How many species of humans were there?

January 28, 2011 • 7:42 am

I must say that while Science’s reporting about evolution and ecology is pretty dire—they’re mostly puff pieces with little critical thought—the work of Ann Gibbons is an exception.   When presenting a new theory or discovery, she always seeks out any significant dissent in the scientific community, and weighs in critically herself.  Gibbons was, for example, one of the first print journalists to note the problems with Darwinius (“Ida”), the putative missing link that turned out to be not so link-y.

The latest issue of Science contains a short but comprehensive piece by Gibbons (free, I believe) on the latest findings in human evolution, including the second discovery of hybridization between archaic humans and “modern” Homo sapiens.  (This is the Denisova fossil that I wrote about recently.)

Do read Gibbons’s piece, as it’s a good way to get up to speed in 20 minutes or so.  She discusses the controversy about whether modern H. sapiens evolved “multiregionally,” transforming itself in many different places, or via replacement from a band that left Africa less than 100,000 years ago.  The hybridization data—Gibbons notes that up to 10%  of some modern human genomes came from mating with “archaic” ancestors—makes it impossible to clearly demarcate the two theories.  Gibbons ends her article by describing a wary rapprochement between the two main advocates of those competing theories:

As for Stringer and Wolpoff, both now in their 60s, their battle has mellowed. Their views, while still distinct, have converged somewhat, and they shared a beer at a Neandertal meeting last year. “The reason we get on well now,” says Stringer, “is we both think we’ve been proved right.”

But I want to discuss briefly a shorter second piece by Gibbons in the same issue of Science, “The species problem.”  Here she brings up the controversy about whether modern humans, Neandertals, and Denisovans were members of different species or the same species.   This question is far more important in dealing with humans than, say, with fruit flies, simply because there were far fewer types of hominins, and anthropologists’ careers depend on whether or not they name a new species.  That’s why there are so many species names in the hominin family tree—names that turn on characters as tiny as a few millimeters in the measurement of a tooth.  It’s likely that, several million years ago, three or four species of hominin did exist at the same time—and maybe at the same place—giving rise to fanciful scenarios about war and (especially) inter-group mating.  There’s nothing more salacious, in an evolutionary sense, than imagining a burly, hairy, and robust hominin male copulating with a female from a more modern-looking species.

But that was more than a million years ago.  What about the more modern groups of Homo, like Neandertals?  According to Ernst Mayr’s biological species concept, which Gibbons describes, individuals are members of the same species if they can mate with each other when they encounter each other in nature, and, critically, produce fertile, viable hybrids.  If they can’t, then there must exist genetic barriers to mixing of genes, the so-called “reproductive isolating barriers” that maintain the integrity of species.

But there is some slack in how biologists use the definition, for hybridization of this type can range from being very rare (e.g., the medium and small Darwin’s ground finches, which hybridize at a rate of about 2%) to more pervasive (e.g., the black duck and mallard form hybrid swarms when they meet).   If there’s only a small amount of hybridization, and the species’ gene pools stay pretty separate, most biologists consider the case to involve different species (this is the situation in the Darwin’s finches I just mentioned).

A couple of other points:

  • Mating between different groups is not enough to deem them conspecific: those matings have to produce viable and fertile hybrids.  And “viable and fertile” means not only that the hybrids can have offspring, but that they do have offspring in the wild.  Some interspecific hybrids in birds, for example, are viable and fertile, but are not recognized as proper mates by members of either parental species because those hybrids look weird or have strange mating behaviors.  That is a form of reproductive isolation, too: it’s analogous to sterility, but sterility on the grounds of not being attractive as a mate.
  • Therefore, if one just sees hybrids in the wild, that’s not enough to claim that the parental groups must be members of the same species. Those hybrids have to constitute a genetic bridge between the parental groups; that is, they have to be fertile and mate with the parental species.  Often we simply don’t know this.  I believe Peter Grant once had a paper in Science noting that about 10% of bird species are known to form hybrids with closely-related species.  That’s been used to claim that hybridization is pervasive in birds, and that bird species might not be so “real” after all.  But that’s a wrong conclusion.  We simply don’t know anything about the fertility and viability (and sexual attractiveness!) of most of those hybrids.  And the hybrids have to occur in nature, not in zoos or other artificial enclosures that might break down reproductive barriers that exist in nature.

Gibbons reports an opinion by Jean-Jacques Hublin, who works on human evolution:

In the real world, [Hublin] says, Mayr’s concept doesn’t hold up: “There are about 330 closely related species of mammals that interbreed, and at least a third of them can produce fertile hybrids.”

Not so fast!  That 330 number means nothing, since most of these produce sterile or inviable hybrids.  Too, how much of that “interbreeding” was observed in the wild versus zoos? (I don’t know the answer).  And only the cases seen in nature count. Lions and tigers hybridize in zoos, producing fertile “ligers” and “tiglons”, but they didn’t hybridize when they once co-occurred in India. Zoos break down reproductive barriers: animals hybridize because they’re bored, horny, and there’s simply nothing else to mate with.

Finally, if we assume that 110 mammal species produce fertile and viable hybrids that interbreed with the parents in nature (that’s a generous estimate), this constitutes only 2.4% of all mammal species (there are about 4500).  It’s misleading to claim that the biological species concept “doesn’t hold up” because it’s ambiguous at best 2.4% of the time. Think of all the other 97.6% of species where it’s not ambiguous.  That’s the problem with using rare exceptions to invalidate a concept that works nearly all the time.  Beware of these “anecdotal horror stories,” as one biologist called them.

So I don’t agree with Svante Pääbo, who is quoted by Gibbons as saying “I think discussion of what is a species and what is a subspecies is a sterile academic endeavor.”  This may be true when arguing about human fossils in the past, or in a single lineage, but is not generally true in sexually-reproducing plants and animals.  As I’ve noted before, species are real, objectively delineated entities in nature, and discussing why they are distinct, and how to diagnose them, is hardly a sterile exercise.  Species are not arbitrary divisions of an organic continuum.  In sexually reproducing taxa they form distinct groups, well separated in “morphospace” and “genospace.”  We need to understand why that is.

  • Finally, it’s dubious to define species based on the degree of morphological difference between them.  As Gibbons notes, “There’s also no agreed-upon yardstick for how much morphologic or genetic difference separates species.”  She’s absolutely right—to settle on such a yardstick is to make an arbitrary decision.  The black duck and mallard, for example, look pretty different, but are probably members of the same species.  If two species of geographically isolated birds were as different in appearance from each other as, say, Asians and Sudanese humans, they’d be called different species. Yet we don’t think of human “races” as different species because, despite their different appearances, they crossbreed easily and form mixed populations when they encounter each other.  Morphological (and genetic) differences are often a poor key to species status, particularly when we’re so attuned to small differences, as is the case in human fossils.

So what about “modern” H. sapiens, Neandertals, and Denisovans?  Clearly they hybridized, and some of the hybrids were fertile, for traces of Denisovan and Neandertal genes remain in our genomes.  On this basis, anthropologist John Hawks deems Neandertals, modern humans, and Denisovans members of the same species; Gibbons quotes him as saying “They mated with each other.  We’ll call them the same species.”  (I hope by “mating” he means “mated and produced fertile offspring”.)

But a little bit of gene flow isn’t enough to convince most of us that these groups were conspecific. On that basis, the Darwin’s finches would be deemed conspecific, but nobody does that.  The question is whether that gene flow reflected lack of opportunity for mating (in which case they might be the same species), or pervasive hybridization (between, say, modern humans and Neandertals) but only weak viability or fertility of the “hybrids” (in which case they’d be different species).   We will probably never know the answer to this.

Does this make the species status of these three groups purely arbitrary?  I don’t think so. What we can do is get a “yardstick” by seeing whether other species of primates that were separated for as long as Neandertals, Denisovans, and modern humans—roughly half a million years—have evolved into reproductively isolated groups.  I’m not sure what the answer is (it’s probably sitting there somewhere in the literature), but I’d guess that they wouldn’t be separate species, especially because humans have much longer generation times than other primates and so would speciate even more slowly.  If it were my call, I’d agree with Hawks (but for somewhat different reasons), calling Neandertals, Denisovans, and modern humans all members of Homo sapiens.

But as for the hobbits, H. floresiensis, I’d stick with calling them a different species.  They diverged from modern H. sapiens much further in the past, although they may have been contemporaneous with us.

54 thoughts on “How many species of humans were there?

  1. In all of the species posts that have gone up recently I don’t know if this question has already been asked, so I apologize if I am being repetitive. How does one define a species in asexually reproducing creatures (particularly single-celled organisms)? Is it a percentage of similarity?

    1. No, that’s a good question. I hate to pull a Mooney here, but to answer it would take reams of space. I’d suggest you have a look at Chapter 1 of my book with Allen Orr, Speciation, where the topic is discussed at length. The short answer is that we’re not sure if “species” even exist in asexuals, but if they do they might be defined ecologically.

      1. Thanks Dr. C. I need to buy your (other) book! This is one that I’ve wondered about for a long time. Seems to me that it would be hard to make biologically-significant definition of “species” amongst bacteria. Like you said: Ecologically based chararcterization seems like the only thing that would amke sense.

    2. I believe a classic study of the number of “species” of bacteria in soils in Scandinavia DID use an arbitrary yardstick – the percentage of DNA sequence similarity (these bacteria are not culturable). But everyone recognized that this definition is not biologically meaningful. I think they found that, with this yardstick, a pinch of soil had an average of 6000 different “species” of bacteria. With all the types of promiscuous gene exchange in bacteria, perhaps it is better to ‘punt’ on the species question?

      1. Yes–70/30, I believe (>30% of genes different=different species was the yardstick used, but of course drawing the line in a different place would result in a completely different final species count).

  2. It seems that, with the time scale we are talking about here (not millions of years but not a few thousands of years either), and with the degree of phenotypic divergence we are talking about here (there are usually small but clear morphological differences and stone-tool differences between Neanderthals and anatomically modern humans), there WILL be some arbitrariness in species designations -whether Neanderthals and AMH should be considered separate species or not. Given a continuous increase in divergence once lineages have had little or no gene flow, how can there NOT be a point in time when species designations are somewhat arbitrary (unless there is instantaneous or quantum speciation, as in some plants)? Neanderthals were once considered an archaic subspecies of H. sapiens, then they were considered a separate species, and now evidence of hybridization make some want to go back again. Nature has multiple examples of recently diverged species that OCCASIONALLY hybridize in certain parts of their overlapping geographic ranges, but not others. Why should our lineage be any different? Paabo may have a point, even if it is not a purely academic endeavor.

    1. Pretty much what occurred to me, too. When JEC wrote, “Species are not arbitrary divisions of an organic continuum,” I wondered “but what about a chronological continuum?”

  3. Occasionally one sees the designations ‘homo sapiens sapiens’ and ‘homo sapiens neandertalis’, perhaps now ‘homo sapiens denisovius’. Is that reasonable, or unnecessary nit-picking? What is your opinion, Jerry?

    1. I have no problem with using Latin trinomials to designate subspecies, as they do convey information. But sometimes subspecies status is pretty arbitrary, too!

    2. I could be wrong, but I believe that I used to see references to Homo sapiens neanderthalensis a lot. Then, clear consistent differences in morphology (shape of the braincase, jaw lines, slight browridges in Neanderthals, etc.) and the first mitochondrial DNA sequences from Neanderthals in 1997 (which were very different from modern humans) made most anthropologists use H. neanderthalensis instead. Evidence of some hybridization outside of Africa may make anthropologists go back to the first convention? But I think that, notwithstanding the recent evidence of hybridization, sequencing of a huge amount of the nuclear genome of Neanderthals suggests that modern humans and Neanderthals don’t share full common ancestry until about 500,000 years ago. that seems like a long time.

  4. The latest issue of Science contains a short but comprehensive piece by Gibbons (free, I believe)…

    Sigh, not free.

  5. Why must differentiation between two species be purely dichotomous, i.e., why must species boundaries always be categorical and distinct? It sounds like what you’re saying is that any two groups of organisms will fall on a range of interfertility — the vast majority of group pairs will not be interfertile, but a small number will successfully interbreed, and those will vary as to how often one sees successful interbreeding in nature. In these latter cases, it seems that the appropriate approach to the notion of species is one of fuzzy sets rather than rigid boundaries, and that species in these cases is not the best “natural kind” to use.

    1. Because otherwise, you’d have dogs and cats living together…mass hysteria!!

      Seriously, I think Dawkins made much the same point in The Greatest Show on Earth with regard to distinct fossil species, and how the urge to group fossils on the basis of small morphological differences is merely a human conceit.

      Never was a mother that didn’t recognize her own child.

      1. Sure, but my point is that genetic isolation is not absolute in many cases, and indeed comes in seemingly continuous gradations. We seem to be trying to find some arbitrary standard of just how much natural interfertile breeding we’ll allow before we call two groups the same (or different) species. That standard doesn’t seem to actually exist in nature, at least not in some notable cases.

        1. I like your “fuzzy sets” terminology. And is it not also the case that some of the fuzziest sets are the most interesting evolutionarily? I’m thinking of those undergoing “adaptive radiation,” if that’s still a term in use…

    2. I agree. I think a lot is to be gained if one realizes that discussing species and speciation is not necessarily for the sake of ‘naming species’. It can also be about describing and understanding populational diversification and its causes. In this light, I think it would be a good idea to treat the term species as a variable and not as a category (or rather talk about the amount of speciation.)
      Related to this, I think it is important to notice that all species concepts can be quantified. In the case of the biological species concept, for example, one could talk about the amount of speciation, depending on the amount of interbreeding that occurs or can occur. In this way one can discuss speciation, but one is not obliged to reach a dichotomous decision; to either lump or split.

  6. Although slightly off topic, the issues, disputes, and debates about what the term species actually means in the particulars (and thus has some effect on the value we get from our scientific inquiries using these different meanings) reminds me strongly of Harris’ thesis about the term morality and his presumption that science can be a valuable contributor in pursuing inquiries into it regardless of which specific definition we use.

    It just seems to me that the same kind of criticisms (lack of a unified and exacting definition of the term, the problem of prescriptive/descriptive starting points, etc.) can and perhaps should be leveled against evolutionary biologists who seem to have little problem finding value in the scientific pursuit of speciation regardless of the gradients of meaning that defines the term.

    Although no doubt many evolutionary biologists think Harris should just chuck the whole notion because of the problems in practice that science can help determine human values, I’m thinking the same kind of criticism seems obviously weak when the shoe is on the other foot. There is still, I think, tremendous value to be gained.

  7. You are the expert, and I am clearly not, but when you say

    As I’ve noted before, species are real, objectively delineated entities in nature

    do you mean that a species has a real existence as a single thing or that species is a useful collective term for individual organisms sharing certain characteristics?

  8. The question you raise about other primate groups was actually just addressed, in part, in a recent Nature paper (Locke et al, 2011, Nature 469:529-533) which looked at the Orang genome, with specific attention focused on Bornean and Sumatran orangs. But I’m not sure what is true for Orangs will be true for humans. By the time you are talking about later Homo groups, like “modern” humans, Denisovans, and Neandertals…you are talking about something that is many very important ways not an ape. The ecology and likely population structure of late Pleistocene humans groups is likely dramatically different than apes…the tripling of human brain size over the past two million years being just one sign of that difference. One of the challenges in paleoanthropology is that we might not have any truly representative extant comparative taxa to base hypothesis testing regimes on. Modern humans are all screwed up by post-Pleistocene population explosions and the demographic shifts associated with agriculture/sedentism, and apes are just really quite a different kind of organism.

    I believe one of the strengths of paleoanthropology is its potential to inform our understanding of the relatively unique evolutionary patterns which have characterized Pleistocene humans. Having a really large brain creates a lot of interesting opportunities for evolutionary change not seen in other organisms.

  9. The genomic sequence data, when complete, will address many of the issues here. First, the identification of chromosomal differences would indicate plausible constraints on hybridization fertility. Second the overall level of nucleotide difference should allow comparison with species of similar or greater levels of difference; if they are able to hybridize then it is unlikely that humans would be so different.

  10. I’m no expert in this field and this probably reflects my bias, but if we do classify Neandertals, Denisovans, and modern humans as Homo sapiens then I’d prefer some further differentiation between them, perhaps as sub-species. Perhaps something like Homo sapiens Neandertal and Homo sapiens sapiens (if that’s allowed in the naming conventions).

  11. Thanks, Jerry —

    Yeah, funny how long interviews get turned into truncated sentences, but clearly fertile offspring were involved!

    My feeling is that, when we’re talking about some populations of people having 8 percent of their genome from these ancient groups, we can start to say that reduced hybrid viability is no longer the null hypothesis. That is, call them the same species, because by all rights it looks like they fit the BSC, while acknowledging that our information is not as good as we might like.

    In this case, I take inspiration from you, and I add my voice to your frequent calls for people to read Speciation.

    1. One should note that most modern plant and animal species were described and named solely on morphological features of dead individuals on museum shelves or herbarium sheets.

      1. True, and if closely related species were found and described this way from the same area, then they certainly conform to the BSC (maintaining diagnostic differences in the face of gene flow indicates reproductive isolation). But lots of species have been sunk because they were allopatric and had slight morphological differences. In some cases morphology will “work” (i.e., correspond to the BSC), in others it won’t.

        If human “races” had been described by Martian zoologists when they were still allopatric, they would have been described as separate species.

    2. But is it not worth noting that matun between the moderns and the neandertals (leading to fertile offspring) was a verty rare occurence? The ancestors of modern Europeans had tens of thousands of years to mate with the neandertals, yet the eurpeans do not share any more genetic material with neandertals than east Asians. It seems this “mixing” may not have happened more than once despite ample opportunity. Wouldn’t this bolster the hypothesis that the moderns and the neadertals were indeed separate species?

  12. So, when they sequenced the Neanderthal genome and compared it to humans, and found the ~4% similarity, did they mention which sorts of genes we retained from them? I’m curious if that cross breeding gave any genetic benefit to the resultant descendants, or if we’re just talking about microsattelite markers.

  13. Lions and tigers hybridize in zoos, producing fertile “ligers” and “tiglons”, but they didn’t hybridize when they once co-occurred in India.

    Excuse me but it is my information that male ligers and male tiglions are sterile but female ligers and female tiglions are fertile and can interbreed with lions and tigers and produce fertile offspring.

  14. I have first cousins with heavy brow-ridges who actually live in caves. Since I’m from West Virginia, I, of course, married one.
    The five-year-old kids are already making decent stone tools. The wife and I are very proud.
    What was the question?

  15. This reminds me of the Crimson, Yellow and Adelaide Rosellas. They have viable hybrids where their territories overlap. They’re considered a superspecies or alternatively the Yellow and Adelaide species have been subsumed into the Crimson species. What decides where we have a super species or we subsume species? Is there any difference?

  16. On a lighter note, I’ve been attempting to obtain a DNA sample from Ms. Bachman or the Beck-ster to determine once and for all if american right-wingers evolved from Neanderthals. Anyone know where they get their hair cut?

  17. I don’t see how species can be anything but “arbitrary divisions of an organic continuum”, even given your definition. You note yourself the frequency of mating varies continuously, and I bet the frequency of producing viable hybrids does too, as does the frequency of producing fertile hybrids. After all, viability and fertility are bound to be due to complex genetic factors.

    Even if it were a stable definition though, it’s so hard in practice to observe the frequency of hybridization, the number of hybrids which survive, the number of those that breed, and the number of those that result in young, that we couldn’t apply it in the vast majority of cases. Surely something easier to use like genetic difference or LITUs (for fossils) would be better.

  18. Do we lump coyotes & eastern (red) wolves? (And then, since coydogs occur, do we end up ultimately lumping C. rufus & C. lupus?)

  19. When describing species, I usually use a rainbow metaphor.

    I can hold up a green crayon (say, Crayola), and we all agree it’s green. I can hold up a yellow crayon, and we agree it’s yellow. I can hold up a blue crayon, and those of us familiar with first grade finger painting will recognize green as being between yellow and blue. But how yellow does blue need to get before becoming green, and green to become yellow?

    We know when a color becomes a distinctly different color, but it’s a gradual transition. It’s the same with species. Every definition of species will put the bonobo and the daffodil as separate species, but if we follow each line down to the split some, 2 billion years ago, we will see, in a human time span, a very, very long and fuzzy joint; the blue-green, red-orange, green-yellow blend of genes.

    So far as the biological species concept goes, that zone of fuzz is the only problem (and prokaryotes, but for different reasons. Let’s stick with the macros). So if we must disregard the BSC, we should also stop labeling crayons, tell police that describing suspects’ cars by color doesn’t really help, dive into the biggest upset in the Arts since the end of the Spittle and Ash Blowing period. The fact that the BSC can’t define a [i]precise[/i] moment of speciation isn’t a problem unless you’re very anal retentive.

  20. Hybridization among sexually reproducing animal species is rare and unusual. It is therefore interesting because it tests the Biological Species Concept.

    There is a large literature on fish hybrids, rare though they may be. I’m presently reading an unpublished manuscript about two wide ranging sister species which; however, rarely occur together (food competition, I think). Some places where they are syntiopic, they form persistent hybrid swarms, in other places there is no hybridization. The paper is an attempt to correlate this range of events with the local ecology. DNA identification is used, and makes study of hybrids much more certain than it was in the past.

    On the other hand, when deciding to describe one of two sister species as new, I was pleased to learn that they would not attempt to breed together under lab conditions where homospecific pairs would breed on a daily basis.

Leave a Reply