In light of the recent results in human evolutionary history some readers have appealed to me to create some sort of clearer infographic. There’s a lot to juggle in your head when it comes to the new models and the errors and uncertainties in estimates derived from statistical inference. Words are not always optimal, and there’s often something left out.
So I spent a few hours creating a series of maps which distill my own best guess as to what occurred over the past 100,000 years. I want to emphasize that this purely my own interpretation, based on what I know. This is naturally going to be biased (I don’t know as much about uniparental lineages as some of my readers, and have a weak grasp of a lot of morphological changes, etc.). But it is a place to start. I’ve put the maps into a slideshow. Please observe that in brackets I’ve put qualifies such as “high”, “medium” and “low” in regards to my assertions. That shows you how confident I am about a given assertion. I’m 100% sure that I’m wrong in a lot of the details here, but this is my best guess as to the shape of things over the past 100,000 years. Feel free to ask more in the comments. Also, take the dates with a little fudge room. If I used exact precise dates for everything there would be too many slides.
Note: You can’t see the slideshow in the RSS browser.
Razib Khan’s degrees are in biochemistry and biology. He has blogged about genetics since 2002, previously worked in software development, is an Unz Foundation Junior Fellow and lives in the western US. He loves habaneros.
September 25th, 2011 at 9:48 pm
Thanks, Razib! Clarifies a lot.
There are a lot of great tools out there for visualizing stuff like this. You might check out Google Maps which lets you draw/share different paths.
September 25th, 2011 at 10:00 pm
don’t have much time to learn new APIs now, but if i have time next weekend i might look at GIS. google maps is usually kinda ugly, if functional though from what i can see.
September 25th, 2011 at 11:26 pm
warning: if you post stuff like people of “semitic appearance” arrived in europe 40,000 years ago, and that albanians are partly african, your comment is so ridiculous that i’m going to delete it (also, if you complain about this policy i’m going to ban you).
September 25th, 2011 at 11:51 pm
I’m speechless. Is this the normal attitude on this blog to delete comments and ban people without even pausing to consider the validity of their arguments? There was nothing ridiculous about my comments. Carriers of haplogroup IJ arrived to Europe circa 40,000 years ago. Most modern descendants of ancestral IJ have semitic appearance. 40% of Albanians have Y-haplogroup E1b1b, which has African origin and is most common in Maghreb nowadays. It would take you all of 5 minutes in Wikipedia to check these facts.
But if you’d rather delete comments than to wonder if they are accurate, there’s no need to ban me. Just say so and I’ll go away – I don’t bother commenting on blogs where owners don’t treat me with respect.
September 26th, 2011 at 12:03 am
I’m speechless. Is this the normal attitude on this blog to delete comments and ban people without even pausing to consider the validity of their arguments?
if i think they’re stupid comments, yes. your whole comment was a mishmash of reasonable and farcical. i’d rather spare my readers confusion, they are an innocent and naive lot.
and if you use the term “african,” you might pause to consider that people will assume you mean sub-saharan african. your comment sounds a lot less funny and idiotic now. but yes, in general i still think it’s stupid to assume that people 40,000 years ago looked just like their current descendants. one reason i think this is stupid is that the morphology is very different in the remains we have. we also know that alleles for light skin seem to emerged relatively recently in eurasia among extant populations.
September 26th, 2011 at 1:08 am
That’s not much of an apology, but it will have to do.
Without a doubt, there has been evolution in physical appearances everywhere. But there are substantial similarities in appearance (skin color notwithstanding) between, say, the British (R1b), Tajiks (R1a) and some Pakistani ethnicities (also R1a), even though they probably did not have any common ancestors for 20+ ky.
Skin color is one of the few areas where we’ve been seeing a lot of visible changes, presumably because selection pressures are strong.
September 26th, 2011 at 1:15 am
But there are substantial similarities in appearance (skin color notwithstanding) between, say, the British (R1b), even though they probably did not have any common ancestors for 20+ ky.
in the case of the pakistanis and tajiks this is absolutely false. probably near an order of magnitude false. you live in a different factual universe than i do.
September 26th, 2011 at 2:34 am
You clearly misunderstood my statement. I did not say that Tajiks and Pakistanis were isolated from each other for 20 ky. That is obviously false. But they were both isolated from the ancestors of the British.
Compare
http://upload.wikimedia.org/wikipedia/commons/thumb/0/09/Huckabee-SF-CC-024.jpg/447px-Huckabee-SF-CC-024.jpg
http://www.opfblog.com/wp-content/uploads/2009/06/Tajikistan-President1.jpg
I’d rather drop this, since “semitic” is a rather subjective term, and, anyway, this was all a purely tangential issue.
September 26th, 2011 at 2:57 am
Why do you put your ‘modern humans’ so far north in Africa? We have remains, sites, artifacts etc. of plenty of them in Southern and indeed South Africa, which doesn’t even feature on your map. I’m betting that the great leaps in MSA development took place in the Western and Northern Cape, and recent datings in Wonderwerk Cave and elsewhere seem to bear me out.
September 26th, 2011 at 3:40 am
I like the simplicity of the graphics actually. The internet is full of over-detailed speculation that looks inherently more certain than it is.
September 26th, 2011 at 4:09 am
Rather than a timed slide show, I think this would work better with Forward and Back buttons. I found I was sitting on the Start and Stop buttons, and didn’t ever want to go to the next slide automatically.
September 26th, 2011 at 6:54 am
Or you might just post your maps as a series of plain still images so people can scroll as they like? It’s convenient and doesn’t take any time for you.
September 26th, 2011 at 9:02 am
thanks for doing this. good idea to had the degree of certainty cuz most wouldn’t do that and then i end up going around bragging about “knowing” something that ends being wrong
September 26th, 2011 at 9:39 am
Ditto what #11 said. Also, instead of writing “[high]” “[medium]“, etc., you could reduce visual clutter by using opacity to indicate certainty.
September 26th, 2011 at 10:03 am
Thanks. Very helpful.
“my readers…..are an innocent and naive lot” – We are. Pure as the driven snow, we are.
September 26th, 2011 at 10:27 am
Also I am confused by the year labels.
When it says “3000 years”, does this mean 3000 years ago, or 3000 years from the start 100,000 years ago (i.e. 7000 years ago).
In short, I can’t tell which is the beginning of the timeline and which is the end.
September 26th, 2011 at 11:03 am
as a graphic designer, im offended by the graphic used. really? all the resources of Discovery magazine and thats the best way you could come up with to illustrate this awesome info? can i get a decent/understandable infographic that justifies the content? this looks like scans of overhead projector slides from high school…
September 26th, 2011 at 12:44 pm
Nameless_: I’d rather drop this, since “semitic” is a rather subjective term, and, anyway, this was all a purely tangential issue.
Erm “Semitic” is not terribly subjective, but there would have to be evidence. And if this is another of those “blah-blah is one of the lost tribes of Israel” Bigfoot/Loch Ness/UFO stories THEN please do go away. It’s just attention-seeking on a racial scale.
September 26th, 2011 at 12:46 pm
bob – It’s years before the present. So the slides are a series of snapshots in time, starting with 100,000 years ago, then 75,000 years ago, etc, down to 3,000 years ago.
September 26th, 2011 at 1:46 pm
? all the resources of Discovery magazine and thats the best way you could come up with to illustrate this awesome info? can i get a decent/understandable infographic that justifies the content?
first, it’s discover magazine. second, this is a blog post. the infographic wasn’t commissioned by them. third, you’re pretty to re-purpose it yourself. since you care to complain i assume you care to invest your own time in producing something better :-0)
September 26th, 2011 at 2:32 pm
BTW, doesn’t that look a lot like the Spencer Wells’ “two waves” model – one westward along the South Asian coast, one diffuse wave through the Eurasian steppe?
September 26th, 2011 at 2:37 pm
This is pretty cool. It’s a more interesting theory than Out of Africa, anyway.
Nameless: It’s obvious the British descended from elephants. Just look at the pictures.
September 26th, 2011 at 3:38 pm
buddy – hey, i’ve never seen a cocky, pretentious graphic designer complain about things on the internet before. this was the first time i’ve seen that. no one cares – shut up and go search around for outdated websites to giggle about with your gizmodo friends
September 26th, 2011 at 3:54 pm
Which browser is recommended to see the slides? I tried IE , chrome, and safari. All show the gnxp web in the box.
September 26th, 2011 at 4:23 pm
I tried IE , chrome, and safari. All show the gnxp web in the box.
some ISPs (university) seem to cause this problem. no idea why.
BTW, doesn’t that look a lot like the Spencer Wells’ “two waves” model – one westward along the South Asian coast, one diffuse wave through the Eurasian steppe?
kind of. i think the coastal route is probably valid. but i still am skeptical of the eurasian steppe idea. rather, i think wave #2 still used southerly routes, albeit more inland.
September 26th, 2011 at 5:15 pm
Also, one last significant thing that’s missing here is a counterclockwise wave that goes from China, through Siberia and then back west all the way to the Baltic some time between 8 kya and 4 kya.
September 27th, 2011 at 7:25 am
I appreciate Razib emphasizing that this is just a best guess from the information available. The more I learn about our distant past the less I think is actally known. Out of 105 identified Neanderthal sites (some as small and as uncertain as a single tooth) 102 have been found in limestone caves. The point I’m getting to with this fact is if hominids didn’t die in one specific location, in a temperature zone condusive to bone preservation we have next to nothing in real evidence. Experts in this area very often overstate what they know, in part to counter the goofy creationists. My best guess is we need to draw a big question mark on any place in the world that was warm and moist or lacking in limestone caves pre agricultural revolution. If southeast asia had the environmental equivalent of the rift valley in Africa then we could at least talk with more cetainty about the world wide movement of pre moderns but since we don’t we can’t conclude much of anything.
September 27th, 2011 at 4:01 pm
Any thoughts on the relationship between all this movement and the evolution of language?
September 27th, 2011 at 4:07 pm
#28, i now believe that language is an older evolutionary trait of hominins than i did in the early 2000s. complexity of language is a different issue….
September 28th, 2011 at 1:18 am
Some other comments based on y-DNA haplogroups:
D is today only found “at the fringe” in Asia, so it got wiped out in India and many other places, presumably by a second wave. That second wave must be C&K related, which we know likely existed in the area >~70,000 ya. That makes DE a good candidate for expansion during the ~130,000-110,000 wet phase (“green Sahara”), and thus an initially inner (not coastal) route through the Levant (ancient AMH finds) and Arabia (tools and settlements documented from that time). A coastal route via the Bab-el-Mandeb straight would have been feasible during the Toba-initiated cold phase and low sea levels after 70,000 ya, but would be too late for the Oceanian finds. A better date (and motivation!) for the second wave (involving CF) IMO is 105,000-110,000 ya when temperatures and sea levels plummeted.
September 28th, 2011 at 6:23 pm
If the bushmen split off 130,000 years ago the graphic presumably won’t have much to say about language evolution since presumably it was all over long before 100,000. (Steven Fry is telling the world on TV that language only started 50k ago :-S .) But I’m wondering about that date for the bushmen. Wondering, and fascinated by it.
September 29th, 2011 at 4:46 am
You have to be careful associating y-DNA haplogroups with autosomal genes and/or appearance – often, there is very little relation. Many regions in East Asia carry D, but some have zero mongoloid features. You can go into a small village in Central Germany and pick three men that look like clones. With some luck, one will carry R1a, one R1b, and one I.
September 29th, 2011 at 4:47 am
The archaeological record is that between ~100,000 ya and 45,000 ya there were no AMHs in virtually all of West Asia (instead, there were Neanderthals and perhaps heidelbergensis-like). So it appears clear that all West Asians and Europeans stem from back-migration from South Asia (and less so from Central Asia, and a little from Africa).
I view IJ and G as ultimately coming from a small area were they were “left behind” and did not participate in further migrations, expansions, and diversification of the other F sub-groups – such as, perhaps, SW Pakistan (their brother groups H and K are S/SE Asian and Australian/Oceanian, respectively, and “parents” F and F* are mostly Indian). Of these, G, I, and J2 are “Caucasian” both sensu strictu and sensu lato. J1 men probably migrated to Southern Arabia, mixing with locals there, and acquired their Afro-Asiatic language there. They also clearly went through a long period of isolation.
September 29th, 2011 at 10:23 am
@33
The dates I have seen are AHMs in Southwest Asia from ca. 100,000 ya to 75,000 ya, then gone from 75,000ya to 45,000 ya, and then back again.
There is one site in South India where there appeared to be continuous AMH presence pre-Toba and post-Toba, which would date that presence to 75,000 ya. There are a handful of pre-75,000 ya arguably AMHs remains in one location in China, although there are some archaic tendencies.
A South Asian refugium would certainly be plausible. But, it is hard to rule out the possibilities of refugia also in North or East Africa, the Persian Gulf, or Iran. Certinaly the back-migration scenario you suggest is not implausible.
September 29th, 2011 at 1:23 pm
@11 above,
Set the interval to 50 secs, then hit the Start button repeatedly, to simulate a Next button
September 30th, 2011 at 1:24 am
ohwilleke,
Yes, there certainly could have been AMH pockets in the regions you describe. The main problem with that is that all the “brother” and “parent” haplogroups are in S and SE Asia (and don’t forget about C!). IJ and G are “too far down the tree” to be left behind much west of Pakistan. So, such pockets, if they exited, could perhaps been E, instead.
September 30th, 2011 at 2:09 am
“You can go into a small village in Central Germany and pick three men that look like clones. With some luck, one will carry R1a, one R1b, and one I.”
True – because their autosomal genes were thoroughly mixed due to their living in an area where many lineages crossed paths.
On the other hand, there are ethnic groups in Northwestern Russia (e.g. Nenets) who have undeniably Mongoloid appearances. And sure enough, their Y-genotypes can be traced back to East Asia. There is a long-standing controversy which has been going on long before Y-DNA genotyping was even imaginable, which involves scientists debating whether Finns are “mostly Mongoloid”, “mostly of European origin”, or something in between. And it is only within the last 5-10 years that genomics was sufficiently developed to conclude that, Y-DNA wise, Finns are, in fact, on average, about 60% N (East Asian) / 40% I (what I called Semitic above).
@31, the 130kya number is indeed quite curious, as there’s no good reason to think that humans possessed the full spectrum of speech abilities at that time. It could be that some speech-related mutation occurred, say, 70 kya ago, and then its survival advantage allowed its carriers to replace speech-deficient A’s and B’s so totally that there are no pre-mutation populations left, without destroying A and B haplogroups entirely.
September 30th, 2011 at 4:16 am
And it is only within the last 5-10 years that genomics was sufficiently developed to conclude that, Y-DNA wise, Finns are, in fact, on average, about 60% N (East Asian) / 40% I (what I called Semitic above).
Finns are ~60% N on average, but what really matters is autosomal genetics, as only autosomal genetics gives information about overall genetics and thus about racial composition. In autosomal genetic analyses such as ADMIXTURE analyses Finns on average have ~6% Mongoloid components in toto. Lithuanians are ~40 N on average, but in ADMIXTURE analyses they have no Mongoloid components, and thus they have no Mongoloid racial admixture. These show that Y-DNA haplogroup N isn’t exclusive to Mongoloids and Mongoloid-admixed and can be seen in high amounts in 100% Caucasoids (e.g., Lithuanians), just as Y-DNA haplogroup E isn’t exclusive to Negroids and Negroid-admixed and can be seen in high amounts in 100% Caucasoids (e.g., Greeks). BTW, Finns aren’t ~40% but ~30% I on average. Finns are ~60% N, ~30% I and ~10% R (more R1a than R1b) on average. Other Y-DNA haplogroups are in negligible amounts in Finns.
September 30th, 2011 at 4:54 am
“Finns on average have ~6% Mongoloid components in toto.”
How is it logically possible to have 60% Mongoloid Y-DNA but only 6% Mongoloid total genome? Even if you have a purely European original population and 60% Mongoloid all-male invader force that comes, kills all men, and mates with all women, you’d still end up with 30% Mongoloid offspring.
September 30th, 2011 at 5:30 am
Again, y-DNA is a horrible indicator of autosomal DNA. After just 10 generations, autosomal contribution may be only ~0.1% – even if the y-DNA haplogroup survived.
Also, it seems to me that your peculiar picture of eastern introgression does not conform to history, nor genetics, nor archaeology.
September 30th, 2011 at 5:32 am
How is it logically possible to have 60% Mongoloid Y-DNA but only 6% Mongoloid total genome? Even if you have a purely European original population and 60% Mongoloid all-male invader force that comes, kills all men, and mates with all women, you’d still end up with 30% Mongoloid offspring.
Those haplogroup N carriers going to what is now Finland were already admixed with Caucasoids before arriving in what is now Finland (after all, Finland isn’t adjacent to East Asia or Siberia). I’ve given you the more obvious examples of Lithuanians and Greeks. If you are still not convinced by the evidence I have presented to you, you are denying all the evidence, data, i.e, facts, and fabricating your own “facts”.
September 30th, 2011 at 5:35 am
Again, y-DNA is a horrible indicator of autosomal DNA. After just 10 generations, autosomal contribution may be only ~0.1% – even if the y-DNA haplogroup survived.
Also, it seems to me that your peculiar picture of eastern intogression does not conform to history, nor genetics, nor archaeology.
Completely agree.
September 30th, 2011 at 5:42 am
“True – because their autosomal genes were thoroughly mixed due to their living in an area where many lineages crossed paths.”
Not at all. The area south of the Harz mountains has huge continuity for the past ~7,000 years. The same 3,000 year old y-DNA haplotypes of Lichtenstein cave have been found in the same exact area, today.
September 30th, 2011 at 5:44 am
“your peculiar picture of eastern introgression does not conform to history, nor genetics, nor archaeology.”
I’m not quite sure what you mean.
“If you are still not convinced by the evidence I present, you are denying all the evidence”
I’m merely observing that I don’t see a logical way to construct a population with characteristics you describe. You are quoting results from an unspecified study which I can’t find on scholar.google.com, and I see these results as inconsistent with reality. What am I supposed to think?
“The area south of the Harz mountains has huge continuity for the past ~7,000 years”
That does not matter. If you have a village whose population is 33% R1a / 33% R1b / 33% I, that means that, at some point in the past, you have an admixture of contributions from wildly different autosomal lineages going back tens of thousands of years.
September 30th, 2011 at 5:58 am
I’m merely observing that I don’t see a logical way to construct a population with characteristics you describe.
I’ve already explained it:
“Those haplogroup N carriers going to what is now Finland were already admixed with Caucasoids before arriving in what is now Finland (after all, Finland isn’t adjacent to East Asia or Siberia). I’ve given you the more obvious examples of Lithuanians and Greeks.”
You are quoting results from an unspecified study which I can’t find on scholar.google.com
My ADMIXTURE analysis figures are from the Dodecad Project (one of the most comprehensive ADMIXTURE analyses in terms of the number of samples and populations you can ever find):
https://docs.google.com/spreadsheet/ccc?authkey=COCa89AJ&key=0ArAJcY18g2GadDUyeEtjNnBmY09EbnowN3M3UWRyNnc&hl=en_US&authkey=COCa89AJ#gid=0
My Y-DNA haplogroup figures are from Eupedia:
http://www.eupedia.com/europe/european_y-dna_haplogroups.shtml
I see these results as inconsistent with reality. What am I supposed to think?
Are Finns racially at least ~30% Mongoloid, are Lithuanians racially at least ~20% Mongoloid, are Greeks racially at least ~10% Sub-Saharan African (including East Africans)? Is this your “reality”?
September 30th, 2011 at 8:14 pm
Suppose you have a small group of men entering a region, men who for whatever (e.g., technological) reason have a slight advantage of making a living. They all take local wives. To avoid inbreeding, their children and grandchildren also take local wives but don’t intermarry. Because they are rich, they may have more than one wife or may have additional “illegitimate” children. In the third generation, there may be 50 times as many of these men than initially, but their autosomal signature is down to just 12%. It really doesn’t take much time at all for the autosomal portion to shrink, while the frequency of that particular y-DNA in the population grows exponentially.
September 30th, 2011 at 10:35 pm
Suppose you have a small group of men entering a region, men who for whatever (e.g., technological) reason have a slight advantage of making a living. They all take local wives. To avoid inbreeding, their children and grandchildren also take local wives but don’t intermarry. Because they are rich, they may have more than one wife or may have additional “illegitimate” children. In the third generation, there may be 50 times as many of these men than initially, but their autosomal signature is down to just 12%. It really doesn’t take much time at all for the autosomal portion to shrink, while the frequency of that particular y-DNA in the population grows exponentially.
Yes. The process you described and the fact the N carriers were already admixed with Caucasoids until arriving in the regions around the Baltic Sea from so far in the east (you have take into account both the distance and the probably very long time period that passed while traveling that distance) explains why Finns and populations in the Baltic countries have so high amounts of haplogroup N while they have little or no autosomal and thus racial Mongoloid admixture. BTW, N carriers in the Ural region and some parts of Siberia (I am referring to the indigenous populations in those regions) are indeed admixed with Caucasoids today. The full or partial Caucasoidization of some of the haplogroup E clades must have happened also via these two processes.
September 30th, 2011 at 10:51 pm
you have take into account…
you have to take into account…