Or, to be less crude, did modern humans, having already evolved in Africa, interbreed with the local Europeans who were Neanderthals, and if so, did they produce fertile offspring … and, did this happen in sufficient degree to have mattered at all to the genetics of later (but not necessarily living) people?


Blogging on Peer-Reviewed ResearchIn my opinion, the answer is, of course they interbred. There are many reasons to believe this if almost no way to prove it. Indeed, the evidence of this interbreeding is virtually nil. With every additional test of the interbreeding hypothesis using DNA, the null hypothesis of no interbreeding is not falsified. Every morphological hybrid that is put forth is eventually, it seems, retracted.

I can’t believe that they would not have interbred, yet there is not a shred of evidence that their DNA mixed, either in the form of the DNA itself or in overlapping or mixed phenotypes.

The latest study, that I’d like to report on here, is coming out shortly in the Journal of Human Evolution. It carefully looks at remains from Cioclovina Cave in Romania. This would be an excellent candidate, based on its age and location, for evidence of meaningful (as in produced offspring, not “oh, yes, our relationship is very meaningful”) interbreeding between local Neanderthals and incoming Africans or Asian-Africans.

Here is the Cioclovina Calveria:

i-af8c540a159a23c8d9c2910460908e47-Cioclovina.jpg

This is Figure 2 from the original paper, and the caption reads: “The Cioclovina partial cranium. a) Lateral and b) frontal views. c) Close up of the Cioclovina nuchal region. d) Posterior view of La Chapelle-aux-Saints, showing the occipital torus and suprainiac fossa.”

The trick is to compare this skull with the skulls of the indigenous Neanderthals and the incoming Africans (or Asian/Africans) and see if this skull (Cioclovina) fits into one, the other, or is perhaps in between.

One could look at this skull and easily say “This is a modern human” or at least “this is not a Neanderthal.” If that is acceptable as a starting point, then one could modify the above to the following question:

Is Cioclovina a full fledge modern human of the type that existed in those days, or does it have Neanderthal Traits, as though its parentage included Neanderthals?

One can formulate these statements into various hypotheses and null hypotheses, but in my view that is a misguided effort at this level. Hypothesis testing requires a much greater degree of control and understanding of the variables for the negative results of such a test to truly falsify something. What we are really looking at here is an exercise in “alpha-taxonomy” … the primary classification of specimens into basic categories … where we are trying to fit a particular specimen into a taxonomic category that would exist if Neanderthals and modern humans a) “did it”; b) produced offspring; and c) the offspring were intermediate or had a mixture of traits. This is not necessarily how the authors of this paper think of their comparison, but I think it is useful.

Of course, there is not really a “modern human” sample of skulls and a “Neanderthal” sample of skulls to make the comparison. There are some materials that are probably ancestral to Neanderthals (but maybe not), the “classic” Neanderthals, and several different groups of modern humans, including the living ones (like you and me) and various more out-of-date modern humans. Mixing samples together would likely cause the loss of important information, and keeping the samples more separated results in better behaved statistics and prettier graphics (these two results not entirely unrelated).

You will have to read the original paper for all the details, but these are the categories of material used in the analysis:

NEA
Classic Neanderthals from Europe and West Asia. This includes all the usual suspects such as Amud, La Chapelle-aux-Saints, La Ferrassie, Shanidar, and Tabun.
MPE
Middle Pleistocene Europeans … These are early, dating to prior to European modern humans or Neanderthals. this includes P{etralona and Sima de los Huesos 5.
MLP
Middle and Late Pleistocene modern humans from Africa and the Levant. These are the “African morphology” modern humans. and are probably most like the “Modern Humans” but typically, we would expect, with increased robusticity and bone thickness (a general trend in all lineages, even non-primate, through the middle and late Pleistocene). This includes Qafzeh, Saldahna and Skhul.
UPE
Upper Paleolithic Europeans. These are the European skills which are most likely to either look like modern humans or show hybridization, depending on whether or not hybridization occurred. The total number of these remains has been reduced in recent years because of redating of some of the specimens. This sample includes, for instance, Cro Magnon.
Recent Humans
This sample includes a wide range of Africans, Asians, New World, European, Australian material.

The study is very detailed, uses several different statistical techniques, and is very thorough. The conclusion of the study is

Our description and analysis rejected all four tests of the hypothesis that Cioclovina represents a Neanderthal-early modern human hybrid. The specimen does not exhibit any of the morphological patterns predicted for a hybrid. … it can be accommodated quite plainly within the variability of the Upper Paleolithic sample. Cioclovina is a fully modern human in all respects examined and analyzed and there is no hint in its anatomy that it may be a Neanderthal-early modern human hybrid.

The details are myriad and beyond the scope of a mere blog post such as this one. This paper would make an excellent exemplar for research in Physical Anthropology for a methods course. Here, I can only touch on a couple of the key results that relate to this conclusion.

The following shows the relationship between the shape of the skull based on all the available data for a “modern human-like” (blue) and a “Neanderthal-like” (red) template.

i-5541a4b338a46454a8b2fe5a70343a1b-Fig_6_ShapeDifferencesAlongPC1.jpeg

This is based on the first principal component of an extensive and intensive morphological study comparing these taxa. Print this out and bring it with you on your next field trip just in case you find a skull … you’ll be able to distinguish between Neanderthals and Modern Humans.

In all the comparisons and analyses presented in th is paper, the Cioclovina skull groups with the Modern Humans and not the Neanderthals. Within the modern humans, it groups especially among the Africans and somewhat less so among he New World and Asian material. It also does not group with the “MLP” set particularly well.

One of the better representations of this finding is seen in the following figure, which is a graphical representation of a set of Mahalanobis Squared Distances … a measure of dissimilarity among metrics, essentially … applied to the samples.

i-04a32241fa12590c5669def5a19a9f50-MSD_Data.jpg

This shows Cioclovina among the moderns, very distant from the Neanderthals and their presumed ancestors. The MLPs are themselves intermediate between the the group that includes the UPEs, Recent humans, and Cioclovina.

Note that the MLP group includes individuals that in the past have been suggested as possible “hybrids” or at least intermediates. These specimens rest in both time and space between Classic Neanderthals and the earlier African Moderns. If Neanderthals and moderns of 40 thousand years ago shared a common ancestor just under a half million years in age (which they probably did) and had very little intermixing of genes during the previous 100 thousand years (i.e., since the rise of Neanderthals as a distinct morphology) then we would expect the MLP group to be where it is on this chart or in similar data sets.

In the end, Romania is a romantic place. This is obvious. Romance languages, Gypsy violinists playing the background music, Mysterious mountains and ancient, beautiful cities. But either this did not apply several tens of thousands of years ago, or sex among the Neanderthals and Moderns was not productive, in and near the vicinity of Cioclovina Cave. Or, most of the Neanderthal alleles that result in distinct phenotypes were recessive. (Which we know in the business as “Wolpoff’s Worst Nightmare…”)


HARVATI, K., GUNZ, P. & GRIGORESCU, D. (2007): Cioclovina (Romania): Affinities of an early modern European.. J Hum Evol, Advanced Online, .




Comments

  1. #1 Martin R
    November 18, 2007

    “… of course they interbred. … no way to prove it. Indeed, the evidence of this interbreeding is virtually nil. With every additional test of the interbreeding hypothesis using DNA, the null hypothesis of no interbreeding is not falsified. Every morphological hybrid that is put forth is eventually, it seems, retracted.”

    I think you were aiming for an odd number of negations there.

  2. #2 RPM
    November 18, 2007

    I can’t believe that they would not have interbred, yet there is not a shred of evidence that their DNA mixed, either in the form of the DNA itself or in overlapping or mixed phenotypes.

    What about the evidence for introgression of archaic alleles into the modern lineage? Granted, we can’t know for sure that these archaic alleles came from neanderthals, but there is genetic evidence for hybridization between moderns and at least one archaic lineage. Also, there is the issue of our power to detect introgression from an archaic lineage into the modern lineage. John Hawks wrote a post on this (and a paper too, I believe) a short while back. Basically, it’s unlikely we’ll be able to detect introgression from genetic data unless it’s accompanied by a selective sweep following the introduction of the new allele (ie, adaptive introgression).

  3. #3 Anne Gilbert
    November 18, 2007

    The question of “did they or didn’t they” is much more complicated than it would appear to be. First of all, a note on the studies themselves: all the writes of these papers, no matter what “side” they come down on regarding this issues, have a priori assumptions(perhaps largely unconscious), that tend to affect the questions they ask, the methods they use, or the “framework” of the research. As you probably are aware, you can get different results from measuring the same fossil material, depending on what methods of measurement you use and the questions you ask. And note well: I’m not letting any “side” or paleoanthropological/archaeological faction off the hook here. So these studies mean b>something, but exactly what they mean is as yet unclear.

    Second, it’s my understanding that the Cioclovina fossil is a “modern” human, with purported Neandertal traits, not an actual Neandertal, though I may have misread the studies on this. So if Neandertals and “moderns” interbred in this particular part of their overlapping range, it wold be quite natural for some early “moderns” to have “Neandertal traits”. After all, there doesn’t seem to have been anything to prevent such interbreeding, as far as their respective behaviors were concerned; at least that’s what such archaeological evidence as exists, tells me. Culturally, it’s another question, but at this time period, whether we’re talking about early “moderns” or late Neandertals, there’s no way of telling.

    Finally, given the (probably) low numbers of both populations, in and around Cioclovina and everywhere else in Western Eurasia at the time — and Neandertals may have been a very small population indeed —it isn’t too surprising that contemporary “modern” populations don’t seem to have any “Neandertal genes”(what is a “Neandertal gene”, anyway?). As a population, “modrn H.sapiens is 6 billion strong and counting. . . the mahhematics from 30,000 years ago until now, just don’t favor that kind of surivival. But again, that’s another story.
    Anne Gilbert

  4. #4 Greg Laden
    November 18, 2007

    Martin: One could never say that you were anything other than not at all in correct.

    RPM: Maybe, but since we are talking about very spotty data derived from individuals all of whom are members of the same species, most likely, this is not too impressive. The divergence between N’s and amHS is much MORE impressive.

  5. #5 Azkyroth
    November 19, 2007

    All I have to say that if Neanderthals and modern humans never interbred, certain of my former classmates demand an alternate explanation…

  6. #6 razib
    November 19, 2007


    RPM: Maybe, but since we are talking about very spotty data derived from individuals all of whom are members of the same species, most likely, this is not too impressive. The divergence between N’s and amHS is much MORE impressive.

    could you unpack this for me? i mean, isn’t the adaptive introgression inference based on very long and deep branches which are pushed back beyond the neandertal-human separtion or around thereabouts?

  7. #7 greg laden
    November 19, 2007

    Razib: Adaptive introgression (or just introgression, really) is the transfer of alleles from one population to another where there is usually separation. In mammals this is expected to happen in relatively closely related species.

    Assume Neanderthals and contemporary moderns have a divergence of about 0.5 mya. Now, you look at a segment of the genome where you suspect there is an “introgressive” allele. You somehow statistically figure out that there is an allele that is shared with a very low probability … in other words, this gene has alleles with a coalescence point that is subsequent to 0.5 mya (so the allele likely arose in one lineage) and the base pairs match exactly (so it looks like it “transferred).

    There are three explanations that have to be considered:

    1) You are not really seeing anything, it is the product of chance. The info provided above suggests that this is unlikely but it is possible.

    2) The allele is similar because of common ancestry; and

    3) The allele “introgressed” (by whatever mechanism that is supposed to happen).

    If the difference between the coalesence for diversity in this allele in the donor population is recent, and the split between the two populations is old, then 2 is very unlikely and 3 is more likely. As the coalesence for the alleles for this gene gets older and/or the split between the lineages gets younger, 2 becomes more likely and 3 becomes less likely.

    Also, for any divergence point, you’ve got to figure on biogeographical grounds alone there is some fuzziness, but the fuzziness remains stable as the divergence extends back into the past. In other words, there is a fixed period of time during which there is a lot of gene flow, then once that “stops” and time continues marching on, that becomes a smaller and smaller percentage of the total divergence time and becomes less important.

    With a divergence time of 0.5 mya, and a lot of biogeographical overlap, I would assume that this period of increased gene flow is a large percentage of that time (maybe 10 or 20 percent?).

    Therefore, the choice between options 2 and 3 is pretty unclear. That’s what I meant.

  8. #8 greg laden
    November 19, 2007

    By the way, here’s John’s Paper:

    http://www.paleoanthro.org/journal/content/PA20060101.pdf

  9. #9 cuchulainn
    November 19, 2007

    right, presumably africans didn’t interbreed with them. does that make the races different species?

  10. #10 Warren
    November 19, 2007

    Consider the number of people who continue, today, to “do it” with nonhuman animals. Did our ancestors do the same with Neanderthals? Heh, well.

  11. #11 Fred
    November 19, 2007

    I’m no expert, but being a modern homo sapien, I find Neanderthals butt-ugly. I suspect my forbearers did as well.

  12. #12 Justin Moretti
    November 19, 2007

    Yes, but imagine the practical attractions of a partner with whom sex can be guaranteed never to produce any viable offspring. And as for the butt-ugly, I’m sure kinks existed even back then.

Current ye@r *