That's the least I can say after looking at the matter with some attention: that the claims made by a recent paper are not sufficiently justified to say the least.
Katerina Harvati et al., The Later Stone Age Calvaria from Iwo Eleru, Nigeria: Morphology and Chronology. PLoS ONE 2011. Open access.
They tell us of a recent pair of skulls from SW Nigeria. They were discovered in 1965 and at least one of them belongs to a whole skeleton, including a critical piece: the mandible. Sadly this paper totally ignores anything but the calvaria (the top of the skull). They have been dated by uranium series to c. 12-6 Ka ago.
Sloppiness seems too common in this paper: not just the two skulls and skeletons are not properly shown in full but, in fig. 1, they claim to be showing both calvaria, they only show one from four angles:
Then we are thrown into PC analysis, which seems to be their main working tool:
Of course PC analysis is limited in its ability to discern and can only provide a very rough view. In any case, here we have PC2/CV2 occupied by intra-modern variability, while only PC1/CV1 show anything that resembles modern-archaic differences: a totally unilinear analysis in the end.
As we know, sample size matters; but instead of balancing it by reducing the number of modern samples (grey dots), the authors have totally ignored this matter.
So we get an (unclearly legitimate) linear archaic-modern horizontal dimension in which some individuals typically considered Homo sapiens of archaic traits (magenta and black triangles) tend to occupy intermediate positions between the erectus-neanderthal-heidelbergensis conglomerate (left) and the modern human oversized sample (right). Iwo Eleru (IE, black cross) is within this "ambiguous" group and resembles other archaic H. sapiens from the Middle Paleolithic (magenta triangles), as well as similarly aged Upper Cave 101, from North China.
UC 101 (UC1, pictured left) is a great counter-example because with Jebel Irhoud or the Palestinian skulls, there's always someone claiming archaic admixture, typically with Neanderthals. But Upper Cave individuals are invariably described as modern in spite of their low vaults, so similar to those of "archaic" Homo sp. A good reason is that they have clear, unmistakably modern, chins.
To me, the calvarium looks very similar in shape to Iwo Eleru. To be safe I compared them both using fig. 4-A as tool:
|Inner drawing: Iwo Eleru, outer drawing: modern mean|
The comparison may be a bit imperfect but it is clear that the shape of both calvaria, of similar age, one from West Africa and the other from East Asia, are very much alike - and both are somewhat different from the modern mean.
I think therefore that at least serious doubt must be casted on the conclusions that some want to extract from this paper: that there was some sort of archaic hominin in West Africa until recently. Nothing in this data suggests this, instead it does suggest that a better methodology is much needed.
Your photographic comparison of Iwo Eleru with Upper Cave notwithstanding, the actual data (Table S2) indicates that Upper Cave is well within the range of modern human variation, exhibiting distances between 7.48 with Iberomaurusian and 20.83 with Near East. Upper Cave is thus about 10 times closer to Iberomaurusians than they are to Iwo Eleru (73.27). Indeed, even the Neandertal group is closer to UC (44.15) than IE is.ReplyDelete
Iwo Eleru do not exhibit any close distances to any other human groups; their closest distance is to Upper Cave (73.27) which is, however, substantially greater than even that between Eskimos and Khoisan (20.92), the two most divergent modern groups.
Claiming that IE is not archaic would inflate the range of modern human variation multiple times, and would, by necessity include almost every Homo skull of the last several hundred thousand years into H. sapiens
Well, these are all statistical approxximation (I mean: not the raw data but something else) and I'm a bit unsure about how good or limited each of these means are.ReplyDelete
Anyhow, using Mahalanobis D² the closest thing to UC1 is EAM (Qahfez 6 y 9 and Skhul 5), with a negative D² value, and then European Upper Paleolithic with a D² value of 4.77. In comparison, the 73.2 D² value in relation with IE seems very high. But this one is also the lowest value that IE achieves in relation to any other skull or group of skulls.
"Indeed, even the Neandertal group is closer to UC (44.15) than IE is".
Doesn't it render this D² measure useless? After all it's a gauge of other measures, as good as any other simplification.
I personally prefer to put one skull on the other, using transparencies, and gauge their similitude or difference visually. It seems better considering that we do not know how the authors have gauged D²: the results are posted but not the raw data nor the whole process, so we basically have to trust (or distrust) their hidden methodology.
I choose to not trust.
"Claiming that IE is not archaic would inflate the range of modern human variation multiple times, and would, by necessity include almost every Homo skull of the last several hundred thousand years into H. sapiens".ReplyDelete
Yes, this is a potential issue and that is why I'd like to see the face and jaw and not just the calvarium.
Let's not forget (and sorry for the multiple posting) that Lagar Velho kid had an issue with a deformed calvarium (deformation caused by the burial itself). We'd benefit therefore from a wider analysis: in LV (as in UC1) everything else yelled H. sapiens, but in this Nigerian case we lack the relevant data, which is hidden from us by the authors of the paper in a way that I deem highly suspicious.ReplyDelete
Doesn't it render this D² measure useless? After all it's a gauge of other measures, as good as any other simplification.ReplyDelete
I find it strange that you prefer your non-expert eyeball scan of a poorly made picture superposition of the side of the skull, to a standard statistical distance measure applied to state of the art digital 3D scans of the fossils made by world-renowned experts.
I've always preferred geometry to algebra, specially when the facts of the algebra are not even clearly specified.ReplyDelete
Obviously visual geometry offers a very immediate intuitive (yet correct) approach to the matter. You may want to point out if something is missing from this approach but you cannot deny it is valid.
It is also valid from the viewpoint of PC/CV analysis, which renders the same results, always placing UC1 in the intermediate zone between "modern" and "archaic" skulls.
Why D² produces a somewhat different result? We would need to know which basic measures were taken and how and if were applied to the Mahalanobis equations. Only then we could discern why these differences.
Why does UC1 cluster with EUP (after the Palestine skulls) in D² values but not in the PC/CV graphs? Why does that even if it's obvious that the vault is much lower than EUP ones? I do not know nor I can't say unless I get the data and process producing table S2.
All the rest of the data is consistent in classifying UC1 as intermediate with the "arguable" H. sapiens heads marked as purple triangles (that are anyhow generally considered H. sapiens even if of archaic or "less evolved" features).
"and would, by necessity include almost every Homo skull of the last several hundred thousand years into H. sapiens"ReplyDelete
That may not be an unreasonable conclusion though.
"I think therefore that at least serious doubt must be casted on the conclusions that some want to extract from this paper: that there was some sort of archaic hominin in West Africa until recently".
Perhaps so. But it is very interesting that after our long disagreement as to how desirable tropical rainforest was as a habitat for humans we now have apparently 'archaic' humans surviving until comparatively recently in a region described as being in the rainforest zone:
"As we know, sample size matters; but instead of balancing it by reducing the number of modern samples (grey dots), the authors have totally ignored this matter."ReplyDelete
No - it is justified a posteriori. Yes, the PC analysis is basically derived from modern humans, due to sample size - but then this is justified by the fact that PC1 neatly separates heidelbergensis/Neanderthal from modern humans (and archaic Europeans!), leaving other archaics in between.
I am more intrigued by what this tells us about the origin of spread in modern populations...
An oversized sample is never justified, except maybe when it's the center of your research. In this case the center of the research is a single dot, so it'd be best compared with much smaller samples.ReplyDelete
I'm a bit unsure but I suspect that a much smaller modern sample would have emphasized the differences between archaics and let the "intermediate" skulls fall to the modern cluster.
Instead here we get an artificial modern vs. archaic gap that is rather unreal (for example Neanderthals should not cluster with H. erectus, etc.) and fails to properly describe the diversity at play.
Yes, I could imagine in particular PC2 to be actually meaningful with a smaller modern sample.ReplyDelete
this is www.proto-germanic.com. I spend three hours writing to a comment and as usual, I cannot even make comments on my own blog. I tried your email, it does not work. Ken email@example.comReplyDelete
My email works but you have to remove the capitalized 'DELETETHIS' anti-spam protection first. ;)ReplyDelete
As I said, I know nothing about Basques, but where I grew up in California, I knew some Basques, I guy (surnames) Biscayne, Ayerza,..
I see you are a Boasnian. I completely reject Franz Boas and all his, self-admitted communist or leftist sympathizers.
I am a right-leaning libertarian, I am wildly, insanely anti-statist and even more virulently anti-communist/marxist/leninist. Lenin and Stalin alone in the USSR murders over 100 million, innocent, men, women and children, civilians and we NEVER hear about them. Apologist for communism say that Stalin, Mao, Che, Castor, Il, Pol Pot were abberations...NO, they follow the communist manifesto to the letter. Marx and Engles called for, demanded the frequent, arbitrary mass-killings of innocents. This is in on early Ashkenazic-conceived, written, evangelized, supported. It's something that modern Ashkenazim like to hide, that they personally killed about 20 million Russian Christians, they killed Sephardic Jews, they hated Hebrew, they killed their own Ashkenzi brothers. Of course there were gentiles. But the percentage of Jewwish authorship of communist writings, leaders of leftists organizations, communist regimes, except for east Asia were, in top party leadership, NKVD, KGB, Chekist, etc, were about 90% Ashkenazic. Now, I love Gilad Atzmon, he is or was an Israeli, and he is openly against Zionism/Marxism. Communism under Mao, Stalin, Lenin, Pol Pot has killed about 200 million civilians, executed, starved, burned, tortured people. Hitler was horrible, but NOTHING compared to the amount of people he killed and the length he was in power. Even if Hitler had not attacked Poland in 39, the Soviets were planning attacks.
I believe in all races/ethnic group, language group, if members of any given race/group choose to be separatist, I am ALL for that. If people want to mix let them. I am against enforced diversity. I believe the Basque and Palestinians should have completely independent states/nations.
In a message dated 9/24/2011 12:22:49 A.M. Pacific Daylight Time, lialdamizDELETETHIS@gmail.com writes:
Maju has left a new comment on your post "What is the "urheimat" of light hair and eyes?":
So you took a post from a forum and reposted it here without even making that clear. I see.
I'm not going to go that forum (which I helped found and abandoned in few days, after the smartass self-appointed leader, some Portuguese guy whose name I prefer not to remember, put me to "patrol" a subforum with an ultrazionist fascist, knowingly we had problems before).
I just thought you'd be interested in knowing that it's junk.
"Maju, I see you are Basque and a Libertarian...?"
Yes. And I do not mind speaking English (which is a way to snub the Spanish invaders, mind you).
"algún pensamiento debido a su proximidad a África del Norte y los dialectos bereberes que vasco podría estar relacionado con la rama camitas de los afro-asiático"
Not in principle. It is possible that languages related to Basque have been spoken at some time in North Africa, leaving maybe some substrate influence in Berber, but there's zero relation between Afroasiatic and Basque. IMO the closest thing to Basque (after Iberian and other possible dead Vasconic languages like Ligurian or Pictish) could be Indoeuropean - but it's a very remote relation.
I don't think there's much to say about "races", much less re. subtle distinctions in ultra-homogeneous Europe. Just that Nordic Europeans (and by this I mean people from North-NW Europe, regardless of individual phenotype) have accumulated a greater concentration of blondisms in their gene pool, surely as part of the process of adaption to the extreme latitudes where they live (thanks to the warming by the Gulf Stream). If they'd be less blond, they'd have greater risk of neuronal and physical damage to their children (specially if they did not eat lots of fish).
"they follow the communist manifesto to the letter. Marx and Engles called for, demanded the frequent, arbitrary mass-killings of innocents."ReplyDelete
I have no idea what is going on in here, but so much historic ignorance makes me cringe. :(
Kenneth: rcved and get lost.ReplyDelete
"I have no idea what is going on in here"...ReplyDelete
Loose Freak warning. My fault for making a couple of comments at HIS blog.
This comment has been removed by a blog administrator.ReplyDelete
"NO, they follow the communist manifesto to the letter. Marx and Engles called for, demanded the frequent, arbitrary mass-killings of innocents".ReplyDelete
Isn't that exactly what the Old Testament of the Bible demands of its true followers also?
This comment has been removed by a blog administrator.ReplyDelete
Don't feed the troll, please.ReplyDelete
What about these European skulls:ReplyDelete
I think that they refer to the Neanderthals of Spy (aka Canstadt).Delete
"Canstadt type" really means neanderthal. But the skulls depicted aren't neanderthals, but very recent humans, and are being considered "atavistic"/approaching neanderthals by the author (that wasn't uncommon at that time, it seems).ReplyDelete
Regarding sample size, a quote from a related article that may be pertinent: "We would like to point out that a smaller modernReplyDelete
human sample is actually biased against our finding that almost all early modern human fossils connect to a recent human. Increasing the modern human sample would only increase the chance that a fossil is close to a modern human in shape space."
This is from "Early modern human diversity suggests subdivided population structure and a complex out-of-Africa scenario".
That text is from the time of Piltdow man or even older maybe. I can't judge that unless you can provide more accurate data, sorry. I never heard before now of Homo sapiens in Spy.ReplyDelete
Regarding what you say of "sample size", that is only true for as much as the archaic skulls are modern, because Homo sapiens skulls are very different in almost everything from other Homo sp. Our species is the least conservative in cranial traits of all the Homo genus. Also modern humans tend to cluster with each other in these PC analysis, so really, no.