at 16,051; 16,086; 16,129; 16,148; 16,223; 16,362; 16,519; 73; 152; 195; 207; 263; and 269 (RW36). The latter shares mutations that are characterized for haplogroup M28 but lacks the mutation at 16,468. Thirty-one individuals are assigned to haplogroup P (33%), whereas 59 are assigned to haplogroup Q (63%) (Table 1). Among haplotypes assigned to P, one is identiﬁed as P2 with mutations at 16,362; 16,519; 73; 185; and 263. The frequencies of P and Q are similar to those that have been previously reported in New Guinea (e.g. Friedlaender et al. 2005). All identiﬁed haplotypes and their mutations are shown in Supporting Information Tables S1–S4. We compared published HV1 sequence data from the Paciﬁc with our samples for each haplogroup medianjoining network. In addition to the three populations examined in this study, the remaining populations were divided into Lowland New Guinea, Highland New Guinea, and IM, which includes all islands outside of New Guinea. Figures 2 and 3 show median-joining networks for haplogroups P and Q. The median-joining net- TABLE 1. Distribution of haplogroups from the three populations analyzed in this study Haplogroups Population Region P Q P1 P2 Q1 Q3 B M28 Other Total Dreikikir Torricelli Mt 12 Jama-Sepik Sepik Plains 9 Kubalia Sepik Plains 10 Total 31 1 1 12 2 14 6 18 6 2 44 14 2 1 1 1 1 28 30 36 94 work for haplogroup P is not clearly resolved due to a number of reticulations, which have been reported previously (Friedlaender et al., 2005). Haplogroup Q (see Fig. 3) shows two distinct clusters, which differentiates subhaplogroups Q1 and Q3. Within haplogroup P, while some individuals from Jama-Sepik share one of the central nodes and are separated by just one mutation, others are branched off at three or four mutations away from the central node. Individuals from Dreikikir are not more than two mutations away from the central node while some individuals from Kubalia are separated by three mutations from the central node. Haplogroup Q shows a similar pattern, in which the three populations are present in the central nodes, both in Q1 and Q3, while others appear as derived haplotypes in the network. To further examine shared and derived haplotypes, Tables 2 and 3 list haplotypes that are shared and derived between geographical regions for haplogroups P and Q. Overall, the number of shared haplotypes does not appear to be signiﬁcantly different between any regions, though haplotypes for haplogroup P in Kubalia are only shared with lowland New Guinea and not with other regions or populations. The number of shared haplotypes between IM and New Guinea populations ranges from two to four in both haplogroups. In both haplogroups, there are more derived haplotypes in highland New Guinea from lowland New Guinea (six) than vice versa (three and one; Tables 2 and 3). In haplogroup P, highland New Guinea shows a high number of derived haplotypes from IM (six) and Dreikikir (ﬁve). Lowland New Guinea also has a high number of derived haplotypes from IM (ﬁve) in haplogroup P. The highest number of derived haplotypes in haplogroup Q (nine) is in Fig. 2. Median-joining network of haplogroup P from the Paciﬁc (only HV1). Nodes are patterned as follows: Dreikikir 5 gray; Jama-Sepik 5 black; Kubalia 5 white; lowland New Guinea 5 crossed pattern; highland New Guinea 5 vertical stripes; Island Melanesia 5 horizontal stripes. American Journal of Physical Anthropology 617 mtDNA GENETIC STRUCTURE OF PAPUA NEW GUINEA Fig. 3. Median-joining network of haplogroup Q from the Paciﬁc (only HV1). Coding follows Figure 2. TABLE 2. Shared and derived haplotypes for haplogroup P between the three populations and Lowland New Guinea, Highland New Guinea, and Island Melanesia (IM) Shared/derived1 Dreikikir Jama-Sepik Kubalia Lowland NG Highland NG IM Dreikikir Jama-Sepik Kubalia Lowland NG Highland NG IM – 1/1 0/2 3/2 2/5 3/1 1/0 – 0/2 2/0 2/0 2/0 0/0 0/1 – 2/1 0/2 0/3 3/0 2/1 2/2 – 5/6 4/2 2/0 2/0 0/3 5/4 – 2/1 3/0 2/1 0/3 4/5 2/6 – 1 Regions listed on the rows are derived from the regions listed on the columns. For example, the last column for the ﬁrst row, Dreikikir, 0 is interpreted as number of derived haplotypes from Island Melanesia that are found in Dreikikir. TABLE 3. Shared and derived haplotypes for haplogroup Q between the three populations and regions in the Paciﬁc Shared/derived Dreikikir Jama-Sepik Kubalia Lowland NG Highland NG IM Dreikikir Jama-Sepik Kubalia Lowland NG Highland NG IM – 1/2 3/0 1/5 2/2 2/0 1/0 – 2/0 3/5 1/2 1/1 3/0 2/2 – 2/5 2/3 4/1 1/3 3/2 2/1 – 3/6 4/3 2/0 1/3 2/0 3/1 – 2/0 2/1 1/3 4/1 4/9 2/4 – lowland New Guinea, which is derived from IM haplotypes (Table 3). Tables 4 and 5 show shared and derived haplotypes between populations by language groups. Papuan languages were further categorized into individual phyla, East Papuan, Trans New Guinea (TNG), Sepik-Ramu, and Torricelli. The Torricelli phylum is only represented by Dagua. Haplotypes within haplogroup P in the TNG phylum show a high number of shared haplotypes with the Austronesian language family (ﬁve) as well as derived haplotypes nine from the language family. The number of derived haplotypes from Dreikikir is also high (nine). In haplogroup Q, on average, more haplotypes are shared between the Austronesian language family and other groups, ranging from two to four (Table 5). There is also a higher number of derived haplotypes from the Austronesian language family found in Papuan languages, ranging from ﬁve to nine. Using a mutation rate of 20,180 years for HV1, we calculated the haplogroup ages for P, Q, Q1, and Q3 (Saillard et al., 2000). Ages were also calculated separately by geographical region and language family: IM American Journal of Physical Anthropology 618 E.J. LEE ET AL. TABLE 4. Shared and derived haplotypes for haplogroup P between the three populations and language groups Shared/derived Dreikikir Jama-Sepik Kubalia Torricelli Sepik-Ramu TNG East Papuan Austronesian – 1/1 0/2 0/1 3/0 2/9 1/2 3/1 1/2 – 1/2 0/0 3/0 2/0 1/2 1/1 0/0 0/1 – 1/0 1/1 0/0 0/0 1/2 0/0 0/0 1/0 – 0/0 0/0 0/0 0/0 3/1 3/0 1/3 0/1 – 2/2 1/2 3/0 2/2 2/0 0/2 0/0 2/1 – 2/3 5/3 1/1 1/0 0/0 0/0 1/0 2/0 – 1/2 3/2 1/2 1/2 0/0 3/0 5/9 1/4 – Dreikikir Jama-Sepik Kubalia Torricelli Sepik-Ramu TNG1 East Papuan Austronesian 1 TNG, Trans New Guinea phylum. TABLE 5. Shared and derived haplotypes for haplogroup Q between the three populations and language groups Shared/derived Dreikikir Jama-Sepik Kubalia Torricelli Sepik-Ramu TNG East Papuan Austronesian – 1/2 3/0 1/1 2/3 1/2 1/0 2/0 1/0 – 2/0 2/1 2/4 2/2 1/0 3/0 3/0 2/1 – 0/1 3/5 1/2 1/1 3/0 1/0 2/0 0/0 – 1/1 3/0 2/0 3/0 2/2 2/1 3/1 1/1 – 3/0 1/0 2/0 1/2 2/0 1/1 3/2 3/1 – 2/0 4/0 1/0 1/0 1/0 2/0 1/2 2/0 – 4/1 2/2 3/1 3/2 3/5 2/9 4/6 4/5 – Dreikikir Jama-Sepik Kubalia Torricelli Sepik-Ramu TNG1 East Papuan Austronesian 1 TNG, Trans New Guinea phylum. TABLE 6. Rho estimates1 for haplogroups P and Q (Q1 and Q3) Haplogroup q q (years) r r (years) P IM-AN IM-Papuan PNG highland PNG lowland Papuan PNG lowland AN Q IM-AN IM-Papuan PNG highland PNG lowland Papuan PNG lowland AN Q1 IM-AN IM-Papuan PNG highland PNG lowland Papuan PNG lowland AN Q3 IM-AN IM-Papuan PNG highland PNG lowland Papuan PNG lowland AN 3.0533 2.5313 1.4000 3.4074 3.7647 2.7188 3.6997 2.9750 3.8166 3.1364 3.0783 3.9487 1.6810 1.2031 1.6719 2.2162 1.7500 1.5200 2.3860 1.0455 0.6271 1.6000 1.7222 0.6364 61615 51081 28252 68761 75972 54864 74660 60036 77018 63292 62119 79685 32653 24279 33738 44723 35315 30674 48149 21097 12655 32288 34754 12842 0.80308 0.92333 0.52493 0.914 1.2062 0.84433 1.0546 0.89739 1.2028 0.89014 0.99621 1.2027 0.39662 0.48235 0.67874 0.52407 0.35755 0.5246 0.95618 0.40401 0.37557 0.61644 0.42673 0.35209 16206 18633 10593 18444 24340 17039 21282 18109 24272 17963 20104 24270 8004 9734 13697 10576 7215 10586 19296 8153 7579 12440 8611 7105 1 Estimates are calculated for the entire haplogroup and broken down by geographical/linguistic region. Austronesian (IM-AN), IM Papuan, New Guinea Highland (all Papuan), New Guinea Lowland Papuan, and New Guinea Lowland Austronesian (Table 6). Haplogroup Q is slightly older than P with q values showing 3.6997 for Q while 3.0533 for P. Within haplogroup Q, Q3 appears to be older at 2.386 compared with 1.681 for Q1. For haplogroup P, Papuan-speaking populations from lowland and highland PNG appear to be the oldest. This is similar to that seen in subhaplogroups Q1 and Q3. On the other hand, Austronesian-speaking populations from lowland PNG and Island Melanesian PapuanAmerican Journal of Physical Anthropology speaking populations seem to be as old if not older in haplogroup Q. Gene ﬂow inferred from AMOVA, pairwise comparisons, and two-dimensional analysis Using F-statistics and AMOVA tests (Excofﬁer et al., 1992), gene ﬂow was inferred by grouping populations by geographical region and by language group (Table 7). F-statistics were estimated by AMOVA between New Guinea and the surrounding islands and between highland and lowland New Guinea populations. Furthermore, populations were grouped by Papuan and Austronesian languages, and within Papuan, further divided into East Papuan languages, which are spoken outside of New Guinea, and TNG, Sepik-Ramu, and Torricelli languages, which are spoken within New Guinea. The highest FST value from AMOVA was between Papuan speakers in New Guinea versus Papuan speakers outside of New Guinea, who are classiﬁed in the East Papuan phylum (0.337), indicating relatively high-genetic division. This value is slightly higher than comparing Papuan and Austronesian-speaking populations (0.2894) or between geographical groupings of New Guinea and other islands (0.3116). The lowest genetic division comes from comparisons between populations speaking Papuan languages of different phyla within New Guinea (FST \ 0.09). In addition, FST value between highland and lowland New Guinea is relatively low at 0.1267 also suggesting a low level of population genetic division between populations from the regions. Table 8 shows pairwise population comparisons between the three populations and other populations in the Paciﬁc, which were carried out by calculating the F-statistics from shared haplotype frequencies (Slatkin, 1995). F-statistics with P values below 0.05 are in bold. A complete table with values between each population is available upon request. The three populations examined in our study show similar genetic afﬁnities with each other (0.031–0.052), but other populations appear to mtDNA GENETIC STRUCTURE OF PAPUA NEW GUINEA 619 TABLE 7. AMOVA-derived F-statistics between populations by geography and language Grouping FST FSC FCT NG versus Bismarck and Bougainville Highland versus Lowland NG Papuan versus Austronesian languages NG Papuan versus East Papuan phylum TNG versus Sepik-Ramu and Torricelli phylum Sepik-Ramu versus Torricelli phylum 0.3116 0.1267 0.2894 0.3370 0.0873 0.0073 0.2667 0.1257 0.2566 0.2617 0.0789 0.0143 0.0612 0.0012 0.0441 0.1020 0.0091 -0.0071 TABLE 8. F-statistics values estimated between Dreikikir, Jama-Sepik, Kubalia, and other populations in the Paciﬁc with signiﬁcant values (P < 0.05) in bold Populations Dreikikir FST Jama-Sepik FST Kubalia FST Dreikikir Jama-Sepik Kubalia Markham Fringe Highlands Garaina Kayagar Wahgi-Minj Mandobo Rigo Walis St Martin Dagua Boiken Wingei Warabung Witupe Kiniambu Aita Eivo Nagovisi Rotokas Nasioi Teop Saposa Ata Marabu Rangulit Malasait Sulka Kol Tolai Kuot Mamusi Melamela Mengen Mussau Nailik Nakanai Notsi Tigak Madak Anem Kove Fiji New Caledonia Vanuatu Ontong Java Santa Cruz Solomon Islands – 0.03127 0.0359 0.02933 0.13926 0.10569 0.04245 0.04645 0.08388 0.1148 0.01783 0.05968 0.02493 0.02236 0.00882 0.01289 0.03205 0.01747 0.31962 0.0901 0.19722 0.06877 0.0463 0.06968 0.03038 0.14341 0.16812 0.10712 0.16739 0.06225 0.0864 0.05695 0.16024 0.12778 0.07078 0.05151 0.01803 0.0174 0.15362 0.03957 0.06186 0.04953 0.06778 0.03247 0.03841 0.0306 0.04575 0.14749 0.09874 0.06626 0.03127 – 0.05209 0.04421 0.15173 0.12113 0.05539 0.06277 0.09875 0.12728 0.05059 0.07101 0.02693 0.04472 0.01956 0.03119 0.04224 0.04137 0.3303 0.10269 0.20943 0.08139 0.05859 0.08246 0.04314 0.15512 0.17944 0.11997 0.17903 0.08782 0.09827 0.06983 0.1717 0.14048 0.08409 0.06864 0.03117 0.03 0.16462 0.05277 0.07441 0.06178 0.08063 0.04568 0.05601 0.04096 0.05985 0.15956 0.11043 0.07863 0.0359 0.05209 – 0.04612 0.15137 0.12304 0.07175 0.06411 0.09945 0.12534 0.05311 0.05704 0.03528 0.03697 0.03381 0.02989 0.03983 0.0417 0.31101 0.08992 0.18855 0.07701 0.06078 0.08429 0.04565 0.15441 0.17805 0.12085 0.17718 0.08634 0.09552 0.07151 0.17008 0.14022 0.09298 0.06526 0.03388 0.03263 0.16392 0.05521 0.07523 0.06391 0.08251 0.04823 0.05075 0.04385 0.05657 0.15068 0.10502 0.07407 have closer or similar genetic afﬁnities. For example, Dreikikir shows the lowest F-statistic value with Nailik, an Austronesian-speaking population from New Ireland (0.017), and Kiniambu (0.017) from the Sepik plains. Dreikikir also shows close genetic afﬁnities with Dagua (0.024) and Boiken (0.022), both from the north coast of New Guinea in the same province. Jama-Sepik has the lowest F-statistic value with Wingei (0.019), and Kubalia has the lowest value with Warabung (0.029). Both Wingei and Warabung are from the Prince Alexander Mountains in the same province. The two-dimensional scaling of all populations shows that most are clustered together by geographic proximity (see Fig. 4). All scaling had a ﬁtness level higher than 0.85. Examining populations in New Guinea, the three populations from this study are included in the main central cluster while Rigo, Mandobo, and a population from the Fringe Highlands (see Fig. 5). Rigo is an Austronesian-speaking coastal population, and Mandobo is a Papuan-speaking population from the lowland riverine area. Papuan-speaking populations were analyzed separately in Figure 6, and the three populations from this study are clustered with the main central cluster. Two Bougainville populations appear to be outliers (Aita and Nagovisi) while some populations from New Britain and New Ireland clustered together apart from the main central cluster. Fringe Highland is clustered more closely to this smaller cluster of New Britain and New Ireland populations. Furthermore, populations such as Nasioi (Bougainville), Kol (New Britain), and Sulka (New Britain) are clustered more closely with New Guinea populations. Nucleotide diversity and neutrality We calculated the nucleotide diversity for populations with a sample size above 10 and carried out Tajima’s D neutrality tests (Tajima, 1989; Slatkin, 1994). Nucleotide diversity was slightly higher in populations of New Guinea compared to populations from New Britain and other surrounding islands, which is consistent with the New Guinea populations being older than the island populations (Table 9). All populations except Tigak and New Caledonia are statistically insigniﬁcant for the Tajima’s D test, thus neutrality is assumed (P \ 0.05). DISCUSSION In this study, we examined several questions regarding the genetic patterns in New Guinea: (1) Are populations in New Guinea older than the surrounding outer islands? (2) Can we see signiﬁcant genetic structuring between highland and lowland New Guinea populations? (3) Can we see signiﬁcant genetic structuring based on language families (Austronesian and Papuan) in New Guinea? (4) Can we predict genetic afﬁnity based on geographical proximity and language afﬁnities in Near Oceania? Previous studies have suggested that inhabitants of the highlands in New Guinea are an older population American Journal of Physical Anthropology 620 E.J. LEE ET AL. Fig. 4. Two-dimensional analysis of populations from the Paciﬁc based on mtDNA HV1 F-statistic values from shared haplotype frequencies. Fitness level is 0.864. Fig. 5. Two-dimensional analysis of populations from New Guinea based on mtDNA HV1 F-statistic values from shared haplotype frequencies. Fitness level is 0.966. who had limited interaction with those in the lowlands or outside of the mainland. Our results agree and show star-cluster age estimates of highland inhabitants of New Guinea to be very old as old as the haplogroup in its entirety. In particular, Papuan speakers represented American Journal of Physical Anthropology by subhaplogroup Q1 are oldest in the highlands of New Guinea, and age estimates are younger in the lowlands and outside of New Guinea. On the other hand, Papuan speakers outside of New Guinea are younger than Austronesian speakers in the same region for haplogroup P mtDNA GENETIC STRUCTURE OF PAPUA NEW GUINEA 621 Fig. 6. Two-dimensional analysis of Papuan-speaking populations based on mtDNA HV1 F-statistic values from shared haplotype frequencies. Fitness level is 0.899. and subhaplogroup Q3. This may suggest that Papuan speakers represented by haplogroups P and Q3 were settled in New Guinea for at least several thousands of years before migrating to the outer islands while Q1 can be an example of the older lineages found in IM seen from previous studies (e.g. Friedlaender et al., 2005). Thus, there seems to be diverse genetic patterns of migration among the founding lineages that arrived in New Guinea. Within New Guinea, gene ﬂow estimates show weak structuring of population between the highlands and lowlands. Two-dimensional analysis shows that most New Guinea populations are clustered together and the few populations that appear to be distant from the main cluster include both Austronesian and Papuan-speaking populations (see Fig. 5). This suggests that the genetic differentiation between highland and lowland New Guinea does not seem to be as signiﬁcant as previously thought, in contrast to previous studies (e.g. Stoneking et al., 1990). This may reﬂect more recent population history and interaction between populations regardless of geographic proximity as well as languages. On the other hand, there seems to be greater differentiation between New Guinea and outer islands, rather than differentiation within New Guinea proper. This differentiation is congruent with languages spoken, as our results show that the structuring between Papuan speakers in New Guinea and East Papuan speakers is equivalent to Papuan versus Austronesian speakers. The separate language histories of Austronesian and Papuan languages have been described by others (e.g. Blust, 1995; Pawley, 2002), and the East Papuan phylum has been suggested to include ﬁve or possibly six families and several isolates (Ross, 2005). Thus, there seems to be a broad range of variation within Papuan languages that is reﬂected in the genetic patterns. Genetic afﬁnity is also inferred from shared and derived haplotypes between different geographical regions. Our results show that more highland New Guinea haplotypes are derived from lowland New Guinea and IM than vice versa (Tables 2 and 3). This may suggest that either (1) more highland haplotypes are derived from populations in lowland New Guinea than vice versa, (2) both share a recent common ancestral population, or (3) there has been extensive interaction between these populations, which may have obscured any prehistoric interaction or lack thereof. Although the ancestral population for haplogroups P and Q can be traced back to around 50,000 years ago (Friedlaender et al., 2005), it seems plausible that the relationship between these haplotypes may reﬂect the interaction in the past few thousands of years. This is also supported by the AMOVA tests, which show relatively high-genetic afﬁnity between populations within New Guinea. Examining the number of shared and derived haplotypes according to language groups, overall the number of shared haplotypes between Austronesian and other language groups are equal to or slightly higher than the number of shared haplotypes between Papuan languages. For example, the number of shared haplotypes between Austronesian and Papuan speakers in haplogroup Q ranges from two to four with an average of three haplotypes, while the number of shared haplotypes between the Sepik-Ramu phylum and other Papuan languages is between one and three (Table 5). This may simply indicate that the commonly used dichotomy between Austronesian and Papuan speakers does not American Journal of Physical Anthropology 622 E.J. LEE ET AL. TABLE 9. Nucleotide diversity and neutrality tests for populations in New Guinea and the Paciﬁc1 Population Region Language Sample size Dreikikir* Dagua Boiken Jama-Sepik* Kubalia* Witupe Kiniambu Wingei Warabung Walis St. Martin Markham Rigo Kayagar Fringe Highlands Wahg-Minj Mandobo Ata Marabu Malasait Kol Sulka Tolai Mamusi Melamela Nakanai Nailik Notsi Kuot Tigak Madak Aita Rotokas Nagovisi Nasioi Saposa Teop Santa Cruz Solomon Islands Fiji New Caledonia Vanuatu Ontong Java Torricelli Mt North Coastal NG North Coastal NG Sepik Plains Sepik Plains Sepik Plains Sepik Plains Prince Alexander Mt Prince Alexander Mt Island NG Island NG Coastal NG Coastal NG Southwest Riverine NG Highlands NG Western highlands NG Lowland riverine NG New Britain New Britain New Britain New Britain New Britain New Britain New Britain New Britain New Britain New Ireland New Ireland New Ireland New Ireland New Ireland Bougainville Bougainville Bougainville Bougainville Bougainville Bougainville Torricelli** Torricelli** Sepik-Ramu** Sepik-Ramu** Sepik-Ramu** Sepik-Ramu** Sepik-Ramu** Sepik-Ramu** Sepik-Ramu** Sepik-Ramu** Austronesian Austronesian Austronesian Trans-New Guinea** Trans-New Guinea** Trans-New Guinea** Trans-New Guinea** East Papuan** East Papuan** East Papuan** East Papuan** East Papuan** Austronesian Austronesian Austronesian Austronesian Austronesian Austronesian East Papuan** Austronesian Austronesian East Papuan** East Papuan** East Papuan** East Papuan** Austronesian Austronesian East Papuan** Austronesian Austronesian Austronesian Austronesian Austronesian 28 22 22 30 36 29 32 27 26 32 32 91 18 34 16 19 26 43 75 38 52 26 67 61 14 74 22 12 60 21 33 38 19 14 31 18 16 69 26 15 35 22 24 Nucleotide diversity Tajima’s D 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 0.64244 20.33091 20.46693 0.49036 20.30192 0.32372 0.53893 20.26964 20.51896 0.20899 0.67344 20.59466 0.36325 0.26386 20.20469 20.70137 20.58126 0.05018 0.47456 1.61876 0.23966 20.6116 20.30458 1.89388 20.43928 0.00705 0.84979 0.28172 21.19116 21.48884 20.36874 20.65946 2.74579 0.17874 1.51875 20.57534 20.30872 0.73126 21.15399 20.21909 21.75746 21.03098 1.73489 0.021831 0.028765 0.024372 0.020873 0.021822 0.02202 0.02491 0.021663 0.022073 0.02704 0.028701 0.027869 0.018246 0.025019 0.017338 0.021934 0.016549 0.01128 0.012992 0.00686 0.020513 0.017363 0.025429 0.015762 0.021388 0.016867 0.017004 0.009282 0.006849 0.011571 0.015607 0.013482 0.025068 0.002013 0.025928 0.012699 0.021747 0.026568 0.021346 0.025113 0.020554 0.024565 0.01539 0.011732 0.015295 0.013114 0.011246 0.011596 0.011806 0.013175 0.011666 0.011884 0.014216 0.015027 0.014309 0.010181 0.013203 0.009803 0.012005 0.009160 0.006442 0.007199 0.004276 0.010885 0.009553 0.013196 0.008567 0.011980 0.009065 0.009457 0.005835 0.004220 0.006753 0.008617 0.007537 0.013568 0.001838 0.013688 0.007390 0.012045 0.013740 0.011522 0.013808 0.011014 0.013210 0.008611 1 Papuan languages are indicated by their phylum name. * Populations from this study. ** Phyla classiﬁed in the Papuan language family. exist and that language has not prevented any interaction between these people. A previous study has argued based on autosomal data that languages were not a signiﬁcant barrier to genetic exchange in northern IM (Hunley et al., 2008). Furthermore, the higher F-statistic value based on the AMOVA tests (0.337) between Papuan speakers within New Guinea versus Papuan speakers outside the main island compared with Papuan speakers and Austronesian speakers (0.2894) suggests that geographical separation may have a stronger inﬂuence in the genetic division than language (Table 7). The high F-statistic value between New Guinea and surrounding islands (0.3116) also corroborates this view. Interestingly, TNG populations have a high number of derived haplotypes from the Austronesian language family in both haplogroup P (nine) and haplogroup Q (six) as well as a relatively high number of shared haplotypes (ﬁve and four) (Table 8 and 9). TNG populations also have a high number of derived haplotypes from Dreikikir (nine) in haplogroup P but a low number in haplogroup Q (two). The TNG hypothesis proposes that TNG speakers American Journal of Physical Anthropology expanded from the central highlands of New Guinea at around 6,000 YBP or as early as 10,000 YBP to the east and west, most likely correlating with the expansion of agriculture (Swadling, 1990; Pawley, 1998; Denham et al., 2003; Pawley, 2005). Among the nine derived haplotypes from Dreikikir found in TNG speakers, six are haplotypes from the highlands, which may suggest that TNG speakers possibly descended from an ancestral population to both TNG and Torricelli speakers who are located in the highlands and expanded across New Guinea including the lowlands. Still, interaction between Austronesian speakers and TNG must have also been very extensive based on the number of shared and derived haplotypes. A previous study suggested that the TNG expansion played a more important role in Y-chromosome diversity of New Guinea (Mona et al., 2007). Further investigation of the mtDNA from different TNG and Torricelli speakers across New Guinea may provide further understanding of this proposed expansion. Our analysis included Sepik-Ramu speakers from six populations located in the Sepik plains and the moun- mtDNA GENETIC STRUCTURE OF PAPUA NEW GUINEA tainous regions (Table 9). The rising sea levels during the mid-Holocene formed an inland sea in the SepikRamu basin, which reached its full extent at around 6,500–7,500 YBP and then slowly transformed to the current riverine ﬂoodplain by sedimentation (Swadling et al., 1989; Chappell, 2005). It has been proposed that the change in landscape shifted the focus of interaction from between people of the basin and the highlands to the basin and outer islands (Swadling and Hide, 2005). This interaction may be reﬂected in the high number of derived haplotypes from the Austronesian language family found in haplotypes assigned to the Sepik-Ramu phylum for haplogroup Q (nine). Some previous studies have suggested limited contact between Austronesian-speaking populations and Papuan-speaking populations (e.g. Vilar et al., 2008), but our study proposes that interaction between people between different language families occurred more frequently than previously thought. SUMMARY AND CONCLUSIONS Our study corroborates the common view that populations in New Guinea are in general older than those in the surrounding islands, and Papuan speakers are in general older than Austronesian speakers. Furthermore, we show the existence of genetic structuring between New Guinea and its surrounding islands, as evidenced in comparison with the East Papuan phylum as well as other Papuan languages within New Guinea. Within New Guinea, our data suggest that highlanders are older inhabitants of the region but also had stronger interaction with lowland populations than perhaps previously thought. Although highland and lowland inhabitants speak a range of different languages, there does not seem to be strong genetic structuring between different language groups in New Guinea. In other words, interaction seems to have occurred with little inﬂuence from the languages spoken or the locations of settlement. Although genetic evidence does show remnants of an ‘‘Austronesian expansion’’ along the route to Oceania as seen in previous studies (e.g. Kayser et al., 2008), there does not seem to have been strong barriers to genetic exchange. Furthermore, our conclusion supports the argument of weak or absent language barriers to genetic structuring in Near Oceania, especially when in close contact, which has also been shown in eastern Indonesia (Hunley et al., 2008; Friedlaender et al., 2009; Mona et al., 2009). Future studies examining more populations of different Papuan languages and information from other genetic systems such as the Y-chromosomal and autosomal data in association with ethnographic data may provide information on other aspects of the complexities in New Guinea prehistory since the Holocene. In sum, we believe that the simple dichotomy of the two ‘‘waves’’ oversimpliﬁes the complex prehistories of New Guinea and its surrounding islands and our results emphasize its rich linguistic, cultural, geographical, and genetic diversity. ACKNOWLEDGMENTS We thank all the donors who provided the samples for this study. We are grateful to Jonathan Friedlaender for sample selection from the IMR collection and helpful suggestions on this paper. All procedures were approved by the institutional review boards. The authors thank Miguel Vilar for assisting in the initial analysis. We acknowledge the two anonymous reviewers for their helpful comments and advice on this work. 623 LITERATURE CITED Bandelt HJ, Forster P, Rohl A. 1999. Median-joining networks for inferring intraspeciﬁc phylogenies. Mol Biol Evol 16:37–48. Blust RA. 1985. The Austronesian homeland: a linguistic perspective. Asian Perspect 26:45–67. Blust RA. 1995. The prehistory of Austronesian-speaking peoples: a view from language. J World Prehis 9:453–510. Chappell J. 2000. Pleistocene seedbeds of western Paciﬁc maritime cultures. Mod Quat Res Southeast Asia 16:77–98. Chappell J. 2005. Geographic changes of coastal lowlands in the Papuan past. In: Pawley A, Attenborough R, Golson J, Hide R, editors. Papuan Pasts: cultural, linguistic and biological histories of Papuan-speaking peoples. Canberra, Australia: Paciﬁc Linguistics. p 525–539. Denham TP, Haberle SG, Lentfer C, Fullagar R, Field J, Therin M, Porch N, Winsborough B. 2003. Origins of agriculture at Kuk Swamp in the highlands of New Guinea. Science 301: 189–193. Easteal S, Whittle B, Mettenmeyer A, Attenborough R, Bhatia K, Alpers M. 2005. Mitochondrial genome diversity among Papuan-speaking people of Papua New Guinea. In: Pawley A, Attenborough R, Golson J, Hide R, editors. Papuan Pasts: cultural, linguistic, and biological histories of Papuan-speaking peoples. Canberra, Australia: Paciﬁc Linguistics, Research School of Paciﬁc and Asian Studies. p 717–728. Excofﬁer L, Smouse PE, Quattro JM. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131:479–491. Friedlaender J, Hunley K, Dunn M, Terrill A, Lindstrom E, Reesink G, Friedlaender F. 2009. Linguistics more robust than genetics. Science 324:464–465. Friedlaender J, Schurr T, Gentz F, Koki G, Friedlaender F, Horvat G, Babb P, Cerchio S, Kaestle F, Schanﬁeld M, et al. 2005. Expanding Southwest Paciﬁc mitochondrial haplogroups P and Q. Mol Biol Evol 22:1506–1517. Friedlaender JS, Friedlaender FR, Hodgson JA, Stoltz M, Koki G, Horvat G, Zhadanov S, Schurr TG, Merriwether DA. 2007. Melanesian mtDNA complexity. PLoS ONE 2:e248. Friedlaender JS, Friedlaender FR, Reed FA, Kidd KK, Kidd JR, Chambers GK, Lea RA, Loo JH, Koki G, Hodgson JA, et al. 2008. The genetic structure of Paciﬁc Islanders. PLoS Genet 4: e19. Gillieson D, Mountain M-J. 1983. Environmental history of Nombe rockshelter, Papua New Guinea. Archaeol Oceania 18:53–62. Gray RD, Drummond AJ, Greenhill SJ. 2009. Language phylogenies reveal expansion pulses and pauses in Paciﬁc settlement. Science 323:479–483. Green RC. 1991. The Lapita Cultural Complex: current evidence and proposed models. In: Bellwood P, editor. Indo-Paciﬁc Prehistory 1990. Canberra, Australia: Australian National University. p 295–305. Groube L, Chappell J, Muke J, Price D. 1986. A 40,000 year-old human occupation site at Huon Peninsula, Papua New Guinea. Nature 324:453–455. Hunley K, Dunn M, Lindstrom E, Reesink G, Terrill A, Healy ME, Koki G, Friedlaender FR, Friedlaender JS. 2008. Genetic and linguistic coevolution in Northern Island Melanesia. PLoS Genet 4:e1000239. Ingman M, Gyllensten U. 2003. Mitochondrial genome variation and evolutionary history of Australian and New Guinean aborigines. Genome Res 13:1600–1606. Jeanmougin F, Thompson JD, Gouy M, Higgins DG, Gibson TJ. 1998. Multiple sequence alignment with Clustal X. Trends Biochem Sci 23:403–405. Kayser M, Choi Y, van Oven M, Mona S, Brauer S, Trent RJ, Suarkia D, Schiefenhovel W, Stoneking M. 2008. The impact of the Austronesian expansion: evidence from mtDNA and Y chromosome diversity in the Admiralty Islands of Melanesia. Mol Biol Evol 25:1362–1374. Maddison D, Maddison W. 2000. MacClade version 4: analysis of phylogeny and character evolution, 4th ed. Sunderland, MA: Sinauer Associates. American Journal of Physical Anthropology 624 E.J. LEE ET AL. Merriwether DA, Hodgson JA, Friedlaender FR, Allaby R, Cerchio S, Koki G, Friedlaender JS. 2005. Ancient mitochondrial M haplogroups identiﬁed in the Southwest Paciﬁc. Proc Natl Acad Sci USA 102:13034–13039. Mona S, Grunz KE, Brauer S, Pakendorf B, Castri L, Sudoyo H, Marzuki S, Barnes RH, Schmidtke J, Stoneking M, et al. 2009. Genetic admixture history of eastern Indonesia as revealed by Y-chromosome and mitochondrial DNA analysis. Mol Biol Evol 26:1865–1877. Mona S, Tommaseo-Ponzetta M, Brauer S, Sudoyo H, Marzuki S, Kayser M. 2007. Patterns of Y-chromosome diversity intersect with the Trans-New Guinea hypothesis. Mol Biol Evol 24: 2546–2555. Mountain M-J. 1991. Bulmer Phase I: environmental change and human activity through the late Pleistocene into the Holocene in the highlands of New Guinea: a scenario. In: Pawley A, editor. Man and a half: essays in Paciﬁc anthropology and ethnobiology in honour of Ralph Bulmer. Auckland: Polynesian Society. p 510–520. O’Connell J, Allen J. 2004. Dating the colonization of Sahul (Pleistocene Australia-New Guinea): a review of recent research. J Archaeol Sci 31:835–853. Ohashi J, Naka I, Tokunaga K, Inaoka T, Ataka Y, Nakazawa M, Matsumura Y, Ohtsuka R. 2006. Brief communication: mitochondrial DNA variation suggests extensive gene ﬂow from Polynesian ancestors to indigenous Melanesians in the northwestern Bismarck Archipelago. Am J Phys Anthropol 130: 551–556. Pawley A. 1998. The Trans New Guinea hypothesis: a reassessment. In: Miedema J, Ode C, Dam RAC, editors. Perspectives on the Bird’s Head of Irian Jaya, Indonesia. Amsterdam: Editions Rodopi. p 655–689. Pawley A. 2002. The Austronesian dispersal: languages, technologies, and people. In: Bellwood P, Renfrew C, editors. Examining the farming/language dispersal hypothesis. Cambridge, UK: McDonald Institute for Archaeological Research. p 251–274. Pawley A. 2005. The chequered career of the Trans New Guinea hypothesis: recent research and its implications. In: Pawley A, Attenborough R, Golson J, Hide R, editors. Papuan Pasts: cultural, linguistic and biological histories of Papuan-speaking peoples. Canberra, Australia: Paciﬁc Linguistics. p 67–107. Pawley A. 2007. Recent research on the historical relationships of the Papuan languages, or, what does linguistics say about the prehistory of Melanesia? In: Friedlaender JS, editor. Genes, language, and culture history in the southwest Paciﬁc. New York: Oxford University Press. p 10–35. Ross M. 2005. Pronouns as a preliminary diagnostic for grouping Papuan languages. In: Pawley A, Attenborough R, Golson J, Hide R, editors. Papuan Pasts: cultural, linguistic and biological histories of Papuan-speaking peoples. Canberra, Australia: Paciﬁc Linguistics. p 15–66. American Journal of Physical Anthropology Saillard J, Forster P, Lynnerup N, Bandelt HJ, Norby S. 2000. mtDNA variation among Greenland Eskimos: the edge of the Beringian expansion. Am J Hum Genet 67:718–726. Scheinfeldt L, Friedlaender F, Friedlaender J, Latham K, Koki G, Karafet T, Hammer M, Lorenz J. 2006. Unexpected NRY chromosome variation in Northern Island Melanesia. Mol Biol Evol 23:1628–1641. Schneider S, Roessli D, Excofﬁer L. 2000. Arlequin: a software for population genetics data analysis. 2.00 ed: Genetics and Biometry Lab, Department of Anthropology, University of Geneva. Slatkin M. 1994. An exact test for neutrality based on the Ewens sampling distribution. Genet Res 64:71–74. Slatkin M. 1995. A measure of population subdivision based on microsatellite allele frequencies. Genetics 139:457–462. Stoneking M, Jorde LB, Bhatia K, Wilson AC. 1990. Geographic variation in human mitochondrial DNA from Papua New Guinea. Genetics 124:717–733. Summerhayes G. 2007. Island Melanesian pasts: a view from archaeology. In: Friedlaender J, editor. Genes, languages, and culture history in the southwest Paciﬁc. New York: Oxford University Press. p 10–35. Summerhayes G, Bird R, Fullagar R, Gosden C, Specht J, Torrence R. 1998. Application of PIXE-PIGME to archaeological analysis of changing patterns of obsidian use in West New Britain, Papua New Guinea. In: Shackley S, editor. Advances in archaeological volcanic glass studies. New York: Plenum Press. p 129–158. Swadling P. 1990. Sepik Prehistory. In: Lutkehaus N, Kaufmann C, Mitchell WE, Newton D, Osmundsen L, Schuster M, editors. Sepik Heritage: tradition and change in Papua New Guinea. Bathurst, NSW Australia: Crawford House Press. p 71–86. Swadling P, Chappell J, Francis G, Araho N, Ivuyo B. 1989. A late Quaternary inland sea and early pottery in Papua New Guinea. Archaeol Oceania 24:106–109. Swadling P, Hide R. 2005. Changing landscapes and social interaction: looking at agricultural history from a SepikRamu perspective. In: Pawley A, Attenborough R, Golson J, Hide R, editors. Papuan Pasts: cultural, linguistic and biological histories of Papuan-speaking peoples. Canberra, Australia: Paciﬁc Linguistics. p 289–328. Tajima F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123:585–595. Vilar MG, Kaneko A, Hombhanje FW, Tsukahara T, Hwaihwanje I, Lum JK. 2008. Reconstructing the origin of the Lapita Cultural Complex: mtDNA analyses of East Sepik Province. PNG. J Hum Genet 53:698–708. White J. 1972. Ol Tumbuna: archaeological excavations in the eastern Central highlands, Papua New Guinea. Canberra: Australian National University. White J, Crook K, Ruxton B. 1970. Kosipe: a Pleistocene site in the Papua Highlands. Proc Prehis Soc 36:152–170.