How did modern humans settle the world?

By Razib Khan | July 9, 2011 11:47 pm

In lieu of lots of text, above is a stylized representation of the routes which Neo-Africans took ~50 thousand years ago from their point of departure to parts unknown. The two colors represent two models. The red lines show two major streams issuing out of Africa, a northern route which pushed into the heart of Central Asia, and a southern oceanic one, which pushed all the way into Australia. The second differs, with eastern and western branches of non-African humanity. The models really start to break down within the last ~10,000 years. For example, by either model India has seen an admixture even between the two branches in the Holocene. Additionally, there may have been “false dawns” and admixtures.

In the early 2000s I accept the probable likelihood of the first model. But today I am more leaning toward the second. What’s your stance, and why? I’ll give my rationale below….

The primary reason I’m skeptical of the northern vs. southern route is that Melanesians are clearly southerners, while East Asians and Europeans are northerners. But the phylogenetic stuff I’ve run myself and that I’ve seen using autosomal data sets indicate that East Asians and Melanesians are far too close for Melanesians to be the distant outgroup to the European-East Asian clade. The archaeology is clear that Sahul, Papua and Australia, were settled ~40-50,000 years B.P., while the molecular evolutionary and archaeological evidence which points to a common origin of East and West Eurasians implies a date of ~20-30,000 years for the last common ancestor (probably the Gravettians or their relatives). Something doesn’t smell right here. It may be that the contemporary Melanesians and Australians are more recent arrivals, but I’m skeptical of that. Additionally, in “Ancestral South Indians” are clearly closer to East Eurasians than they are to West Eurasians. It seems that the West Eurasians are best modeled as an outgroup to East Eurasians and ancient South Eurasians (granted, the latter clade has relatively distinct and distantly related branches).

I could say more. But those are the big issues. I just have a hard time accepting the ~20-30,000 year figure to the last common ancestor for Europeans and East Asians if Melanesians settled Sahul ~40-50,000 years B.P. This is why I’m favorable to a shift of the mutation rate which recalibrates these estimates. Doubling the last common ancestor of East and West Eurasians to ~40-60,000 years would make much more sense of the model in my own head.

  • http://lablemming.blogspot.com/ Lab Lemming

    From a climatalogical point of view, I don’t like assuming important human habitation and migration through central asia anytime between about 60K and 20K (calendar years; Ymutation rateMV). Are there any central asian archeological/fossil sites from that time? I was under the impression that northern/ euro migration was surpesssed by climate and/or inability to compete with neandertals until better tools became available in the 40-30K range. But that is probably based on out-dated work- I haven’t paid attention to the field for a while.

  • carpetanuiq

    My take on this (very simplified) is as follows.

    First the OOA was through Levant not through the Corn of Africa.

    At Levant, a first western-eastern division arised (reflected by Y-chromosome division E and C; let´s call them E and C for the rest, just names). The E branch went to west eurasia (and was responsible of all the archeology in this area until post-LGM); the C branch went to east-eurasia through some northern route (maybe Central Asia, up to Amuria).

    From there this C branch followed the following route: EA>SEA>OCEANIA (I mean PAPUA-NG and Australia)>SA. In Northern SA (possibly Upper Indo) there was a new split of this branch (IJK derived Y-chromosomes explosion) posibly at post-LGM times. Some went north-east, east and south-east, mixing with previous C populations (O, Q, M…)in the east and some to west eurasia, mixing with E populations (I, J, R). This admixture with pre-LGM populations is what explains all the puzzling issues (and also the fact that west eurasians and east eurasians look so different while beeing closely related). IMO, no need to predate anything .

    The rationale under this hypothetical route is that it is the only one I know, up to now, that explains correctly all the known data (I´m very interested in knowing which undisputed data are not explained by this hypothetical route). It allows us also to make two predictions: first, the west-east southern route will turn out to be a hoax; second, if ever Y-chromosome from palelolithic preLGM Europe or West Asia is recovered, it will be E.

  • bob sykes

    I think you need to combine genetic information with climatological information. There are two interesting maps of glacial extent and Central Asian archaeological sites at

    http://www.uib.no/form/aktuelt/polararet/prosjekt/icehusII2.htm

    There are a number of archaeological sites ca. 30,000+ ybp along the Urals.

    There is even a narrow route between the ice sheet maximum and the then huge Caspian Sea ca 90,000 ybp.

    So, the northern routes seem possible ca 50,000 ybp but I would need to see some really detailed info regarding the climate along that path.

    On the other hand, both the red and blue expansions show a track right through the Himalayas and Tibetan plateau. Is that correct? I am highly suspicious of any route through mountainous terrain.

  • vg

    Bob Sykes – On the other hand, both the red and blue expansions show a track right through the Himalayas and Tibetan plateau. Is that correct? I am highly suspicious of any route through mountainous terrain.

    On the map, it looks more like the northern Indian plains. That wide gap that you can see on the map between the red/ blue lines in the south and the red line in the north is the Himalayas – Karakoram – Pamir – Hindukush and the Tibetan plateau.

  • http://lablemming.blogspot.com/ Lab Lemming

    There was {^ck-all ice at 90Ka- maybe 25% of LGM: 18O curve at:
    http://www.museum.state.il.us/exhibits/ice_ages/specmap_graph.html
    But that’s irrelevant, because 90Ka is too early anyway.
    A brief squiz of origins.net suggests neadertals in central asia through 40-30Ka, which is well after Australian settlement.
    From the habitibility POv, the problem is not ice sheets, but loess: too dry for vegetation.

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    (and also the fact that west eurasians and east eurasians look so different while beeing closely related).

    i wouldn’t say they’re closely related. i mean, compared to africans yeah i guess.

  • Insightful

    I just have a hard time accepting the ~20-30,000 year figure to the last common ancestor for Europeans and East Asians if Melanesians settled Sahul ~40-50,000 years B.P. This is why I’m favorable to a shift of the mutation rate which recalibrates these estimates. Doubling the last common ancestor of East and West Eurasians to ~40-60,000 years would make much more sense of the model in my own head.
    Hmmm, it would be interesting to get Dienekes’ take on this to see how he comes down on it..

    @carpetanuiq and also the fact that west eurasians and east eurasians look so different while beeing closely related

    Sometimes you have to look at things in context, such as in the context of nature. Eurasia is a huge landmass it would make sense that their would be such within species variation just as there is elsewhere in nature. Imagine a Eurasia where from the British Isles to Japan all people had the same skin color, same hair and same general facial features regardless of latitude north and south, east and west right across the immense Eurasian landmass. Knowing what we know about nature It wouldn’t make sense now would it? One of the things we do know is that East and West Eurasians independently evolved lighter skin pigmentation. That is why although the Chinese are light their skin is still a different color from the Europeans..

  • Insightful

    Razib, I’m thinking maybe those archaeological dates and finds need to be re-examined with the latest technology before considering a shift of the mutation rates. It might end up bringing the two roughly in line

  • gcochran

    Considering that the evidence for the high mutation rate was always extremely shaky, while the new estimates are based on direct measurements, it’s hard to see what the old estimate has going for it, other than tradition.

  • http://forwhattheywereweare.blogspot.com/ Maju

    I don’t think either model is correct, however the blue one is maybe closer: you just need to make the western blue line beginning not in West but South Asia (and maybe add a buckle, i.e. forth and back migration, between South and SE Asia.

    My model is this map, which I drew ad hoc in few minutes for this commentary.

    Why? Because of the structure and geography of the mtDNA scatter. By considering the phylogeny of M and N (and sublineages), with emphasis on basal branches and their geography, we obtain centers of gravity for M and N in South and SE Asia respectively. These surely approximate the true centers of expansion of each haplogroup.

    Dedicating some time to the details, complementary approaches and necessary revisions (see here for example) we can reconstruct the map linked above with quite a good deal of precision. We can even fit the Y-DNA phylogeny on it with some good degree of consistency (see here).

    Of course, it requires forgetting about the TMRCA nonsense, something that I do quite well because I lack faith in the Molecular Clock (I’m of the Spaghetti Monster sect actually).

  • http://blogs.discovermagazine.com/gnxp Razib Khan

    Why? Because of the structure and geography of the mtDNA scatter. By considering the phylogeny of M and N (and sublineages), with emphasis on basal branches and their geography, we obtain centers of gravity for M and N in South and SE Asia respectively. These surely approximate the true centers of expansion of each haplogroup.

    your model seems similar to stephen oppenheimer’s. i disagree with a lot of his ideas, but a “out of india” after “out of africa” is something i’m kind of leaning toward at this point.

  • http://forwhattheywereweare.blogspot.com/ Maju

    I’m surprised to “agree” with Oppenheimer, who usually makes claims I cannot easily sympathize with, such as a massive extinction in South Asia because of Toba or some “Basque” colonization of Ireland… ignoring all the rest of Europe quite confusingly.

    But whatever.

    South Asia must have been central. The massive star-like structure of M (there’s nothing like that in all the mtDNA phylogeny except possibly H) indicates a very quick expansion, which would have happened as the migrating population got free from the climatic constraints of Arabia, upon arrival to South Asia. (The Y-DNA parallel of M would be F).

    However we cannot ignore that there appears to have been another radiation center in SE Asia: not just I estimate that N probably coalesced over there but also Y-DNA D and C look totally original from that region as well (and MNOPS, although this one is less basal). So I understand that the pulse that got the first Eurasian people to expand in South Asia and beyond towards the East (no real barrier, just a hill/jungle buffer by Assam) was fluid enough somehow to allow for a backflow.

    While I do not fully discard that this backflow is associated to Toba (is this in which you think I agree with Oppenheimer?), the relative low impact of mtDNA N/R in South Asia strongly suggests that there was no “extinction” anyhow. There may have been a “cultural/genetic vector” however in that backflow (Y-DNA MNOPS -> P and mtDNA N -> R) but at this point (with R also flowing eastwards again and IJ being incorporated to the Western flow) it’s not anything neat but in continuous admixture and flow.

    What I am more and more persuaded about is that your ancestors’ homeland, Bengal, played a pivotal role in this backflow westwards from SE Asia. I’d say it is at the origin of Y-DNA P and probably also mtDNA R, which are the core of this late oscillation of what is best named the Eurasian expansion.

  • http://washparkprophet.blogspot.com ohwilleke

    While I would agree that the Blue line in the original post map seems closer than the Red line, there are nuances.

    For example, mtDNA haplogroup X, found mostly in West Eurasia but also a minority regionally limited haplogroup in North America, is suggestive of a thin Northern route presence that is also supported by some of the components of the Northeast Asian genome that don’t have obvious modern origins.

    The Northern route traffic appears to have been two ways, however, with Y-DNA haplogroup N appearing to have made a Southern Route trip followed by a trip back across Siberia from East to West to reach Finland.

    The case that Southern route migration happened in distinguishable waves or pulses also seems quite strong. The older you get, the more homogeneous populations seem to have been genetically.

    But, given the unreliability of mutation rate dating and the paucity of archaeological evidence that can be firmly tied with ancient DNA from these eras, the sequence of the waves is hard to pin down. We haven’t really solved the question of human expansion out of Africa until these waves can be better resolved and understood separately and in sequence.

    For example, in most of the places where Y-DNA haplogroup D is found (the Andamans, Japan, Tibet) it seems as if the non-Y-DNA haplogroup D elements have clear sources in later admixture and migration events much later in time. The initial Upper Paleolithic Y-DNA haplogroup D migration does not appear to have involved a broad mix of haplogroup types. While the data is a bit muddy, it appears that in all three of these cases, mtDNA haplogroup M was present, while no mtDNA haplogroups that were part of N or R were present.

    The presence of what appear to be a number of Y-DNA haplogroup D individuals in West Africa of a few basal branches of the haplogroup, and the fact that Y-DNA DE is the common origin for both Y-DNA haplogroup D and (predominantly African) Y-DNA haplogroup E suggest a common African origin for both with D becoming distinct at around the time of this wave leaving Africa.

    But, we can’t really tell if D/M was a first wave broken up by a latter CF/MN wave, or if D/M was a relative late comer that could only settle where there was virgin territory.

    Another thing that the big picture human expansion map fails to convey is that extent to which there have been massive population replacements and folk wanderings in the Holocene and prehaps als before (e.g. at the Last Glacial Maximum). Much of the world shows little genetic continuity with the Paleolithic population of that region, or even the early Neolithic population of that region.

  • http://forwhattheywereweare.blogspot.com/ Maju

    #13 Andrew: mtDNA X is a West Eurasian and North African haplogroup which surely coalesced in the Levant. A subclade of a subclade (X2) is present in Central Asia and then another subclade of X2 is present in North America. This is totally consistent with the diversity pattern of Y-DNA Q: most diverse in West Asia but dominant in America.

    This is not a “northern route” within the early Eurasian expansion but within a second or even third moment: it derives from the Western subgroup and formed surely only after 48-40 Ka. ago, when we can confirm that Altai was occupied by anatomically modern people (Homo sapiens) with West Eurasian technologies (“Aurignacoid”).

    This is not a parallel and equivalent “northern route” to the southern route but a minor tertiary branch in a very advanced stage. It is comparing the trunk of a tree with a small branch.

    “Much of the world shows little genetic continuity with the Paleolithic population of that region, or even the early Neolithic population of that region”.

    Highly speculative and the way to make sure we cannot arrive to anywhere: by denying a priori the validity of ALL the evidence, you renounce to reach to any conclusions at all (what may be very convenient when a pet theory, wishful thinking, is threatened by the hard facts).

  • http://washparkprophet.blogspot.com ohwilleke

    @14 ***“Much of the world shows little genetic continuity with the Paleolithic population of that region, or even the early Neolithic population of that region”.

    Highly speculative and the way to make sure we cannot arrive to anywhere: by denying a priori the validity of ALL the evidence, you renounce to reach to any conclusions at all (what may be very convenient when a pet theory, wishful thinking, is threatened by the hard facts).”***

    Ancient DNA, physical anthropology, the population genetics of modern populations, and in some cases recorded history all support massive population replacement in much, and perhaps all, of Europe between early Cro-Magnon and latter Europeans, at the Neolithic, and probably at least once and sometimes more times thereafter in various parts of Europe. Gypies and Jews have ancesteral historical roots outside Europe in the historic era that genetics confirm (admittedly with significant local admixture).

    We have historical accounts of multiple waves of major population replacements and relocations in East Asia and Southeast Asia that are confirmed by population genetics. There is genetic and archaeological evidence for at least two waves of Tibetan settlement. We know that there were major population infusions in Japan in the Yaoyi period and that there was probably major demographic transfer from China for centuries after the Yaoyi arrived.

    The notion that there was some sort of ancestral North Indian incursion into the rest of South Asia in an era on the order of the Bronze Age is widely accepted, as is the relatively recent incursion of Siddis and Tibeto-Burmese components of South Asia’s population.

    Population genetics and history confirm major population replacements in the Americans post-Columbian, and in some areas (for example, the Na-Dene in the Southwest of the U.S.) archaeology and linguistics support some level of population replacement in late pre-Columbian America. Ancient DNA clearly support at least two separate instances of near complete demic replacement in circumpolar North America prior to Columbus and culminating in modern Inuit populations.

    Bantu expansion was not demic in all of Africa, but in much of Africa, the population genetics clearly indicate that population replacement took place. There is also strong population genetic evidence that suggests a major pre-Bantu expansion Neolithic expansion in West Africa. Similar signs of Neolithic era or later Nilo-Saharan expansion/migration are also present in genetic evidence. Demic impacts on Ethiopia from Ethio-Semitic incursions are quite visible in the population genetics and linguistics. No one seriously questions that about half of the genetic roots of Madgascar are attributable to Indonesia ca. 1000 CE. There is good reason to believe that the Khoisan migrated to Southern Africa from origins in Eastern Africa.

    The demographic impacts of Russian expansion into Siberia and Han Chinese expansion into Western China in relatively modern times are well established.

    New Zealand’s Maori have been there only a thousand years, as we know from archaeology, and European descent individuals only for hundreds of years. Australia’s aborginines have much deeper roots, but most Australians, of course, descend from immigrants in the last two hundred years.

    There are a few places in the world where there is credible evidence to link modern populations to Paleolithic populations in the same region (e.g. Papuans, Australian Aborgines, Ainu, Andamanese), and others (South India, Indonesia) where the impact of fairly ancient local populations on modern populations mixes appears to be rather great although there are more recent strata evident as well. But, these are the exceptions, not the norm.

    There is nothing speculative about it and it is hardly a denial of the validity of all evidence. Rather it is a recognition that simple models are insufficient. At the very least, waves of migration and replacement have to be wound back as far as possible based on what we know, before serious speculation on what human population structure before “discernable pre-history” looked like can begin. A static model that assumes that the people who live in locations now descended from the people who lived there millenia ago, doesn’t cut it and is inconsistent with multiple lines of evidence almost everywhere.

  • http://forwhattheywereweare.blogspot.com/ Maju

    (@15) Andrew:

    “Ancient DNA, physical anthropology, the population genetics of modern populations, and in some cases recorded history all support massive population replacement in much, and perhaps all, of Europe”…

    Ancient DNA only seems to support massive population replacement in parts of Central Europe (not in Iberia, for example). The population genetics of modern populations don’t tell me any clear history of replacement and definitively not one of massive replacement in all Europe. Not at all.

    “Gypies and Jews”…

    Small minorities that have been largely admixed with host populations. If all migrations in the past were like these two, there is no doubt that I am right.

    “We have historical accounts of multiple waves of major population replacements and relocations in East Asia and Southeast Asia that are confirmed by population genetics”.

    The best known one is the Yaoi migration into Japan, yet modern Japanese are at least 50% pre-Yaoi. So good for the massive population replacements that leaves no trace of whatever was there before.

    “Bantu expansion was not demic in all of Africa, but in much of Africa, the population genetics clearly indicate that population replacement took place”.

    Bantu expansion is the best you can come with but then we find that Mozambicans are “something else” and anyhow most Bantu expansion happened at the expense of hunter-gatherers with Iron Age tech (there was no expansion before steel).

    “The demographic impacts of Russian expansion into Siberia…”.

    How is this different from the colonization of North America, Argentina, Australia…? How do you dare to compare massive industrial processes fed with unprecedented numbers of people supported with unprecedented technologies with what happened in Prehistory? It’s ridiculous!

    And still, both farmer and foragers (but specially farmers) survived these hyper-modern advances. You still find some pre-European DNA in all those places, where there were consolidated farmers, it makes at least 50% of the genetic pool, in some areas it’s still so dominant that it’d seem that almost no colonization took place, though of course invasion and enslaving did happen. This, the Andean result, is the model I can imagine for all “migrations” other than the Neolithic wave (Indoeuropeans, Celts, Slavs, Germanics…) – but in the case of the Neolithic wave we know (by aDNA) that, where they had an impact, they were later replaced, and in other places they had no impact almost at all.

  • Grey

    “The best known one is the Yaoi migration into Japan, yet modern Japanese are at least 50% pre-Yaoi. So good for the massive population replacements that leaves no trace of whatever was there before.”

    Replacement by invasion, especially in ancient times, might have dramatically different effects on the male and female lines of descent.

  • http://forwhattheywereweare.blogspot.com/ Maju

    Absolutely Grey: I was considering only the male line, the one that may have been affected the most. Today’s Japanese show at least 50% of Y-DNA lineages that are exclusive from Japan (D2, C1) and therefore pre-Yaoi. I am not sure of the female apportions right now but surely much bigger in favor of local ancestry (80%?).

  • dnm

    @Maju

    I think you mean Yayoi. Yaoi is something else…

  • http://forwhattheywereweare.blogspot.com/ Maju

    Sure Yayoi. Wonder what was I thinking, ahem (your links are broken but I found what they mean anyhow). :D

NEW ON DISCOVER
OPEN
CITIZEN SCIENCE
ADVERTISEMENT

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 http://www.razib.com

ADVERTISEMENT

See More

ADVERTISEMENT

RSS Razib’s Pinboard

Edifying books

Collapse bottom bar
+

Login to your Account

X
E-mail address:
Password:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »