• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Sep 13, 2005; 102(37): 13034–13039.
Published online Sep 6, 2005. doi:  10.1073/pnas.0506195102
PMCID: PMC1201611

Ancient mitochondrial M haplogroups identified in the Southwest Pacific


Based on whole mtDNA sequencing of 14 samples from Northern Island Melanesia, we characterize three formerly unresolved branches of macrohaplogroup M that we call haplogroups M27, M28, and M29. Our 1,399 mtDNA control region sequences and a literature search indicate these haplogroups have extremely limited geographical distributions. Their coding region variation suggests diversification times older than the estimated date for the initial settlement of Northern Island Melanesia. This finding indicates that they were among the earliest mtDNA variants to appear in these islands or in the ancient continent of Sahul. These haplogroups from Northern Island Melanesia extend the existing schema for macrohaplogroup M, with many independent branches distributed across Asia, East Africa, Australia, and Near Oceania.

Keywords: Island Melanesia, phylogeography, paleodemography, molecular evolution, Papua New Guinea

Portions of Northern Island Melanesia were settled by at least 42,000 years before present (YBP) (1, 2), apparently not long after New Guinea, which was joined at that time to Australia as the ancient Pleistocene continent of Sahul. Non-14C dates from Australia, if they are reliable, may extend the date for first human settlement in Sahul for an additional ≈10,000 years (3, 4). New Guinea and adjacent Northern Island Melanesia constitute the region of Near Oceania, as shown in Fig. 1, and the populations there were at the eastern edge of the human species range until 3,200 YBP.

Fig. 1.
Frequency distributions of mtDNA haplogroups M27, M28, and M29, taken from our series (Table 1). Numbers within each pie are values of N. “Other” haplogroups are P, Q, B, and E, as reported in Table 1 for this series (and discussed in ...

The earliest populations of Northern Island Melanesia were small groups of hunter-gatherers. They were not “strandlopers” restricted to the coasts and lagoons, but settled the interiors of the large islands intermittently (2, 5). The intentional introductions of plants and animals from New Guinea over the following millennia indicate continuing outside contacts at a modest level. Short voyages between islands are inferred (2, 6) because people had successfully made the longer windward crossing to Bougainville from New Ireland by 29,000 YBP, and after 20,000 YBP there was a detectable and repeated trickle of New Britain obsidian to New Ireland and Nissan up to ≈7,000 YBP. The implication is that isolation of these small island populations was an incomplete but persistent condition across the region for tens of thousands of years during the Pleistocene. By extension, movements between Near Oceania and Island Southeast Asia also would have been intermittent and small in scale.

Only ≈3,200 YBP did people from Island Southeast Asia, with advanced sailing and agricultural skills, make a major impact on this region (7, 8). This led, in turn, to the rapid colonization of the formerly uninhabited islands of Remote Oceania as far as Western Polynesia by the “Lapita People” and also to a following period with variable contacts among those populations and coastal groups back in Near Oceania. As a result of this history, Near Oceanic populations are linguistically and biologically extraordinarily diverse (911).

mtDNA continues to be a particularly powerful source for reconstructions of early human demographics because of the effective absence of recombination and its comparatively high mutation rate. In the Southwest Pacific, the haplogroups P and Q have been recognized as old and specific to the general region (1215). They have long branches, and analyses of their control region variation indicate population expansions in the Pleistocene (12, 13). P is more widespread and heterogeneous than Q, and may therefore be the older of the two. With one clear exception, different branches of P occur either in Australia or New Guinea, but not both. Q is absent in Australia but very common in New Guinea and Island Melanesia (12). This pattern suggests substantial isolation of Australia from New Guinea/Island Melanesia since around the time of first settlement. Haplogroup B4a, which is very common in Remote Oceania and coastal Near Oceania but absent in aboriginal Australia, the New Guinea highlands, and interior sections of Bougainville and New Britain, was introduced to Near Oceania in the terminal Pleistocene (1618).

Besides these common mtDNA haplogroups, a number of other mtDNA variants in Island Melanesia have not been possible to characterize without full mtDNA genome sequencing (17). We have now sequenced 14 whole mtDNA genomes of these and can define three additional branches of macrohaplogroup M. They have not been found anywhere to the west of Island Melanesia, including New Guinea. These haplogroups (M27, M28, and M29) are not closely related to each other, although M29 might be distantly related to Q.

Materials and Methods

The samples analyzed were selected from our Southwest Pacific collection. Its core consisted of blood samples collected in three recent field seasons in the Bismarck Archipelago. This primary set was augmented with plasmas and urines from older collections, described elsewhere (12). Information on survey subjects included their language, a short genealogy, current residence, and birthplace (used to assign location), although such details were not available from some of the other collections. One sample from each identified matriline was included in the initial mtDNA control region analysis. The primary samples were collected, and all selected samples were analyzed, with informed consent protocols approved by the appropriate Human Subjects Ethical Committees of Papua New Guinea, the University of Michigan, Binghamton University, and Temple University.

The mtDNA analysis of the selected samples occurred in three phases: (i) sequencing of hypervariable segments 1 and 2 (HVS1 and HVS2); (ii) for those samples not definitely assigned to a known haplogroup on this basis, restriction frangment-length polymorphism (RFLP) screening for the two mutations defining macrohaplogroup M (DdeI 10394, AluI 10397), and depending on the presence or absence of these, additional RFLPs known to identify other haplogroups in the Southwest Pacific (1921); and (iii) sequencing of the coding region on 14 representative samples from each of the major haplogroups that could not be assigned to currently published M sublineages.

DNA was extracted from 100–200 μl of buffy coat, plasma, or urine (depending on the source of the sample) by using either the guanidine-silica based IsoQuick extraction kit (Orca Scientific, Bothell, WA) or the column-based Qiagen extraction kit (Qiagen, Valencia, CA). In preparation for sequencing, DNA was PCR amplified following standard protocols, employing Platinum Taq Polymerase (Invitrogen). The control region was amplified by using primers spanning nucleotide positions 15938–00429. The coding region was amplified by using the PCR primers and conditions of Rieder et al. (22). Successful amplification was verified by electrophoresis on 1% ethidium bromide-stained agarose gels. Samples were prepared for sequencing by an ExoI digest followed by filtration through a Millipore 96-well filter plate (Millipore, Billerica MA) to remove single-stranded DNA and unincorporated nucleotides. PCR product was sequenced by using various versions of the BigDye Terminator Sequencing kits from Applied Biosystems on an Applied Biosystems 377XL automated sequencer using described conditions (23). Custom designed internal sequencing primers were used for all large PCR fragments to increase double-fold coverage. Contig assemblage and sequence alignment was accomplished with sequencher (Forensic Version, GeneCodes, Ann Arbor, MI). The phylogenetic tree was inferred from median-joining networks rooted to L3. The tree was hand-checked to resolve several homoplasies. A few ambiguities remained, and we tended to be conservative in interpreting those cases.


The control region mutations that initially identified the M variants are listed in Table 1 with key mutations in bold. The mtDNA haplogroup incidences in our 1,399 control region results are presented in Table 2 by island and the frequencies of the three haplogroups are indicated in Fig. 1. Besides the commonly occurring haplogroups P, Q, and B described elsewhere (12), the newly defined branches of M were found in certain circumscribed regions in intermediate to high frequencies. The phylogenetic M tree for the old Near Oceanic lineages is presented in Fig. 2 with Australian Aborigine M sequences included for comparison.

Fig. 2.
Phylogenetic relationships of ancient Near Oceanic and Australian Aborigine M lineages. Control region mutations are in bold, and mutations that recur in the phylogeny are underlined. The poly(C) regions in HVS1 and -2 as well as 16519 are excluded. Boxes ...
Table 1.
Defining control region mutations for M27, M28, and M29
Table 2.
mtDNA haplogroup incidences from this study

Haplogroup M27, as shown in Table 1 and Fig. 2, had three branches with no key shared control region mutations, but they were ultimately linked in the coding region sequencing by the shared transitions at nucleotide positions 5375, 9201, and 12358. We found no convincing associations of M27 with other M branches, although 5177 was also found in Q3 and in M21b from the aboriginal Semang of Malaysia (24). Position 5585 in M28a has been found in Indian M6b, but this also appears to be a recurrence. M27 was most frequent in central Bougainville, especially branch M27a, with the other M27 branches detected in Bougainville and sporadically in New Britain, New Ireland, the Solomon Islands, Santa Cruz (17), and Vanuatu (25).

Haplogroup M28 was relatively common. It had 71 constituent haplotypes that were identified in 261 individuals in our series. Its defining control region variants were the transitions at nucleotide positions 16468 and 16148, and each of its major subbranches had other control region distinctions, as shown in Table 1 and Fig. 2. Coding region sequencing on a representative subset reinforced these relationships. Within M, haplogroup M28 shared the transition at nucleotide position 16362 with haplogroups D, G, and M9/E, so we interpreted this as a recurrence.

M28 had its greatest prevalence and diversity in the interior of East New Britain, where it was associated with an ancient population expansion (17). In some sample sets, it had a frequency of 100%. M28 was rare in our New Ireland sample set and almost absent in Bougainville, a distribution contrasting with M27. We found M28 at variable frequencies elsewhere in Island Melanesia (i.e., Santa Cruz, Vanuatu, New Caledonia, and Fiji), reported in studies that relied on short sequencing within HVS1 with different terminologies (refs. 2527, and possibly ref. 28). Because the most characteristic single mutation of M28 (i.e., 16468) lies outside the region of HVS1 most commonly sequenced in early mtDNA sequence analyses, it is possible its distribution is still underestimated. One transversion in a branch of M28b is 16318T, which is a defining mutation for Indian M18. However, we are assuming it is recurrent as there are no other shared mutations.

Haplogroup M29 was initially distinguished by the transition at nucleotide position 211, subsequently complemented by a long series of coding region transitions (an average of 13). It was relatively rare in our series, but, as with M28, its distribution beyond our series could have been underestimated because of the general early reliance on short HVS1 sequencing. In our series, haplogroup M29 was found most frequently in East New Britain and especially among the Tolai (who originally resided in southern New Ireland). It was also identified in single samples from the Solomon Islands and Vanuatu. The M29 sequences share the 13500 transition with Q, which could indicate a remote connection. However, it may also be a recurrence, because it has also been found in a variety of other haplogroups (D, M7a, R2, F, U7, and V). Additional M29 sequencing should help resolve this issue.

Haplogroup E, an Asian M variant with a more recent coalescence (16), was also found in our series, most commonly in West New Britain (Table 1).


Northern Island Melanesia is clearly a relict area, retaining a remarkable number of ancient population genetic signatures. Four old lineages of macrohaplogroup M have now been identified in Near Oceania (Fig. 2). Besides haplogroup Q, which had been shown to have two clearly “star-like” subdivisions (12), the three M haplogroups described here are centered in different locations within Northern Island Melanesia: M27 is most common and diverse in Bougainville, M28 in the interior of east New Britain, and M29 in southern New Ireland and east New Britain. The remarkably limited and different distributions of each of these M branches are the likely result of restricted marital migration rates (refs. 29 and 30, pp. 71–75) within the region from the times of their appearances, possibly coupled with their separate introductions.

Fig. 2 includes the Australian M42 haplogroup (T. Kivisild, personal communication). The Australian and Near Oceanic M branches share no special relationships (1598 is clearly recurrent) and do not overlap in their distributions. This finding only reinforces the remarkable distinctiveness of the different Southwest Pacific populations.

The old Northern Island Melanesian M haplogroups have distributions that tie their original associations to different Papuan speaking populations in the region (17). However, links to specific languages or language families have decayed somewhat over time because of language shift and intermarriage so that correlations of gene and language distributions will inevitably be weak.

On present evidence, the spread of M27, M28, and M29 beyond their Northern Island Melanesian centers was limited to the adjacent island chains in Southern Island Melanesia (also reported in refs. 17, 25, and 26). A selection of particular P and Q haplotypes are also scattered across this region from New Guinea to Fiji, specifically Q2*, Q2b, P1, P1e, Plf, and P2 (see tables s2 and s3 of ref. 12). Because none of these are restricted to Southern Island Melanesia, they indicate population interactions right across the region. They must have all spread to Southern Island Melanesia after its colonization by the Lapita People, who would already have been predominantly haplogroup B4a, like all their direct descendants across the rest of Remote Oceania. A more complex scenario would have all of these haplotypes carried with Lapita groups as far as Fiji, and then have all lost in separate founder events in every subsequent voyage into Remote Oceania, a most unlikely alternative. Also, archaeological evidence does indicate considerable influence from Near Oceania into Southern Island Melanesia after the Lapita period, compatible with the preferred first scenario (7, 3135).

Table 3 gives lineage diversity (ρ and σ) and age estimates for old M branches from the Southwest Pacific employing the commonly used technique of Saillard et al. (36), in conjunction with the estimated mutation calibration rate of Mishmar et al. (37), because these have been recently applied, with an adjustment, to a number of mtDNA sequences from other relict populations in Southeast Asia, including Andaman Islanders and aboriginal Malays (24, 38). The overall (average) age of M had been estimated with the Saillard and Mishmar approach at 64,800 years with a standard error of 7,100 years, taken to conform to proposed dates for a “Southern Route” expansion between 55,000 and 85,000 YBP (39). With our data, the combined estimate for Near Oceanic Ms becomes 72,100 ± 8,000, with the estimate for the M27 branch, taken alone, as 84,400 ± 14,300 years (Table 3). Our calculation for the Australian Aboriginal M42 haplogroup age estimate is in the same range, as shown. However, questions on this approach include the lack of true independence of the branch lengths because they are time-constrained (also commented on in ref. 24), and because of uncertainties on the proper rate calibration (40). A final issue is that the coalescence times of these and other molecular estimates under 1 million years may contain a significant rate artifact (41, 42), which, if verified, would make a variety of absolute date estimates considerably younger.

Table 3.
Coding region diversity and age estimates for M haplogroups in Near Oceania and Australia

The length of the branches between the ancestral haplotype M and the M27, M28 and M29 branches vary considerably (see Fig. 2 and Table 3), and this can be interpreted in different ways. M27 is not so much longer than the others as to violate a rather insensitive test of mutational clock assumptions (comparing two log likelihood scores with paml and the HKY85 model, as in ref. 24), but the mutation accumulation within the M27 clade still follows a distribution that is significantly higher than that of the other Near Oceanic M haplotypes at the 1% level using the Mann–Whitney test. If this difference is accepted as real, it could be explained in two ways. First, all M lineages arose essentially simultaneously from ancestral M (as argued in ref. 24), and the variance in mutations subsequently accumulated is inconsequential. An alternative explanation for the difference would be that the ancestral haplotype M existed for a minimum interval of ≈20,000 years, during which the Near Oceanic lineages branched off sequentially, thereby accounting for the different numbers of mutations accumulated to the present. This scenario would imply that the ancestral M haplotype was close to fixation in the ancestral population, perhaps reflecting an initial expansion into unpopulated territory by a small group. Because the time of allele loss from fixation in nuclear genes can be approximated by 4Ne generations (43), the comparable value for mtDNA can be approximated to 2Nf. Assuming a female generation time of 20 years, this approximation yields a long-term effective population size of 483 ancestral M women for Near Oceania, or even less if female generation times were >20 years. The resulting population estimate would be a minimum value, and an effective population of ≈1,000 M women may be more appropriate to account for the regional M branch origins occurring while the ancestral haplotype M was at high frequency. As a larger set of Southwest Pacific whole mtDNA genome sequences develops, we should be able to test these alternatives with more certainty.

Although the unusual set of Southwest Pacific mtDNA variants does tend to set this region apart from Eurasia as an apparent unit, it is its internal and ancient diversity that is its true defining characteristic, maintained by long-term isolation across the entire region. Their very circumscribed geographical distributions suggest that the old M lineages appeared with the colonization of Near Oceania and Sahul, and their lineage age estimates, if accurate, intimate they first diverged from ancestral M before their introduction here. Their heterogeneous distributions may also suggest they were introduced not only at different times but by different groups. Where they diverged from ancestral M remains problematic, in part because of poor sampling throughout the adjacent regions of Island Southeast Asia. Their highly structured distribution clearly contradicts some earlier notions of a loosely unified Melanesian, Old Melanesian, Australoid, or Australo-Melanesian population with a common ancestry (4446), in favor of a far more complex population history (47, 48).


We thank Heather Norton, Dan Hrdy, Charles Mgone, and the people of Bougainville, New Ireland, New Hanover, and New Britain who participated in the study, Theodore Schurr, John McDonough, Stacy McGrath, Fred Gentz, and our other collaborators at the Papua New Guinea Institute for Medical Research. We also thank Ger Reesink, Eva Lindström, and Gisele Horvat for advice, and particularly Roger Green. Toomas Kivisild has kindly permitted us to include the M42 Australian Aboriginal sequences, and has provided a number of constructive criticisms. The research was supported by grants from the National Geographic Society Exploration Fund, the Wenner-Gren Foundation for Anthropological Research, and National Science Foundation Grants SBR-9796054 and SBR-9601020.


Author contributions: D.A.M. and J.S.F. designed research; J.A.H., S.C., and G.K. performed research; F.R.F. and R.A. analyzed data; and J.S.F. wrote the paper.

Abbreviation: YBP, years before present.

Data deposition: The sequences reported in this paper have been deposited in the GenBank database (accession nos. DQ137398–DQ137411).


1. Leavesley, M. & Chappell, J. (2004) Antiquity Proj. Gallery 78, 301.
2. Allen, J. (2003) in Pacific Archaeology: Assessments and Anniversary of the First Lapita Excavation (July 1952) Koné, Nouméa, 2002, ed. Sand, C. (Le Cahiers de l'Archéologie en Nouvelle-Calédonie, Nouméa, New Caledonia), Vol. 15, pp. 33–42.
3. O'Connell, J. F. & Allen, J. (2004) J. Archaeol. Sci. 31, 835–853.
4. Bowler, J. M., Johnston, H., Olley, J., Prescott, J., Roberts, R. G., Shawcross, W. & Spooner, N. A. (2003) Nature 421, 837–840. [PubMed]
5. Pavlides, C. & Gosden, C. (1994) Antiquity 69, 604–610.
6. Anderson, A. (2000) in East of Wallace's Line: Studies of Past and Present Maritime Cultures of the Indo-Pacific Region. Modern Quaternary Research in Southeast Asia, eds. O'Connor, S. & Veith, P. (A. A. Balkema, Rotterdam), Vol. 16, pp. 13–50.
7. Spriggs, M. J. T. (1997) The Island Melanesians (Blackwell Scientific, Cambridge, MA).
8. Bellwood, P. (1996) in The Origin and Spread of Agriculture and Pastoralism in Eurasia, ed. Harris, D. R. (Smithsonian Institution Press, Washington, DC), pp. 465–498.
9. Friedlaender, J. S. (1987) in The Solomon Islands Project: A Long Term Study of Human Biology and Culture Change, ed. Friedlaender, J. S. (Oxford Science, Oxford), pp. 351–364.
10. Dunn, M., Terrill, A. & Reesink, G. (2002) Oceanic Linguistics 41, 28–62.
11. Green, R. C. (1991) in Man and a Half: Essays in Pacific Anthropology and Ethnobiology in Honour of Ralph Bulmer, ed. Pawley, A. (The Polynesian Society, Auckland), pp. 491–502.
12. Friedlaender, J. S., Schurr, T. S., Gentz, F., Koki, G., Friedlaender, F., Horvat, G., Babb, P., Cerchio, S., Kaestle, F., Schanfield, et al. (2005) Mol. Biol. Evol. 22, 1506–1517. [PubMed]
13. Forster, P., Torroni, A., Renfrew, C. & Rohl, A. (2001) Mol. Biol. Evol. 18, 1864–1881. [PubMed]
14. Stoneking, M., Jorde, L. B., Bhatia, K. & Wilson, A. C. (1990) Genetics 124, 717–733. [PMC free article] [PubMed]
15. Redd, A. J. & Stoneking, M. (1999) Am. J. Hum. Genet. 65, 808–828. [PMC free article] [PubMed]
16. Trejaut, J. A., Kivisild, T., Loo, J. H., Lee, C. L., He, C. L., Hsu, C. J., Li, Z. Y. & Lin, M. (2005) PLoS Biol. 3, e247. [PMC free article] [PubMed]
17. Friedlaender, J. S., Gentz, F., Thompson, F., Kaestle, F., Schurr, T. G., Koki, G., Mgone, C., McDonough, J., Smith, L. & Merriwether, D. A. (in press) in Papuan Pasts: Studies in the Cultural, Linguistic and Biological History of the Papuan Speaking Peoples, eds. Pawley, A., Attenborough, R., Golson, J. & Hyde, R. (Pacific Linguistics, Canberra, Australia).
18. Merriwether, D. A., Friedlaender, J. S., Mediavilla, J., Mgone, C., Gentz, F. & Ferrell, R. E. (1999) Am. J. Phys. Anthropol. 110, 243–270. [PubMed]
19. Ballinger, S. W., Schurr, T. G., Torroni, A., Gan, Y. Y., Hodge, J. A., Hassan, K., Chen, K.-H. & Wallace, D. C. (1992) Genetics 130, 139–152. [PMC free article] [PubMed]
20. Stoneking, M. (1990) J. Mol. Evol. 31, 343–346. [PubMed]
21. Huoponen, K., Schurr, T. G., Chen, Y. & Wallace, D. C. (2001) Hum. Immunol. 62, 954–969. [PubMed]
22. Rieder, M. J., Taylor, S. L., Tobe, V. O. & Nickerson, D. A. (1998) Nucleic Acids Res. 26, 967–973. [PMC free article] [PubMed]
23. Merriwether, D. A., Kaestle, F. A., Zemel, B., Koki, G., Mgone, C., Alpers, M. P. & Friedlaender, J. S. (1999), in Genomic Diversity: Applications in Human Population Genetics, eds. Surinder S., Papiha, R. D. & Chakraborty R. (Kluwer Academic/Plenum, New York), pp. 153–181.
24. Macaulay, V., Hill, C., Achilli, A., Rengo, C., Clarke, D., Meehan, W., Blackburn, J., Semino, O., Scozzari, R., Cruciani, F., et al. (2005) Science 308, 1034–1036. [PubMed]
25. Cox, M. P. (2003) Ph.D. dissertation (University of Otago, Dunedin, New Zealand).
26. Friedlaender, J. S., Gentz, F., Green, K. & Merriwether, D. A. (2002) Hum. Biol. 74, 453–471. [PubMed]
27. Sykes, B., Lieboff, A., Low-Beeer, J., Tetzner, S. & Richards, M. (1995) Am. J. Hum. Genet. 57, 1463–1475. [PMC free article] [PubMed]
28. Lum, J. K. & Cann, R. L. (2000) Am. J. Phys. Anthropol. 113, 151–168. [PubMed]
29. Long, J. C. (1991) Genetics 127, 417–428. [PMC free article] [PubMed]
30. Friedlaender, J. S. (1975) Patterns of Human Variation: The Demography, Genetics, and Phenetics of Bougainville Islanders (Harvard Univ. Press, Cambridge, MA).
31. Green, R. C. (1997) N. Z. J. Archaeol. 17, 5–27.
32. Bedford, S., Sand, C. & Burley, D. (2002) in New Zealand Archaeological Association Monograph (New Zealand Archaeological Association, Auckland), Vol. 25.
33. Galipaud, J.-C. & Lilley, J. (1999) The Pacific from 5000 to 2000 B.P. Colonisation and Transformation (IRD Éditions, Paris).
34. Clark, G. R., Anderson, A. & Sorovi-Vunidilo, T. (2001) Terra Australis (Pandanus, Canberra, Australia), Vol. 17.
35. Bedford, S. (2000) Bull. Indo-Pacific Prehistory Assoc. 19, 159–166.
36. Saillard, J., Forster, P., Lynnerup, N., Bandelt, H.-J. & Norby, S. (2000) Am. J. Hum. Genet. 67, 718–726. [PMC free article] [PubMed]
37. Mishmar, D., Ruiz-Pesini, E., Golik, P., Macaulay, V., Clark, A. G., Hosseini, S., Brandon, M., Easley, K., Chen, E., Brown, M. D., et al. (2003) Proc. Natl. Acad. Sci. USA 100, 171–176. [PMC free article] [PubMed]
38. Thangaraj, K., Chaubey, G., Kivisild, T., Reddy, A. G., Singh, V. K., Rasalkar, A. A. & Singh, L. (2005) Science 308, 996. [PubMed]
39. Forster, P. & Matsumura, S. (2005) Science 308, 965–966. [PubMed]
40. Ingman, M., Kaessmann, H., Paabo, S. & Gyllensten, U. (2000) Nature 408, 708–713. [PubMed]
41. Ho, S. Y., Phillips, M. J., Cooper, A. & Drummond, A. J. (2005) Mol. Biol. Evol. 22, 1561–1568. [PubMed]
42. Penny, D. (2005) Nature 436, 183–184. [PubMed]
43. Li, W. H. & Nei, M. (1977) Genetics 86, 901–914. [PMC free article] [PubMed]
44. Howells, W. W. (1973) The Pacific Islanders (Scribner's, New York).
45. Bellwood, P. (1978) Man's Conquest of the Pacific (Collins, Auckland).
46. Pietrusewsky, M. (1994) Oceanic Linguistics 332, 407–429.
47. Terrell, J., Kelly, K. M. & Rainbird, P. (2001) Curr. Anthropol. 42, 97–124.
48. Green, R. C. (2003) in Pacific Archaeology: Assessments and Anniversary of the First Lapita Excavation (July 1952) Koné, Nouméa, 2002, ed. Sand, C. (Le Cahiers de l'Archéologie en Nouvelle-Calédonie, Nouméa, New Caledonia), Vol. 15, pp. 95–120.

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • MedGen
    Related information in MedGen
  • Nucleotide
    Published Nucleotide sequences
  • PopSet
    Published population set
  • Protein
    Published protein sequences
  • PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...