JAC: Greg sent me some thoughts on the new hominin finds involving what used to be known as Denisovans. I suggested he put up his own post on these thoughts, and he kindly agreed. The result is below.
Greg has asked me to weigh in at the end, and I will, under the JAC initials. Except for that, credit for this post goes to Greg.
by Greg Mayer
In a recent pair of papers, Qiaomei Fu and colleagues show, via both DNA and protein analyses, that the skull known as “Dragon Man” from Harbin, China, is a Denisovan. The Denisovans, first identified in 2010 by DNA from a single finger bone from a cave in Siberia, are a long-separated (several 100 kya) lineage of humans that has interbred with anatomically modern humans and Neanderthals, leaving a fairly strong signature (ca. 6%) in the ancestry of modern humans from East Asia and the Pacific. Although several more Denisovans were identified by DNA after the first one, all were from very incomplete remains– until now.
The Harbin skull was first made known to science in 2018, and was named Homo longi in 2021. I have been able to access only the DNA paper, not the protein one, but Carl Zimmer gives a fine summary in the NY Times.
Zimmer notes that some scientists, including Chris Stringer of the British Museum (Natural History), are now using the name Homo longi, based on the Harbin skull, as the name for the Denisovans. John Hawks of the University of Wisconsin, pushed back on this, telling Zimmer “I’m pretty confident saying these are all Homo sapiens.”
Jerry, Matthew, and myself have discussed Neanderthals (or Neandertals) and Denisovans several times here at WEIT.
About 10 years ago, Jerry, I, and Hawks all agreed on all three forms (Neanderthal, Denisovan, anatomically moderns) being members of one interbreeding species. But when I read David Reich’s book and other papers about five years ago, I was impressed by the paucity of Neanderthal DNA on the X chromosome of moderns, suggesting that Haldane’s Rule had kicked in, and there was partial intrinsic isolation (probably sterility) of hybrids between Neanderthals and moderns. (Haldane’s Rule states that if there is partial reproductive isolation due to infertility or sterility of hybrids, it is the heterogametic sex—XY males in Homo— that is most affected.) So, my view now is that a few hundred thousand years of isolation had moved the Neanderthals and moderns measurably down the road to speciation, though obviously not completely.
My earlier view had been that the low percentage of Neanderthal DNA in moderns was due to social/historical factors, analogous to how there is little European/American Indian admixture in North America, even though there is no intrinsic isolation between these two groups. There’s probably still some social/historical factors involved in the Neanderthal/modern encounter, but part of the disparity in genetic contribution is due to some infertility in male hybrids. [JAC: This infertility is suggested by the observation that introgression of Neanderthal DNA into modern humans is less pronounced on the X chromosome than on the other chromosomes, or autosomes.]
As far as nomenclature goes, I’d be happy with them all being sapiens, but we should recognize that the differences between Neanderthals/Denisovans and moderns are greater, and of a different sort, than differences between the obvious geographic races that exist within modern humans today.
Hawks has a piece on his website, “The humanity of a new Denisovan“, in which he expands on his comments to Zimmer and goes much deeper into the issues of the history of Homo in the last million years; I heartily recommend it. His fourth figure (the second phylogeny) in particular provides a nice summary of his view of that history, including various episodes of interbreeding. He does not mention the distribution of Neanderthal DNA within the moderns’ genome (i.e., the paucity of it on the X), so I’m not sure if this has affected his thinking in any way. I’m hoping that Jerry will weigh in with his current thoughts on the significance, if any, of the distribution of Neanderthal DNA across the moderns’ genome.
JAC: I’m in agreement with Greg (and Matthew, who I hope will also weigh in here) that “modern” sapiens, Denisovans, and Neanderthals should be considered members of the same biological species, Homo sapiens. I don’t know how big the disparity between introgression of X-linked vs. autosomal DNA is with respect to modern human genomes, so I cannot judge if it is substantive evidence for some sterility (or inviability) of the hybrids between modern sapiens and Neanderthals. Since the sterility would have to be strong to even consider these two as different biological species, and because there is a sizable aliquot of Neanderthal DNA in many modern humans, the sterility could not have been nearly complete. (I don’t know if Denisovan DNA in modern humans show the same disparity between that from the X chromosome versus the autosomes. If it doesn’t show that, there’s no reason to split Denisovans off as a new species.).
At any rate, the biological species concept (BSC) regards two populations as members of the same species if they can hybridize when they encounter each other in nature and some of the hybrids are viable and fertile. To the extent that Neanderthals meet that requirement, I would say that they, too, are members of H. sapiens. This kind of “splitting” of groups into different named species is pervasive in human paleobiology, as it is in some other groups, like giraffes, but to me rests on shaky grounds. People like to split human groups into new species, for you get a lot more attention if you can say you found a new species than if you say you simply found a new subspecies or population.
Matthew: The identification of the extraordinary ‘Dragon man’ skull with the Denisovan lineage is probably what most people expected, but is nonetheless an astonishing development, using protein profiles to establish evolutionary relationships (first suggested by Francis Crick in his 1957 ‘central dogma’ lecture, but he never imagined this could be done on a sample over 140,000 years old!). The issue is what we call the Denisovans – the skull was described as Homo longi, and paleoanthropologists, such as the Natural History Museum’s Chris Stringer, accept this, just as they accept Homo neanderthalensis for the Neanderthals. I agree with Jerry that from a biological point of view, these are all members of the same species, because of the existence of fertile hybrids (us!).
This was Chris’s reply to me on Bsky:
This is a link to the article he refers to.
There are two points that make me pause. First, the taxonomic powers that be are unlikely to change their collective minds on this (they haven’t shifted over H. neanderthalensis in the last decade or more, since the first Neanderthal genome was sequenced and the existence of introgression was established. It’s mildly irritating, but that’s the way it will be. Secondly, the points raised by Greg are interesting – we were separated from these lineages for hundreds of thousands of years, and during this time some striking morphological adaptations occurred. They clearly did not lead to the establish of biological species, although we have little grasp of morphological or genetic variation in these groups (all of which were composed of very small (tens of thousands at most), dispersed populations). The role of the X chromosome, as highlighted by Greg, suggests that all three lineages may have been on the way to speciation, although that process was stopped by the demise of the Denisovans and Neanderthals (whether it would have continued in the presence of introgression is another thing to consider).
Finally, our understanding of human evolutionary history is perpetually changing. Stephen Jay Gould used to rewrite his slides each year; I did the same, pretty much, until I found this on Twitter, back when it was good, and now use it in all my human evolution lectures (along with more serious stuff).
Fu, Q., P. Cao, Q. Dai, E. A. Bennett, X. Feng, M. A. Yang, W. Ping, S. Pääbo, and Q. Ji. 2025. Denisovan mitochondrial DNA from dental calculus of the >146,000-year-old Harbin cranium. Cell 188:1-8. pdf
Fu, Q., F. Bai, H. Rao, S. Chen, Y. Ji, E. A. Bennett, F. Liu, and Q. Ji. 2025. The proteome of the late Middle Pleistocene Harbin individual. Science in press. locked
Reich, D. 2018. Who We Are and How We Got Here: Ancient DNA and the New Science of the Human Past. Pantheon, New York
Very interesting. Thanks, Matthew and Jerry!
Hawks writes in the paper linked above, “Based on sequence divergence of introgressed genome segments, the Denisovans might have encountered, coexisted, and mixed with late-surviving groups of Homo erectus.”
And if this is the case, then according to the biological species concept, Homo erectus and Homo sapiens are a single species as well, no?
Indeed, as I continue with Hawks’ paper, he makes this explicit. Hmm….
Yes. There has been a lot of “splitting” in studies of human evolution. Everyone wants the privilege of naming a new species. It’s understandable. The use of DNA to adjudicate matters is a powerful new development.
Very very interesting. Thank you.
Chris Stringer argues here https://www.nhm.ac.uk/discover/are-neanderthals-same-species-as-us.html that we should think of Neanderthals and anatomically modern humans (AMH) as being different species. (He has recently argued that Denisovans should also be assigned their own species too – I just can’t find the reference). He suggests that the significant physical differences between these taxa implies that we should retain the status quo, because the biological species concept has “limitations” in cases like this. Evolutionary gradualism means that, unless very large periods of time have elapsed, separating species can be very difficult – scientists, effectively, are having to decide between different shades of grey. In one sense it would be a good thing, if all these “species” were assigned to Sapiens as it would help the public to understand that, in all the ways that are important, these were people (although perhaps people in a different way to the way we are people). On the other hand the very strong morphological divergence suggests we keep the existing classification. On balance I think I agree with Stringer, and we keep Neanderthals as H. neanderthalensis and, if I’ve got this right, in keeping with the rules of nomenclature, Denisovans should be redesignated H. longi (at least on the assumption that the recent data proves correct and “Dragon Man” is confirmed as a Denisovan). Anyone think biology is easy?
There are different species concepts (biological s.c., morphological s.c., etc.), and the practice is to use whatever concept works on a case by case basis. Extinct species are usually identified based on the morphological s.c. since we don’t usually know about gene flow in them. But when we have information about gene flow among sexual species, even when known only from fossils, it seems, then the biological species concept seems to take precedent.
What is interesting about the whole broader issue is that in closely related species or sub-species, different species concepts can give conflicting results. Modern humans/neanderthals/Denisovans can appear to be 3 species from a morphological s.c., but the b.s.c. says they are all one species.
There are many other examples like this. Wolves and Coyotes give contradictory results depending on what species concept you use, for example, although by tradition and maybe stubbornness, I believe they are still named as separate species even though the b.s.c. says they are one species.
Are you sure the biological species concept conflicts with the conclusion from morphology. in this case? The amouint of interbrfeeding matters. If hybrids are very rare, the two forms can still evolve in different directions
As I joked below, I would have nightmares if my wife has a brow ridge big enough to knock me out. And the Denisovians may have thought that our flat foreheads were hideous. That might be enough of a barrier to mating that an application of the BSC would still count them as separate species. Though as Peter Fisher says above, it is hard to say how much hybridization is too much.
In population genetics it is often argued that one interbreeding event per generation is the critical point (if there is no hybrid infertility). I have argued that there are conceptual problems with this, and I suggest a different criterion, but in any case, maybe this issue can be eventually settled based on genetic data from lots of individuals.
I don’t know, and had not considered frequencies of hybridization. But isn’t that complicated here since Neanderthals and Denisovans were markedly less common than modern humans?
You are right, it depends a lot on population sizes.
Yes, thanks, I agree.
I was obviously aware of the different species concepts, which I suppose underlines my central point about the difficulties of resolving questions like this. There is obviously a judgement call for scientists to make as to which sc is deemed appropriate in this case. I am not a scientist so my two-pence worth must be seen very much as no more than a mere personal opinion. Chris has now entered the conversation and provided to link to his paper anyway. My feeling is that if these hominins have lived, largely, separately from each other, perhaps even occupying different niches – very roughly, Neanderthals cold adapted, Denisovans altitude specialists and AMH plains occupiers – for hundreds of thousands of years, then it would seem appropriate, to me, to adopt Chris’s standpoint. Others are quite entitled to disagree of course.
Sorry, I’m not a biologist. What are s.c. and b.s.c.?
Species concept and biological species concept.
These circumstances seem a perfect fit for using the additional subcategory, subspecies. As a non-expert I’ve always wondered why subspecies have not been widely adopted.
If it is decided that Denisovans, AMH and Neanderthals are all sapiens, then they will be redefined as subspecies such as Homo sapiens neanderthalensis.
Yes, very interesting. What a brow ridge on that skull! I don’t think I would marry someone that looked like that.
In orchids, many different species and even different genera can be freely cross-pollinated without loss of fertility. What keeps them separate in nature are mating barriers, like different fragrances that attract different pollinators. That brow ridge, or the lack of one, and other traits that we can’t see, might well be a deal-breaker for all but an occasional tryst between these two forms…
Yes, but the mating barriers are also forms of reproductive isolation in nature. Whether they can be cross pollinated and produce fertile hybrids under non-natural conditions does not make them the same species according to the BSC. The “under natural conditions” part is important; otherwise lions and tigers, which used to be sympatric but did not hybridize, would be considered the same species because they can produce somewhat fertile offspring (“ligers” and “tiglons”) in zoos
Yes, definitely, but that was my point. That brow ridge and other possible behavioral and morphological differences may keep them from hybridizing much, even if the offspring are fertile, and so these could be different biological species even if they occasionally interbred with our modern forms, right?
This is way outside my competence, but I’m confused by this interchange. Doesn’t the current evidence suggest that there was significant successful interbreeding between Denisovans and modern humans? Enough that up to 6% of Denisovan DNA persists in some modern human groups? And surely it happened “in the wild”? Doesn’t that meet the criteria Jerry described up above for being considered the same species per the BSC?
I frankly don’t know how much hybridization is enough to count as “breaking” the biological species concept. That’s a gray area. If hominid populations were very small, even a single hybridization event could, by chance, leave a fingerprint on the lineage. Also, as Mark Sturtevant pointed out, the relative abundances of the two forms might affect our interpretation. If Denisovians were very rare compared to modern sapiens, then the presence of the rare form’s genes in the common form’s genome suggests a lower pre-mating barrier, while if Denisovians were more common than modern sapiens, and yet only left 6% of their genes in modern sapiens, that suggests there must have been a stronger pre-mating barrier, and that they never freely interbred. Some formulations of the biological species concept require that the populations of a species freely interbreed in the wild where they are in contact.
Many formulations of the biological species concept are too absolute. Both Jerry’s and my points were that the mere existence of occasional fertile hybrids does not necessarily mean that two populations are conspecific. Evolution is a stochastic process and there are often no clear lines that divide categories.
Thank you, Lou.
😂 🤣
And Lou Jost wins the Internet today!
Thank you thank you jejej
I want to extend my most sincere thanks to Greg for the conversation and links, and to Jerry for hosting this. I do have an essay from a couple of years ago that looks at the X chromosome evidence closely, with my thoughts at that time.
https://www.johnhawks.net/p/positive-selection-x-chromosomes-recent-human-evolution
To make a long story short, the X chromosome within modern humans has a much more widespread pattern of selection affecting its variation than other chromosomes. Meanwhile the “genomic deserts” of Neanderthal introgression, much more marked on the X chromosome than others, may be compatible with higher drift on the X due to lower effective size compared to the autosomes. I am willing to be persuaded that Haldane’s Rule might have been at play once we have more evidence from ancient genomes closer to the time of population mixture. But for the time being I think that other explanations for the X chromosome pattern might do the job.
Hello! Big fan here.
Yeah, I thought about drift as well, but I don’t know enough about the X vs. autosome pattern of differentiation (or about models of drift in these species) to make any decision. If it’s purely drift that accounts for the pattern, then the X vs. autosome difference gives no evidence for intrinsic reproductive isolation.
Thanks. I did a little modeling and the autosomal “deserts” are not surprising under drift in the population sizes that we think were characteristic of late Neanderthals/early moderns. It’s the difference between X and autosomes that is interesting, and that difference carries across several primate divergences also. So I think it may be compatible with drift with a bit of excess positive selection/hitchhiking but I’m sure more data will help to clarify!
What an astounding feat to isolate DNA and protein from remains that are over 140,000 years old, I continue to be amazed.
I still prefer distinct species names, given that the level of morphological distance is far greater than we find within our species today, and matches that between closely related species in other mammal groups. But it’s not a simple issue, as we explain here:
Homo sapiens, Neanderthals and speciation complexity in palaeoanthropology https://academic.oup.com/evolinnean/article/3/1/kzae033/7900502
So great to know for a fact.
I believe I have some understanding of Haldane’s rule that it is the XY sex in mammal hybrids that has reduced fitness, but I don’t understand why that would restrict passage of the X chromosome more than, say, the autosomes.
The X does not recombine with the Y, so recombination and preservation of Denisovan X DNA will be restricted to the XX sex — maybe halving the amount of Denisovan X that people have today while both sexes can pass Denisovan autosomal DNA. So I wonder if that is the reason for there being scarce amounts of Denisovan X DNA today.
“… I don’t understand why that would restrict passage of the X chromosome more than, say, the autosomes.”
Good question. Even though I’m only an immunologist and it’s been a long time since I took genetics (dinosaurs roamed the earth then), my understanding is it simply a reflection of the hemizygous state of the X.
But I’m olde and it’s been awhile, so I am probably missing something.
???
Comment by Greg Mayer
Hemizygosity has been discussed as a possibility, but there are other possibilities. Jerry and Allen Orr discuss them here
Coyne, J.A., and H.A. Orr.1989. Two rules of speciation. pp. 180-207 in D.Otte and J. Endler, eds. Speciation and its Consequences. Sinauer, Sunderland, MA.
And Jerry followed up in
Coyne, J.A. 2018. “Two Rules of Speciation” revisited. Molecular Ecology 27:3749-3752.
See also other papers in the above issue of Molecular Ecology, which is a special issue on “Sex Chromosomes and Speciation”.
(I can’t find a free pdf of either paper. And, small correction, it’s the lack of Neanderthal DNA on modern X chromosomes that suggests reproductive isolation; I’m not sure if we know the chromosomal distribution in modern humans of Denisovan DNA.)
EDIT: In comment 7, John Hawks links to a piece of his about the distribution of archaic DNA between modern X and autosomes. His main point is to argue that the distribution can be explained in ways other than partial reproductive isolation, but, importantly, he notes that Denisovan DNA is also sparse on the modern X, so Mark Sturtevant is right. Thanks to John for commenting here and providing the link I was hoping for!
GCM
I may be right?? Woo-hoo!! That does not happen very often! 😀
Reading this with great interest and thankfulness, and commenting just because the host worries that people don’t read the science posts. We do! We eat them up.
I, too, am eating. Thanks!
Yes. I’ve followed this stuff for decades, though not my field. It’s all mind-boggling. When I read articles like these, I often think of Biochemistry ~1973, when DNA sequencing was nearest neighbor analysis and labs where protein sequencing was Edman degradation done manually. The technology has advanced beyond wildest dreams, databases, sophistication of analysis staggering, allowing questions that had been nearly unthinkable.
+1
This discussion has made my day!
Interesting discussion!
Thank you Greg and Matthew, and PCC(E). This is why WEIT is my favorite site along with the fun culture war stuff.
I’ve been watching the whole Denosovians of Our Lives soap opera for some years now and it is incredible. It is surprising there isn’t more of it in the popular press.
best,
D.A.
NYC
Love this stuff. Thank you to all the contributors for taking the time to present this.
Very interesting and informative. I really like hearing your (Greg, Jerry, Matthew) expert commentary on these findings.
Yes, me too. And I really like that some of the principle authors mentioned have also joined the discussion!
Don’t forget Chris Stringer, who has also contributed. He is one of the world’s foremost experts on Neanderthals, and paleoanthropology generally. He has been hinting for a while that the extraordinary Harbin, “Dragon Man”, fossil, which is the largest human skull ever found, was likely to be a Denisovan. This recent evidence strongly supports that hypothesis.