The flux of genes on the South Seas

By Razib Khan | November 22, 2010 12:29 am

Huli Wigman from the Southern Highlands, Painting of Tahitian Women on the Beach by Paul Gauguin

ResearchBlogging.orgMany demographic models utilized in genetics are rather simple. Yet the expansion and retreat of various demes in post-Ice Age Europe seems to be far more complex than had previously been assumed, though I suspect part of the rationale for the original simplicity was a preference for theoretical parsimony in the face of a paucity of data. The landscapes traversed by our species are rich and topographically convoluted. Not only does the land vary, from plains, to deserts, to mountains, but the climate shifts radically over time and space. In the pre-modern age when humans were more dependent on environmental exigencies these fluxes in ecological and climatic parameters were essential in sharping the arc of human demographic expansion and contraction.

Oceanias_RegionsThis is why a closer examination of the prehistory of Oceania is so appealing: here you have a physical geography which is radically constrained and so reduces the degrees of freedom of human movement and habitation. Unlike Europe, South Asia, or much of Africa, the time depth of the residence of the current indigenous inhabitants of Australia is on the order of 40 – 50,000 years. It seems likely that the indigenous people of the island of New Guinea to the north are from the same original settlement of Sahul, the ancient super-continent which consisted of New Guinea, Australia, and Tasmania. After the initial sweep out to the farthest reaches of what became Tasmania, there was a later push to the east of New Guinea, to the Solomon Islands,~30,000 years before the present. Then nothing for tens of thousands of years. The march of humanity seemed to stand still on the shores of the Solomons, just as the hominin lineage had once been cordoned off from Sahul by the forbidding seas between it and Sundaland, the Ice Age peninsula of Southeast Asia which was later submerged and became the western portion of Indonesia and Malaysia. The stasis was shocked by the Austronesians, a seafaring peoples who seem to have exploded out from somewhere between Borneo and Taiwan within the last 10,000 years, likely just on the margins of written history. The most famous of th Austronesian peoples are the Polynesians, who pushed across the Pacific, and likely even had some tentative contact with the New World. A less well known case is Madagascar, whose inhabitants speak an Austronesian language with clear affinities to a dialect of Borneo. The map below shows rough distribution of Austronesian peoples:

Austronesian expansion

Of particular interest for the purposes of this post is the expanse to the east: Melanesia and Polynesia, Near Oceania and Far Oceania. A new paper in Current Biology , Demographic History of Oceania Inferred from Genome-wide Data, examines the genetics of this region of the world in light of history utilizing a ~1 million marker SNP-chip:

We developed a new approach to account for SNP ascertainment bias, used approximate Bayesian computation simulations to choose the best-fitting model of population history, and estimated demographic parameters. We find that the ancestors of Near Oceanians diverged from ancestral Eurasians 27 thousand years ago (kya), suggesting separate initial occupations of both territories. The genetic admixture in Polynesian history between East Asians (87%) and Near Oceanians (13%) occurred 3 kya, prior to the colonization of Polynesia. Fijians are of Polynesian (65%) and additional Near Oceanian (35%) ancestry not found in Polynesians, with this admixture occurring considerably after the initial settlement of Remote Oceania. Our data support a greater contribution of East Asian women than men in the admixture history of Remote Oceania and highlight population substructure in Polynesia and New Guinea.

Like Dienekes I think there’s something off with the dates they’re generating here. The archeology tells us New Guinea was settled by humans 15-20,000 years before this paper finds that they diverged from other Eurasians! We know that Aborigines are the closest to Papuans genetically, so if they separated from Eurasians less than 30,000 years ago, that would mean that the original inhabitants of Sahul were replaced after that period by the current groups. Far simpler I think to assume that something is off with their timing. Below are the primary figures, a frappe bar plot, PCA, and tree 7b which illustrates the most supported pattern of population branching and admixture.

We cannot display this gallery

There’s nothing too revolutionary in this paper. Rather, it seems to be an exploratory analysis of Oceanian genetics, a precursor to what may come soon. They can not, it seems, differentiate between the slow-boat and express-train models of the settlement of Polynesia, though the consistent pattern of Melanesian admixture seems to lean toward some form of slow-boat, because that is the theory which emphasizes a longer interaction with Melanesian populations.

800px-Area_of_Papuan_languages.svgI know I emphasized the relative simplicity of Oceania in relation to other parts of the world in terms of interpretation because of the geographical constraints, but even here there are layers and twists in the genetic and cultural bedrock. To the left is a map of the Papuan languages. From what I can tell Papuan languages are actually a negation of Austronesian and other well supported language families. The key is to notice that some parts of Near Oceania, Melanesia, have been shifted toward Austronesian languages, though New Guinea is a general exception to this pattern. Remember that humans did not move past the Solomons for ~30,000 years. Large scale settlement of Madagascar seems to have occurred only with the arrival of the Austronesians within the last 2,000 years (after a likely sojourn in East Africa!). This was a genuine cultural revolution which radically shifted the terrain of the possible. And yet by and large the Papuans resisted assimilation to the Austronesian cultural toolkit, which seems to have been otherwise so successful. Why? The Papuans were well equilibrated to their own local ecology, and the Austronesians had no comparative advantage. Rather, the Austronesians, in the form of the Polynesians, struck out into unknown waters and innovated. They found low hanging fruit by discovering trees which had been neglected.

The second interesting point is the bias toward Austronesian mtDNA, and substantial admixture on the Y lineages from Papuans even among Polynesians. The standard explanation of this is that the Austronesians had some aspect of matrilineal descent and matrilocality in terms of communal fission. I think that the Austronesians are arguably a perfect example of the leap-frog pattern of migration, and yet unlike most continental leap-frogs the genetic signal seems to be stronger on the female than male side. There is some evidence of the same in Madagascar. This indicates to me that the Austronesian maritime expansions were qualitatively different from continental leap-frogs, which often were based on the mobility of men on horses.

As I said, the picture remains broadly the same. But there are some touch ups and clarifications on the margins, and that is worthwhile. And definitely an appetizer for what is to come.

Citation: Wollstein A, Lao O, Becker C, Brauer S, Trent RJ, Nürnberg P, Stoneking M, & Kayser M (2010). Demographic History of Oceania Inferred from Genome-wide Data. Current biology : CB PMID: 21074440

Image Credit: Nomadtales, Wikimedia

CATEGORIZED UNDER: Genetics, Genomics
  • Pingback: Tweets that mention The flux of genes on the South Seas | Gene Expression | Discover Magazine --

  • German Dziebel

    As I’ve been arguing on Dienekes for quite some time now (see, e.g.,, recent linguistic research seems to have established genealogical relationship between Austronesian and Jarawa-Onge in the Andaman islands. If looked next to the Austric proposal, based largely on Nicobarese data, this finding casts some doubt on the recent out of Taiwan theory of the origin of Malayo-Polynesians. On the one hand, the amount of “Papuan” admixture in Melanesians and Polynesians (Y-DNA O, which is frequent in Taiwan is rare east of the Wallace Line) requires a rather significant time depth. On the other hand, Y-DNA C*/C2/C2a lineage, which is frequent in Polynesians doesn’t seem to be derived from non-Austronesian (NAN) speakers in southern Wallacea and island Melanesia (it’s much more common in AN than in NAN, which means that gene flow carrying this lineage went from AN to NAN, not the other way around). Y-DNA hg C is phylogenetically ancient and stratigraphically comparable to Y-DNA hg D, which is found in Andamanese. (It’s fixed in Andamanese, just like Y-DNA C is nearly fixed in some Polynesian populations.) It’s also virtually absent in aboriginal Taiwanese. This means that the Andamanese-Malayo-Polynesian axis seems to be of Pleistocene age, while the Taiwan is squarely Holocene. This suggests that Austronesian is an old language family that existed for a lengthy period of time as a dialect chain (think of Ainu, who are Y-DNA C, D and slightly O), not as an expansive language family, and that East Asian/”Mongoloid” gene flow in the Holocene didn’t bring the Austronesian languages with it, but rather spread across the already differentiated Austronesian dialect chain.

  • Scott the mediocre

    Re the Austronesian expansion map – Polynesians reached Clipperton :) ? (I assume it’s shown that way in the map because Clipperton is part of French Polynesia and it made the map easier to create).

    @German Dziebel –

    Fascinating – will have to read more. The link you give to Dienekes is broken, BTW. As a summary for the amateur, is the idea that the Malayo-Polynesian languages for some to-be-explicated reason differentiated much more rapidly in Taiwan than elsewhere (or is the hypothesis of greater diversity in the Taiawanese aboriginal languages vice elsewhere in the standard language family contested?).

  • Simon

    German – The recent linguistic research “establishing” links between Austronesian and the Andaman languages is interesting – but far from conclusive. As for Austric, well that’s even more debatable. If anything, linguists in the are seem to be leaning more towards linking Austronesian with the Tai-Kadai and Sino-Tibetan languages – which would fit quite nicely with a Taiwanese origin.

    Also – look at the linguistic data sometime ( is a good start), there’s just no way the Malayo-Polynesian languages are that old. The differences are much more consistent with a 4000 year divergence than with a 15000 year divergence.

    As for the time-depths that you’re citing – these are known to be overinflated by crappy methodology (anything using “rho” dating is broken – period).


  • Scott the mediocre

    Oops, I was wrong (check first, post second) – Clipperton is not part of French Polynesia; it’s directly under the Minister of Overseas France. But it was administered from French Polynesia until 2007 – maybe the map was created from data predating then.

  • Peter Marsh

    Re; “Fijians are of Polynesian (65%) and additional Near Oceanian (35%) ancestry not found in Polynesians, with this admixture occurring considerably after the initial settlement of Remote Oceania”.
    The official story is that Polynesians dispersed out from Fiji, yet the Melanesian/Fijian admixture occurred considerably after the settlement of remote Oceania! Other DNA studies agree with this; (Stoneking, Bing Su) indicate this admixture occured merely 1,000 years ago.
    The mtDNA haplotype B of Western America and Polynesia both show a dispersal from East Asia 6-8,000 years ago (these are the Tai/Hai culture – people from Taiwan, Thailand, Tibet, Tlingit, Haida Gwai’i people of Canada, Hawai’ians, Pima, Maya and Taino people of the Virgin Islands). Their arrival in Eastern Polynesia was 2,200 years ago, where they evolved quite separately from Western oceania.
    Hawaiian royalty have a history describing their ancestry back to chief Nuu (Nuutka)who arrived with his family from the coastal islands of Canada/Alaska 2,200 years ago, which agrees with the genetic evidence, so why keep trying to push the fast train/slow train/three I’s models which make no sense at all.
    To believe Eastern Polynesians evolved out of Lapita people living amongst Melanesians for over one thousand years without interbreeding is to believe in the unbelievable. Wake up people, use your logical brain and weigh up the facts that keep being swept under the carpet – you have been hoodwinked!

  • Spike Gomes

    Hoo boy, I ain’t even gonna go into what’s wrong with *that* one. We still have followers of Heyerdahl around?

  • Razib Khan

    tx for bringing this to my attention. no more weird revisionism on this thread for now. kind of getting old.

  • German Dziebel

    @Scott the mediocre

    Strange that the link is broken. Here it is again Apologies if this doesn’t work either.

    “the idea that the Malayo-Polynesian languages for some to-be-explicated reason differentiated much more rapidly in Taiwan than elsewhere (or is the hypothesis of greater diversity in the Taiwanese aboriginal languages vice elsewhere in the standard language family contested?).”

    Austronesian languages are divided into 9 small Formosan branches and 1 huge Malayo-Polynesian branch. Notably, Malayo-Polynesian languages aren’t found in Taiwan. It means, linguogeographically, there’s no trail from parts of Taiwan out into the Philippines, etc. The Formosan branches are established on the basis of sound patterns. If, for instance, a new sound law(s) is discovered that will explain the divergence of Formosan branches from Malayo-Polynesian, then Formosan will become derived, albeit very diverse. E.g., Ilya Peiros argued that the alternation of c and t in Formosan dialects is not an ancient relic but is caused by a stress pattern that took place after the breakup of the Austronesian family.


    “The recent linguistic research “establishing” links between Austronesian and the Andaman languages is interesting – but far from conclusive. As for Austric, well that’s even more debatable. If anything, linguists in the are seem to be leaning more towards linking Austronesian with the Tai-Kadai and Sino-Tibetan languages – which would fit quite nicely with a Taiwanese origin.”

    I should’ve softened the “established” language, but Blevins’s research ( comes from a very respectable institution, hence I give it a bit more credit than probably warranted at this point. It’s noteworthy, however, that Ongan and Nicobarese are neighbors and the Ongan-Austronesian link was suggested without any Austric connection in mind. The Austric hypothesis, too, received new life in the works of Lawrence Reid and it’s rather intriguing and independent of the Ongan-Austronesian research. What struck me is the fact that Ongan Y-DNA D (D is also rather frequent in Tai-Kadai, up to 10%, which is high for this haplogroup) seems to be coeval phylogenetically with C*/C2/C2a, which is present in Malayo-Polynesians and absent in Taiwanese. Coincidence? Possible. But also note mtDNA M31/M32 in Andamanese and Q in Melanesians and Polynesians – they are also phylogenetic neighbors and again with no parallels in Taiwan.

    “Also – look at the linguistic data sometime ( is a good start), there’s just no way the Malayo-Polynesian languages are that old. The differences are much more consistent with a 4000 year divergence than with a 15000 year divergence.”

    I guess it’s true but, as Blevins pointed out, languages with lots of monosyllabic roots (and Ongan and Austronesian seem to that kind of languages) may maintain traces of genealogical relatedness for longer than linguists usually expect. So sometimes stability is a sign of antiquity, while diversity is a sign of recency. Think of Ainu: unlike Formosan languages, it’s a family of one, not nine, hence it’s not diverse at all. But there are strong reasons to believe that it’s of Pleistocene age. How did it maintain it’s stability? Probably because innovations have been diffusing back and forth across the dialect chain without much population movement. I suspect that could’ve been the case with Austronesian before it expanded so widely.

    @ Peter Marsh

    “To believe Eastern Polynesians evolved out of Lapita people living amongst Melanesians for over one thousand years without interbreeding is to believe in the unbelievable.”

    I don’t believe that Polynesians came from America, but it’s strange indeed that 9 bp deletion, which is often fixed in Remote Oceania, was not detected in Lapita remains. See Hagelberg here ( and here (

  • Sandgroper
  • farmiddle

    I’m just amazed that there was no Austronesian expansion into Australia. They hit all these small and distant islands, but miss a giant continent!

  • Simon

    farmiddle – maybe they just had good taste.

  • Peter Marsh

    With regard to a commonly held assumption that the Lapita people were proto-polynesians.
    Firstly the DNA was different;
    1. Lisa Matissoo-Smith in her interview on TV NZ (Tagata Pasifika Lapita special 3 2005) said; “We were able to look to see whether the individual possessed a particular mutation that we see at a very high frequency in Polynesians. It is a 9based pair mutation of Mitichondrial DNA and we found that the Teouma material – the first samples that we analysed did not have that mutation, so they did not look like 98% of the people we see living in Polynesia today.”
    2. A recent study by Johnathon Friedlaender et al, titled THE GENETIC STRUCTURE OF PACIFIC ISLANDERS shows that Polynesians have no genetic relationship to the genetically diverse group collectively called Melanesians.
    3. S.W. Serjeantson “The Colonization of the Pacific – A Genetic Trail 1989 pp 135,162-163,166-7. SW Serjeantson comments with regard to the assumed Lapita/Polynesian connection; “It seems quite implausible that a group supposedly evolving within Melanesia could have acquired, by chance, so many non-Melanesian genes! The following genes set them apart: Polynesians lack HLA-B27 , wheras it is common amongst Melanesians. HLA-Bw48 is commonly found in Polynesian populations, but occurs only sporadically in Melanesia. The only other known population with an appreciable frequency of HLA-Bw48 is that of the North American Indians or more specifically the Tlingit (Prince of Wales Island).

    Secondly, their skeletal structure was different; In the paper; WHO ARE THESE PEOPLE? Van Dijk says;”Polynesians actually show more differences than similarities, and it is these differences we should concentrate on. It appears fairly clear that the Lapita people were quite phenotypically distinct (Pietrusewsky 1989, Katayama 1990) from what we idientify as Polynesian today.” Pietrusewsky notes; “Skeletal and dental features which clearly differentiate the Lapita remains from other Pacific groups include wide low mandible shapes, small teeth and slender long limb bones.”
    Van Dijk concludes; “In a cluster analysis based on the results of mandibular measurements the Lapita remains were isolated and furthest removed from Polynesians”.

    Thirdly their toolkits were different. Polynesian culture used; two piece fish hooks, trolling lures, harpoon head, whale tooth pendant, tattooing needle, ground stone pounders, tanged polished stone adzes and gourds. Lapita culture used pottery, buried their dead in urns and used shell money. All these above items were mutually exclusive to each culture.

    Fourthly the chronology of the two cultures are separated by 700 years, there is no overlap!

    So when people discuss Lapita culture as if they were the proto-Polynesians, ask yourself; Where is the evidence? Make fun of the messenger if you like, but if you truly believe the two cultures were related in some remote manner, then SHOW ME THE EVIDENCE!!!

  • TGGP

    Like farmiddle, the absence of expansion into Australia really stands out to me on the map.

  • Peter Marsh

    There is some evidence of temporary visitation to Australia, by Macassar sea merchants along the northern coast, (Bechedemere, sandalwood etc) then there is the stepped pyramid at Gympie, according to Max Gilroy, there was another up Trinity Inlet and one near Atherton. Also dry stone walls up the Burdekin river. Possibly the same people who made stepped pyramids on Mauritius and Canary Islands (The Sea people??).
    The activity up the Mary River and near Noosa is associated with gold mining and aboriginal names which means “Where the sun gods came to land”. These people probably had more to do with Persian, Tamil Nadu, Berber (Bell Beaker) and Lapita sea traders, than with Austronesian expansion. So yes it is interesting that although Australia was visited by these people it was not colonized by them.

  • Simon


    You have some very selective “evidence” there.

    1. the 9 base pair deletion is just one marker of Polynesian prehistory. There are many others. Don’t read too much into a radio interview. (p.s. Lisa’s name has one s)

    2. Reread the Friedlaender paper again: “Polynesians are closely related to Asian/Taiwanese Aboriginal populations, while they are very weakly associated with any Melanesian groups (the closest association there appears to be with New Ireland populations).” (p. e19).

    3. Citing a genetics paper from NINETEEN EIGHTY EIGHT is hardly keeping up with the latest results.

    Finally – and your biggest problem – NOBODY treats Lapita as “proto-Polynesian”. Lapita is the ancestor of the Oceanic societies – proto-Oceanic if you will. This includes Polynesian, Micronesian, along with all the people in Vanuatu, New Caledonia, Fiji, as well as many groups of Austronesian peoples in Near Oceania. Proto-Polynesian. If you trace the linguistics, archaeology, and genetics back it becomes clear that Polynesians are descended from a Central Pacific stage (including Fiji), which are descended from Proto-Oceanic.


  • Lassi Hippeläinen

    Peter Marsh must be a Mormon. Their scriptures claim that Polynesians are Labanites.

    The reason why Proto-Lapitans missed Australia is probably that the coast wasn’t clear. They were sailing maybe in single boats, probably north of Papua. Their small groups were not strong enough to make a permanent beachhead against the incumbent population. Only small islands that were not suitable for Melanesian agriculture were available for them. But their boats were more seaworthy than what the Melanesians had, and therefore they could reach current Polynesia, where nobody had been before. The rapid passage through Melanesia explains why they didn’t mingle much.

  • Peter Marsh

    Ooh – thats a lot of speculation. May I speculate a bit more. Austronesians of Indonesia were active sea traders in all directions – The Borneo Ibans used to be active sea traders. Archaeology suggests obsidian from New Guinea found its way back to Borneo. So were the Torajia and Bugis. Then the sea trade route between Kiribati and Moluccas and down to Samoa. Then there is the old seaport of Nan Madol, and signs of a very significant megalithic culture on Tinian. It is clear that South East Asia has seen many changes. Personally I believe Austronesian has very ancient links in SE Asia during the last Ice Age, when human populations were compressed into the tropical zones. Austronesians of Taiwan may well have been a northward moving group as the the world warmed up, pushing the previous Jomon (Ainu?) people northward. As the final flooding of the ice age precipitated a mass exodus of people from the flooding coastal plains 6-8,000 years ago, many people migrated via the Kuroshio current to America (mt DNA haplotype B) – common to native Americans and Polynesians. HLAbw48 of the Tlingit and Kwakuitl and Haida people (from Haida Gwai’i suggests the departure point of Polynesians to Hawai’i – the Polynesian homeland.

  • Peter Marsh

    In response to Simon;
    The 9 based pair deletion is not found in Melanesia, but is found amongst almost all true Polynesians as well as a number of native American tribes – Tlingit, Kwakuitl, Haida, Chumash, Pima, Angiate and Maya. So to me it is a very important marker – so is HLA Bw48 common to some native Americans and almost all Polynesians. Lisa Matissoo-Smith as agreed to me that it too is an important marker. Who is stopping much needed research in this direction?
    I have no problem with a Polynesian exodus from Taiwan 6,000 years ago, but where was their homeland for 3,800 years before their arrival in the central Pacific?
    My argument is that there is absolutely no evidence of their unique mix of DNA passing through Melanesia and surviving intact.
    Furthermore, most DNA evidence shows that almost all Polynesian DNA found in Melanesia only arrived 1,000 years ago, and when they did they had no problems assimilating, creating an harmonious blend between both cultures.
    I also disagree with Simons comment that “NOBODY treats Lapita as “proto-Polynesian” Much of Patrick Kirch’s and Roger Greens work is based on this assumption. They have both been very influential in governing the direction of Pacific research over the last fifty years.
    Besides, Lapita only arrived 3,900 years ago in the Bismark archipelago – without any formative phase, sure these sea merchants with shell money from the Western Indian ocean influenced Melanesian culture – especially in pottery making, but there absolutely no evidence to show Lapita culture was imparted into Polynesian culture as their tool kit was completely different; two piece fish hooks, trolling lures, harpoon heads, tanged adzes, polished stone grinders and tattooing needles – much the same as the people of the Canadian NW.

  • Simon

    Peter – you are being fast and loose with the evidence.

    1. There is no evidence of B4a1a1a (the Polynesian Motif) existing in Native American tribes.

    2. There is also a LOT of evidence that “Polynesian” DNA “passed through Melanesia” – there is huge amounts of admixture in both the Mitochondrial genomes, Y chromosomes, and whole genomes. Reread that Friedlaender paper you cited earlier. There’s been at least 10 papers in the last few years showing this.

    3. Re: Kirch and Green – that’s absolute rubbish. They were very careful to disentangle Lapita (what linguists would call proto-Oceanic) from proto-Polynesian. Proto-Polynesian is a few major steps down the chain: Proto-Oceanic -> Proto-Central-Pacific -> Proto-Polynesian. Lapita is relevant to Proto-Oceanic and Proto-Central Pacific (i.e. Fijian). Pat Kirch and Roger Green did not assume otherwise. I had many long conversations with Roger Green on this matter, and Roger was ALWAYS very careful to distinguish these.

    4. Lapita ‘had no formative’ phase – what about the red-slipped pottery in the Philippines? There’s a number of papers by Peter Bellwood on this.

    5. Of course there was no whole-sale “imparting” of Lapita culture into Polyesian culture. Lapita died out a thousand years before Polynesian culture was born. There ARE enough similarities to show that these are linked though.

  • Clark

    Peter Marsh must be a Mormon. Their scriptures claim that Polynesians are Labanites.

    No idea if Marsh was Mormon but there’s nothing in LDS scripture about Polynesians being Lamanites. Not even remotely. There are some folk traditions about that but they don’t have much to them, beyond a lot of Mormon polynesians liking the idea of having a special place in Mormon thought. It all comes from a single verse of scripture about some people going north on a boat and never being heard of again. Not a lot to build a big claim off of. I think the folk tradition is more due to Mormon naivete and perhaps a bit of latent racism in early 20th century Mormons. (i.e. anyone not European must be Indian)

    My personal belief is pretty much what’s been outlined in this thread.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar