David Hillis on speciation

July 6, 2023 • 11:35 am

In October of last year David Hillis, a well known evolutionary and systematic biologist at UT Austin whom I admire, published what I saw as a good post about speciation—on Facebook, of all places!  I asked him at the time if he minded if I put it in a post here, but I got no response. It turns out that he saw my request only the other day after he was reviewing his correspondence. And he said it was fine if I put it up.

So it’s below, indented. I wanted to post it for two reasons. At the time, I wanted people to understand that the definition of species is different from the features we use to recognize species. The “Biological Species Concept”  (BSC), which I use (but not Hillis), defines species as groups of populations that are unable to exchange genes (i.e., produce fertile hybrids in nature) with other such groups.  But we recognize species, in general, by seeing groups of animals and plants that maintain their phenotypic distinctness from other groups when they live in the same area (in “sympatry”). As Allen Orr and I explain in our book Speciation, the definition is more fundamental than the recognition because the definition ultimately explains why there are species—why nature is lumpy rather than continuous. Why isn’t there just one kind of bird that is a spectrum ranging from hummingbirds to condors, instead of what we really see: a series of distinct groups of birds? (If there weren’t, field guides would be useless.) The recognition just reiterates that nature is lumpy, but the origin of reproductive barriers between these groups, which involves the definition of species, explains why the lumps exist. (This, by the way, is the real “problem of speciation” that Darwin didn’t solve. It was solved in the 1930s and 1940s by people like Ernst Mayr and Theodosius Dobzhansky). The definition thus leads directly to a research program that explains the lumpiness: what are the genetic barriers between groups and how do they arise?

(By the way, you needn’t tell me that there are gray areas in species delimitation: some species that can produce fertile hybrids. We discuss all the possible exceptions in our book.)

It’s important to note that Hillis uses a species definition different from mine, but as I argue in my book with Orr, the reason that lineages remain distinct from each other when sympatric is due to barriers to gene exchange. (David may not agree, but there you go.)

The second reason I’m posting this discussion is because in many ways the distinction between definition and recognition we see in speciation also applies to biological sex. Biological sex is defined as whether an organism has the equipment to make either large and immobile gametes (females) or small and mobile gametes (males). But in general we recognize the sexes by sex-related criteria that are not quite as binary as the definition itself: genitalia, chromosome constitution, and so on.  But I want to make it clear, since some people maintain this claim, that biologists did NOT define the sexes as binary because somehow they wanted —for reasons seen as bigotry—sex to be a binary when it’s really a spectrum. No, sex is defined as a binary because it is a binary in nature: in all plants and animals. (In humans there are a few exceptions, about 0.018%, but that’s close enough to a binary for me). The binary nature of sex is a result of natural selection, but I won’t go into the messy mathematical details about why the condition of two sexes with different gametes is an “evolutionary stable strategy,” and that a third sex, or more sexes, simply cannot invade the system.

And for reasons similar to why the definition of species is more fundamental than the recognition of species, the definition of biological sex is more fundamental than other characters we use to recognize sex.  The differential investment of males and female animals in gametes, with females investing far more than males (an investment that continues with pregnancy or laying eggs) is what leads to many of the fundamental differences we see between males and females.  Those include, in particular, differences in behavior, size, ornamentation, and other features produced by sexual selection, all of which begin with that differential investment in gametes. (See my paper with Luana Maroja in the Skeptical Inquirer.)

But this second reason arose after I’d already decided I wanted to publish David’s thoughts on species definition (i.e., species “concepts’) vs. species recognition. So here it is:

I see a lot of people, including many experienced biologists, confusing “species concepts” with “methods of species delimitation,” as well as with the study of “speciation mechanisms.” These three areas of study are very different from one another.

Species concepts provide a conceptual explanation of what we mean when we use the word “species.” They are a description of what species are, not how we can recognize them in nature, nor why they exist. For example, I would describe a species as “An ancestor–descendant lineage of populations that maintains its identity from other such lineages over time and space.” That is what I’m talking about when I use the word “species.” I think that is identical or close to what most other biologists mean as well, although they may choose to emphasize the methods they use to recognize species (species delimitation) or mechanisms that produce species when they are asked to define what they mean by “species.” They also might choose different words to describe the same basic idea of lineages isolated through time and space.

My preferred conceptual definition of species (or anyone else’s CONCEPTUAL definition of species) doesn’t say anything about why species remain distinct from one another through time, and it doesn’t say anything about the operations one might use to detect lineages that remain distinct through time and space. There are many different methods that people use to identify and delimit species. There is a lot of evidence (from morphology, behavior, genetics, etc.) that can be used to show that two lineages are (or are not) remaining distinct through time and space. Often, these operations require close examination of contact zones between potential species or geographic variants of species, to determine if they are isolated from one other or not. This diversity of delimitation methods is a strength, not a weakness, of the study of species.

Conceptual definitions of species also don’t tell us WHY species exist. Mechanisms of speciation constitute another important area of study, but this is distinct from species concepts and from species delimitation. In different groups, there are many different behavioral, ecological, and genetic reasons that species remain distinct from one another. The existence of distinct species in life is an observation, but we are still studying all the explanations for this observation.

When people say that there is lots of controversy among biologists about “species concepts,” what they actually mean is that there are lots of arguments about the methods we use to delimit species (species delimitation), or arguments about the relative importance of different speciation mechanisms, or arguments about different conclusions of species delimitation from different analyses. I think there is far less conceptual disagreement among biologists about what we mean by the word “species.”

28 thoughts on “David Hillis on speciation

  1. What, are you saying that birds are not on a spectrum? Why, that is sheer birdophobia, and makes Larus gulls feel unsafe. The ACLU will no doubt call for suppression of this website.

  2. The distinction between how a concept is *defined* and how instances representing that concept are *recognized* is super important. Thank you both for calling this out explicitly.

    1. …and I would just chime in to endorse the third intellectual task which is how a concept is *determined*, in the mechanistic sense of what causes speciation (and sexual distinction) to occur.

  3. Maybe sometime you will connect the female investment in offspring as opposed to that of the male so as to explain how social practices, historically, relegated women to a “secondary” role in society. Before infant formula women were tied down by breast feeding for up to several years in some places, and after that were responsible for raising, feeding and protecting the young for many more years. Small wonder that they never became soldiers or professionals where their work required full time attention. Even later there were few professions open to women who didn’t marry: tutors, housekeepers, nuns, prostitutes. It wasn’t until the 20th century that women’s responsibilities were eased and reduced. That they are still perceived as less important than men is something that males have to deal with and discard. Evolution needs to be understood as the underlying cause but in no way supports continued secondary roles for women in an industrial society.

    1. Interesting point, with profound implications. The greater female investment in offspring presumably underlies the mothers’ predominant role in rearing/protecting the young in mammalian species (illustrated in 3/4 of nature documentary films) and some other animals. Of course, this is merely a fact of nature, not a moral injunction. It does not enjoin against, for example, appointing female bears to be CEOs of major bear corporations. But it does suggest that certain psychological differences between males and females may have evolved in those species where differential functions in rearing the young are observed. [Of course, merely suggesting such a thing in regard to humans disqualifies an individual for the presidency of Harvard. Safest, therefore, to keep this discussion focused—or “centered” as we say these days—on bears and elephants.]

    2. The connection between parental investment and traditional roles of men versus women is often argued, and I don’t know if it really needs reminding. But that stuff may be a recent thing borne out of human societies living in (and ruled by) large population centers of humans. There was that interesting article that Jerry posted recently that found that in present day hunter-gatherer societies, women tend to do about as much hunting as men. Maybe not so much the more dangerous kind of hunting, and they would do it with kids in tow, but the differences are less stark than would be assumed if we hypothesize that these different roles harken back to the Stone Age. So the roles we see today could be something that is pretty recent.

    3. This is an important topic relevant to lots of other discussion on this site such as affirmative action. My wife fed and cared for our kids and while she did that had only a part-time academic life for almost ten years. When she was ready to get back to full-time university work she was objectively less experienced and qualified than some women who didn’t go on a long maternal hiatus. How should she have been treated by people reviewing her for a university teaching job: as a member of a disadvantaged class of moms, with a thumb on the scale to compensate for her stay-at-home years? or as a less meritorious applicant whose hiring might short-change her students and colleagues? An emphasis on merit favours the latter, but an evolution-informed understanding of these roles and effects might favour the former.

  4. This is the reason dummies like me read this site. It avoids doing all the research to come to the correct answer.

    1. Just wanted to agree with you, and say that even folks like me who do research on species and speciation also benefit from this kind of clear distinction between definition of a thing, how to recognize the state of the thing, and mechanism that generates the thing. We all need this stuff. It’s great!

  5. Thank you for posting this. This has always been a sticking point for me when discussing human-caused extinctions and the ‘sixth mass extinction’. Climate activists seem to me to use the loosest, vaguest definitions of species to claim humans have caused a dramatic increase in the rate of extinctions. It took me awhile to even understand that some people claim that two populations that can interbreed and which actually do interbreed can be considered separate species if an apparent physical barrier exists between them. Same with behavior differences which only lessen the chance of interbreeding.

    1. There are examples where we give species names to strongly distinct varieties that can still interbreed (i.e, wolves vs coyotes), but that should not be sufficient to absolve us from the findings that we are either in the early stages of a 6th mass extinction, or are about to enter it.

      1. lets just define all organisms as being the same so we can end extinction once and for all!

      2. Yes, I am a bit puzzled by Matthew’s point here. It doesn’t seem to me that estimates of the current rates of species extinction are being inflated by a looser definition of species than used to measure past rates of extinction.

        Having said that I would argue that an anthropogenic loss of genetic diversity within species is also something to be deplored. There might for example be distinctive geographical races of a species of falcon, say, and the extinction of one or more of these races is still a serious loss even though the species overall persists.

        1. Rates of extinction from previous periods are estimated using the fossil record.

          Previously categorized species are occasionally recategorized into new species that live in proximity to each other and are physically indistinguishable (e.g. Bavayia geckos from New Caledonia which went from 13 to 41 species recently). The lost of any of these 41 species would be counted as an extinction in the current period, but would not be detectable in the future fossil record.

          This leads to the likelihood that the estimate for the current rate of extinction cannot be compared to rates of extinction for previous periods, because they do not measure the same ‘concept of species’.

          1. I’m not sure that this makes very much difference. Estimates of current extinction rates are statistical projections and although the estimates vary widely they are at least a couple of orders of magnitude greater than the estimated ‘background rate’ and I don’t believe that taxonomic revisions will significantly affect that.

  6. I had been trying to come up with how to compare the confusion about biological sexes to confusion about species, since it seemed to me that these must somehow use similar muddled thinking. I believe the structure provided here oughta clarify things!
    But I am waiting for the sex-is-a-spectrum crowd to manage to misunderstand the point, even so. They will instead seize upon the point that since there are lots of different species concepts, and since the separate-ness of species can sometimes be hard to pin down, so too, by analogy, there are lots of different biological sex concepts, and so biological sex is also multifaceted and hard to define.
    Its only a matter of time, I think.

  7. great post! Reading this helps me think about species delineation vs. speciation as a process. I always wondered why different types of black bass are referred to as separate species even though they can reproduce with each other, the same is true of the white mulberry and the red mulberry which readily form hybrids but are still considered distinct species. Dog breeds, on the other hand, show huge variation (physically and behaviourly) but are all considered to be the same species.
    People can debate whether or not some of these are actual species or just variation within a species but what is often more interesting is why they differ and what kinds of uses we have for these different classes. Like so many debates, people often agree conceptually but disagree semantically.

  8. For whatever reason, I thought Darwin went pretty far in understanding speciation. I knew he didn’t know about genes/heritability and the “peacock feather made him sick,” but I didn’t realize speciation had him stumped as well. Either way, thanks for this post: clear and edifying.

  9. My personal lightbulb moment for the “species problem” (delimitation) is that it was a problem of human expectations and human bias. For whatever reason, we humans like discrete boxes to put stuff in, and we sometimes expect nature to be that way in all areas. The movement between ancestral population, incipient speciation, and discrete species does not fit that neat little box.

    Sex, on the other hand, does fit into those discrete boxes. If your question is which two people you need to make offspring, the two populations you draw from are binary.

  10. Sorry for this semi-technical (maybe too technical?) question, and apologies if it is dealt with in your book- which I have not read.
    Thinking about the biological species concept, it appears from hybridization studies undertaken in many different animal species that some inter-species hybridizations are truly genetically incompatible (no viable offspring), other species hybrid pairs are completely compatible (all offspring viable), while yet other species hybridizations are intermediate in viability (often with female heterogametic Haldane Rule abnormalities and non-viability). In nature, all of these species typically do not mate with other species; in insects this reproductive isolation is mediated by sex-specific pheromones.
    Is there a general corellation between cladistic “closeness” and interspecies hybridization viability? Can we say that two species that can create viable progeny if forced in vitro to hybridize must have separated from each other later in the cladistic tree for that genus?
    Maybe that’s too obvious?

    1. It’s a great question and I think the answer is: not necessarily. Imagine three related species A, B, and C. The first speciation event separated A as a lineage from the ancestor of B and C. Maybe they’re allopatric and strong breeding barriers don’t develop. Later, B and C diverged from their ancestor and clearly they are closer related to each other than either is to A. But say C went on to develop a bunch of genetic differences that are strong barriers to fertile hybridisation with A and B. In such a case, A and B might be more able to hybridise and they might even look more similar to each other. It’s a genetic version of residual similarity due to autapomorphy.
      Also, I want to add my admiration for the clarity of this post.

  11. Jerry: “In humans there are a few exceptions, about 0.018%, but that’s close enough to a binary for me”. While I tend to agree with Jerry’s view on sex (definition) this bothers me. If you consider we are ~7 billion that amounts to ~126 million people; or over 13 Israels. Not something to sniff about. I get this is not the intention, but may need some elaboration.

    1. I think you did your math wrong because you’re taking percentages as fractions. The actual number (unless I erred) is 1.26 million. And who sniffed about the exceptions? What I said is that 99.982% of people conform to the binary. If you toss a nickel 7 billion times, it will land on its edge about 1.26 million times.

      At any rate, what you fail to consider is that these “intersex” people ARE NOT MEMBERS OF OTHER SEXES. They are of indeterminate sex. So insofar as sex can be determined, yes, it is a binary. Nor am I denigrating the intersexes, who can have hard lives.

      Thanks for the comment about Israelis, which seems a bit snarky to me, but you need to learn to do math and also learn what a “binary for sex” means.

    2. The important thing to the sex binary is that no individual produces any gametes other than spermatozoa (which define male) and eggs (which define female.). No matter what developmental disorders an individual might suffer that make it hard to identify his/her sex at a glance with the naked eye, no individual, no matter how ambiguous the genitalia, produces “ambiguous gametes”, like an egg with a flagellum or a spermatozoon that kept all its cytoplasm and contributes its mitochondria to the zygote. Even if an individual produced both sperm and eggs—no humans have been observed to do so—their gametes would still be typical sperm and eggs. We can say this because the differentiated tissues that give rise to gametes can’t produce intermediate or totally different third forms.

      Even if there really were 126 million, or 1.26 billion, such people in the world, it wouldn’t change the fact of the sex binary: the number of non-binary sexed individuals producing “spergs” or “speggs” (Colin Wright) would still be zero.

      People with differences in sexual development have medical conditions. They aren’t on a glorious rainbow spectrum of “normal” the way “gender-fluidity or nonconformance” is alleged to be. Whether they are rare or common they need to be regarded as individuals with specific physical and psychological concerns. But from a biological point of view, it doesn’t matter if their bodies can be shoehorned into an “Are you ‘really’ male or female?” dichotomy. What matters for this discussion is that they all make eggs, sperm or neither, just like everyone else. Of course this may not be the most important consideration to the person discovered to have a DSD but I’m here speaking biology, not medicine.

  12. The correct statement (in my opinion) is all (actually almost all) humans are born with exactly one sex. A very few living humans have 46,XX and 46,XY cells in the same body (and can be said to have two sexes). My guess is that the 0.018% statistics refers to chromosome abnormalities (for example, XXY) and/or ambiguous genitalia at birth. However, folks with chromosome abnormalities and/or ambiguous genitalia have one and only sex.

Leave a Comment

Your email address will not be published. Required fields are marked *