I won’t say that philosophers in general have nothing to contribute to debates about the nature of biological species, but this philosopher certainly does: Henry Taylor, a fellow in philosophy at the University of Birmingham. His paper in The Conversation (click on screenshot below) not only says that the most used species concept in evolutionary biology—Ernst Mayr’s “biological species concept” (henceforth “BSC”)—is not only wrong, but that we should in fact have no species concept. Ignoring nature completely (has he been outdoors?), he concludes that nature is not divided into the discrete groups that gave rise to the notion of species. Rather, he thinks, nature—like some ideologue’s notion of biological sex—is a continuum. In fact, he concludes that “there is no such thing as ‘the human species’ at all.”

Well cut off my legs and call me Shorty! My whole life I’ve been interacting with (and mating with) what I thought were specimens of Homo sapiens. Now I find that I’m mistaken: we form a continuum with other species. Could I have mated with a chimpanzee or a badger by mistake?

Taylor’s list of publications gives exactly one (forthcoming in Synthese) related to the notion of species, and manages to make a big to-do about a geographically isolated population of brown bears that can hybridize with brown bears and polar bears. That is one of his beefs about the BSC, which I discuss below the screenshot:

The BSC is not really a definition, but, as I emphasize in my book Speciation—written with Allen Orr—an attempt to encapsulate in words the palpable lumpiness in nature that we see before us.  And nature, at least in sexually-reproducing species, really is lumpy: it’s not the continuum, or “great interconnected web”, that Taylor sees. In Chapter 1 of Speciation, I give three lines of evidence for the reality of species: they aren’t just artificial constructs, or subjective human divisions of a continuum, but real entities in nature. Yes, there is some blurring in both sexual and asexual organisms, but by and large species exist as “lumps” in the pudding of Nature. If this were not so, biologists would be wasting their time studying species, and field guides would be of no use. There is no blurring, for instance, between our species, chimpanzees, and orangutans, nor between starlings, hawks, and robins on my campus. And so it goes for most of nature. Some hybrids may be formed between species, but they are often sterile or inviable, and so don’t blur the boundaries between groups.

What Mayr and others (e.g., Theodosius Dobzhansky) did was simply to describe what kept these lumps separate from one another where they coexist in the same location. And that factor was reproductive isolation: the existence of genetic barriers to hybridization that kept two species living in one place from forming fertile hybrids, and thus kept their gene pools separate. The BSC is this:

Members of different species are unable, when they live together in the same area, to hybridize and form fertile offspring: they are “reproductively isolated”.  Members of the same species are able to mate and produce fertile offspring with other members of the same species. 

Coexistence, or “sympatry”, is important in this determination because geographically isolated populations that show some differences can’t be fully tested under the BSC since they don’t encounter each other in a state of nature, and some species that hybridize in captivity don’t do so when they encounter each other in the wild (e.g., lions and tigers, which used to coexist in India).

There are of course intermediate cases—groups that are more or less “species-like”—depending on how much hybridization and gene flow they experience. But for sexually reproducing organisms, these cases are the exception (see Chapter 1 of Speciation). And of course, as we emphasize in the book, the BSC cannot be applied to species that lack sexual reproduction—like many species of bacteria. In those groups one may have to use other species concepts.

The advantage of the BSC is that it gives us an empirical program for studying how lumpiness arises in nature: it arises by the formation of genetic barriers, almost always between isolated populations that experience divergent evolution to the extent that, eventually, gene flow becomes impossible. (The barriers to gene flow aren’t directly selected for in most cases: they are simply byproducts of divergent evolution.) As I pointed out in my chapter, virtually everyone studying speciation in biology (as opposed to defining species), studies the origin of reproductive barriers. That’s a tacit admission that speciation does have something to do with reproductive isolation.

I won’t go on here: I recommend Chapter 1 of Speciation (it’s accessible to the layperson who knows a bit of biology), and, if you want to see the failures of other species concepts, read the Appendix.

Now, why does Taylor reject the BSC, and along with it all species concepts? He gives two reasons.

1.) Polar bears and grizzly bears, once living in different places (“allopatric”) are now meeting each other in nature due to the global-warming-induced disappearance of the cold habitat to which polar bears were once restricted. There is some hybridization between the two groups that now meet, and some of the hybrids are fertile.

Taylor says this shows that the two bears weren’t reproductively isolated, and thus weren’t species. But this is bogus: the two groups were biological species, isolated by what we call “ecological isolating barriers”: genetically based preferences for different habitats that kept two species from encountering each other. (The genetic basis of habitat segregation is important here: two groups isolated simply because they’re on different islands aren’t necessarily biological species because their spatial segregation is due to the contingencies of geography and not to evolution.) Thus the polar and grizzly bears were separate species, but their genetic barriers broke down due to climate change, making the differential habitat preference nonfunctional.

Species may not be permanently different: all of us recognize that groups that remain distinct in nature can, in the future, exchange genes because their genetic barriers have been circumvented by environmental change. Plants kept apart in nature because they are serviced by different pollinators (“pollinator isolation”) may, in the future, suddenly begin hybridizing if one of the pollinators goes extinct. Changes in habitat can efface genetically based ecological preferences, and so on. If you put lions and tigers together in zoos, this breaks down both the geographic and sexual preferences that kept them separate when they used to coexist in India. They can then hybridize and form fertile “ligers” or “tiglons”. Does this mean that lions and tigers are the same species? No, because the change in habitat (artificial confinement in this case) has broken down their genetically-based isolating barriers.

To say that the BSC is bogus because polar bears and grizzly bears now hybridize in some places is to throw out the baby with the ursine bathwater. And this isn’t even an intermediate case: it’s a case where a barrier has been effaced by climate change.

2.) Taylor then trots out the old canard (if ducks can trot) that organisms that don’t interbreed can’t be subject to the BSC. DUH! This is something I discuss at length in Speciation. Taylor:

The [BSC] definition makes use of the notion of interbreeding. This is all very well with horses and polar bears, but smaller organisms like bacteria do not interbreed at all. They reproduce entirely asexually, by simply splitting in two. So this definition of species can’t really apply to bacteria. Perhaps when we started thinking about species in terms of interbreeding, we were all just a bit too obsessed with sex.

Indeed, it’s hard (but not entirely impossible) to imply a reproductively-based species concept to bacteria. But different species do exchange genes, and there have been several attempts to discern bacterial species using reproductive criteria. The question hinges on whether there’s a problem to explain in bacteria: are they “lumpy,” like sexually-reproducing species, or do they form more of a continuum, and thus there’s not a biological observation that needs explaining? This question isn’t yet settled.

And that’s it: Taylor’s lame effort to topple the BSC—a concept that was not even meant to apply to asexual organisms.  He then throws into the mix Darwin’s own confusion about what species really were (this is well known) and on that basis wants us to deep-six all species concepts and all ideas that species even exist as discrete entities independent of human judgment. (Tell that to a robin who is courting other robins but not pigeons! Animals are themselves good taxonomists!)

Here you go:

Scrapping the idea of a species is an extreme idea: it implies that pretty much all of biology, from Aristotle right up to the modern age, has been thinking about life in completely the wrong way. The upshots of this new approach would be enormous, both for our scientific and philosophical view of life. It suggests that we should give up thinking about life as neatly segmented into discrete groups. Rather, we should think of life as one immense interconnected web. This shift in thinking would fundamentally reorient our approach to a great many questions concerning our relation to the natural world, from the current biodiversity crisis to conservation. [JAC: Yeah, what would we now conserve if all of nature is one interconnected web? Would we need to conserve everything?]

And, in a way, this kind of picture may be a natural progression in biological thought. One of the great discoveries of evolutionary biology is that the human species is not special or privileged in the grand scheme of things, and that humans have the same origins as all the other animals. This approach just takes the next step. It says that there is no such thing as “the human species” at all.

That last sentence is risible: there is no species Homo sapiens?! Does Taylor know that we cannot form fertile offspring with any other species (yes, it’s been tried with our closest relative: inseminating female chimps with human sperm produces bupkes). And it’s not the “next step” in dethroning humans as the pinnacle of evolution to then say that they don’t exist as a group.

The danger here is that those who don’t know much about biology and evolution will read Taylor’s piece and think he’s onto something. He isn’t: these criticisms of the BSC have been made many times before, and dispelled equally many times—I do it in my book, which is 15 years old. The palliative for Taylor’s nonsense—and here I have to be a bit self-aggrandizing—is to read Chapter 1 and the Appendix of Speciation.

h/t: coel

At least a dozen readers have called my attention to a new paper in Nature Communications by Anna M. Kearns et al. (reference at bottom, pdf here), supposedly showing “reverse speciation” in ravens. The paper has received a lot of public attention because it claims to show that two distinct species of ravens have fused back into a single species. And that has excited people because a.) they don’t think this has happened before, and b.) it shows that speciation is not the simple bifurcating tree that Darwin portrayed. Rather, some of the branches can grow back together, fusing into a new single branch.

But, as the authors note, we’ve seen fusion of species before. Often this is connected with anthropogenic habitat change, such as change in climate, local ecology, or the introduction of predators (see a summary paper here). We may be seeing this now with the polar bear and brown bear (“grizzly”). Previously isolated by ecological preference and adaptation (a genetically based “reproductive isolating barrier”), their ecological separation may disappear with climate change. As the brown bear moves north with warmer climate, it will invade the territory of the polar bear, and the two species can hybridize and have done so repeatedly in the wild. If the hybrids are fertile (and I can’t find data on this), the two species may well fuse into one.

But there are other cases of species fusing when humans weren’t responsible. After all, ecological change, climate change, and introduction of predators can occur without the intervention of humans. The problem with finding hybrid species in nature is that one can’t easily detect that they resulted from hybridization if the parental species are both extinct. (The hybrid nature of species is usually detected by seeing that they’re a genetic mosaic of the parental species, and if the parents don’t exist that’s hard to detect.) But we have plenty of example of “hybrid species” that haven’t replaced the parental ones, including many allopolyploid plants as well as diploid hybrid species in butterflies and sunflowers. Further, species do exchange genes more frequently than we used to think, and that “horizontal gene transfer” can mess up phylogenies.

But none of this invalidates the generalization that species nearly always form from geographically isolated populations that genetically differentiate to the point where they can no longer exchange genes, when they’ve evolved barriers to gene exchange like hybrid sterility, ecological preference and adaptation, mate discrimination, and so on.  Except in plants, hybrid speciation is the rare exception rather than the rule, and Darwin’s “bifurcating tree” of life, drawn in his notebooks, is still a good description of life:

Now, what did Kearns et al. find? They found that two old lineages of Common Ravens (Corvus corax), one widespread (“Holarctic”) and one from the West coast of North America (“Californian”), had diverged, probably after geographic isolation, about 1.5 million years ago, but then didn’t become two extant species because they began exchanging genes—repeatedly. (The geographic isolation may have resulted from the common ancestor being isolated in refugia during times of glaciation.) So now, in the Western US and Mexico, one finds a raven that looks just like other worldwide ravens, but carries an ancient lineage of genes that must have diverged from the ancestor of “Holarctic” ravens a long time ago. That’s detectable because you find, in that area, birds that carry two gene copies that diverged anciently—far more divergent that the normal variation within Holarctic ravens.

At present, the California and Holarctic forms are not different species because they hybridize readily in Western North America, and there are no two distinct “types”. You can see them in the map below: Holarctic is purple, California is orange, and the hybrids, carrying genes from both, are striped purple-and-orange (the orange type is not found by itself; its genes have simply merged with those of the Holarctic ravens).

The situation is complicated because there’s another species of raven that is “sympatric” (lives in the same area as) both the Holarctic and California ravens but remains distinct: the Chihuahuan Raven, inhabiting the black area below. It’s regarded as a different and full species (Corvus cryptoleucus) because there’s no evidence that it hybridizes with any other group; it appears to be fully reproductively isolated from other ravens.

The distribution of the two lineages (and two species: Common and Chihuahuan) from the paper:

What the authors discovered from DNA sequencing was that in fact the Chihuahuan Raven is more closely related to the California lineage of the Common Raven than to the Holarctic Raven, and apparently split off from the California isolate more recently than the divergence of the Chihuahuan and Holarctic lineages of the single Common Raven. The Chihuahuan branch split off from from the California lineage between 0.6 and 1.5 million years ago.

So what we have is shown in the diagram below: a non-bifurcating family tree. It reflect the ancient divergence of the lineage that produced the Holarctic branch of the Common Raven on one side and the Chihuahuan Raven + California lineage on the other. That branch then split again and more recently, producing the full species the Chihuahuan Raven and then the lineage of Common Raven that fused back with the Holarctic lineage. This should be clear from the diagram:

The upshot: The big question, and the reason this paper got so much publicity, is the claim, echoed in the paper’s title, that it showed “speciation reversal.” That is, the authors assume that the California lineage was once a species of raven distinct from the Holarctic Common Raven, but then fused with it later. Somehow the reproductive barriers (genetic ones) that kept them apart became ineffectual.

That, of course, assumes that there were once reproductive barriers between the California and Holarctic lineages. But we have no evidence for that! The fact that they fused so easily, and without human intervention, argues against substantial reproductive barriers, though there could have been some ecological ones. The authors simply assume that, because the Chihuahuan raven became a full species in less time than the Holarctic-California divergence, then the latter divergence must ALSO have involved full speciation.

But divergence time tells us very little about speciation. The key is whether a lineage evolves reproductive barriers from another one, not how long they’ve been separated. And those reproductive barriers are probably byproducts of selection, which can be either strong or weak. That’s why, when judging whether two populations are biological species, evolutionists prefer to use indices of reproductive isolation (observations of no matings, no evidence of genetic admixture) rather than divergence times. It may be, with the Holarctic and California lineages, that they simply didn’t diverge genetically enough to produce reproductive barriers as a byproduct. That’s what happened to Homo sapiens: our own geographic isolates in Polynesia, Australia, and the New World weren’t separated long enough from the rest of the species to become new species of humans. Now that we have transportation and migration, we’re in the process of slowly fusing into one big gene pool, which may never be fully mixed but mixed enough to keep us able to mate with people from every other place.

The authors recognize this problem but, perhaps recognizing the extra attention that “speciation reversal” will get (as opposed to “lineage reversal”), make the flawed “time argument” for speciation (my emphasis in the quote):

Thus, it is not clear-cut whether we should call the situation in ravens ‘speciation reversal’ or view it as a case of ‘ancient lineage fusion’. This contrasts with most other examples of speciation reversal, where there is strong evidence for the strength and nature of reproductive isolation prior to speciation reversal despite a very shallow divergence between lineages (e.g., sticklebacks). Two lines of evidence suggest that California and Holarctic lineages could have been reproductively isolated prior to secondary contact and lineage fusion. First, the timing of divergence of the Holarctic lineage and the ancestor of the California and Chihuahuan lineages between 0.9 and 2 mya (Supplementary Fig. 1) is approaching the limit where most bird taxa (especially those in the northern hemisphere) have evolved reproductive isolation (~2 mya). Second, life history traits, as well as mtDNA, intron, SNP and ENM analyses all support reproductive isolation between Chihuahuan Ravens and Common Ravens despite the more recent divergence of the Chihuahuan Raven and the California lineage 0.6–1.5 mya (Supplementary Fig. 1). This shows that ravens can develop reproductive isolation and maintain strong species boundaries after a more recent divergence than that between California and Holarctic lineages. We argue that our findings represent the strongest support possible for the conclusion of speciation reversal given the inability to measure ancient prefusion reproductive isolation.

I am not convinced by either line of evidence that the Holarctic and California lineages were separate species. Speciation times can vary among taxa, and if selection is weak, lineages may not progress to full species for a very long time. Using data from other birds doesn’t settle the issue in this particular case. The second line of evidence isn’t really independent: the authors simply say that because the Chihuahuan Raven speciated in less time than the divergence between Holarctic and California lineages, the latter two must also have been species. That’s not convincing either. “Time to speciation” varies widely, even within a group, depending on various factors that include the strength of selection, the degree of geographic isolation, and so on.

So what we see here is lineage fusion, not species fusion. Even the authors recognize the problem in the bolded bit above, but they try to obviate that. Pity that the science journalists who wrote about this didn’t know much about speciation.

h/t: Michael, Jószef, Tom

___________

Kearns, A. M., M. Restani, I. Szabo, A. Schrøder-Nielsen, J. A. Kim, H. M. Richardson, J. M. Marzluff, R. C. Fleischer, A. Johnsen, and K. E. Omland. 2018. Genomic evidence of speciation reversal in ravens. Nature Communications 9:906.

I’ll take “speciation” in this post, as do all the authors involved, to mean “the origin of reproductive barriers between populations that live in the same area, preventing them from either cross-mating or producing fertile hybrids if they do.” Most biologists think that speciation—the acquisition of these barriers—requires a prolonged period of geographical isolation between populations, allowing them to diverge through natural selection or genetic drift without contamination of genes between the groups. When that differentiation has proceeded to a certain point, reproductive barriers can arise as a byproduct of evolutionary divergence, and thus we have new species. (If we’re to be sure they are genuine “biological species”, they should be able to coexist after coming back together in the same area, or, in a one-way test, produce sterile or inviable hybrids when forcibly mated in zoos. If they’re cross-fertile in zoos, we can’t tell, for lots of animals that coexist in the wild without hybridizing can do so under the artificial conditions of confinement.)

This process, which Allen Orr and I described at length (and adduced evidence for) in our book Speciation, takes time—often lots of it. We estimated that producing a new species in this way in sexually-reproducing organisms takes on the order of a million years.

Yet there are ways that speciation can go faster—much faster. One way is “hybrid speciation”, in which two species have a rare bit of gene exchange, and that leads to the formation of a hybrid population that’s genetically heterogeneous. That population, faced with an odd genetic admixture and perhaps strong selection on a novel genome, might itself evolve to become reproductively isolated from both parental species, thus forming a third hybrid species.

This happens a lot in plants through the process of allopolyploidy (see here), which has accounted for a few percent of speciation events in some plant groups (and more in ferns). This form of speciation stands out for two reasons: it’s quick, often taking just a handful of generations, and it occurs without the need for geographical isolation of populations, since it’s begun by a rare hybridization event between coexisting species.  No evolutionists doubt the importance of allopolyploidy as a way of forming plant species.

Another and similar method, but not involving doubling of chromosome complements, is “homoploid hybrid speciation”, in which two normal diploid species, like the case described below, hybridize and, because some of the hybrids may be fertile or semifertile, those diploid hybrids can evolve quickly into a new species that is reproductively isolated from both parent species. (Note that in these cases the parental species can’t be absolutely reproductively isolated, as they have to mate and some of the hybrids have to be fertile. But, as Allen and I noted, the definition of “species” can allow for some trivial gene flow.)

In recent years homoploid hybrid speciation has been highly touted by some biologists, but, as I noted about a year ago, hard evidence for the process is rare. A 2014 paper in Evolution, using stringent criteria to examine possible cases, found only four reasonably convincing instances of this kind of speciation: three in a single genus of sunflowers, and one in butterflies. So the process, while interesting, has yet to be shown sufficiently common to constitute an important new take on how species arise.

But a new paper in Science by Sangeet Lamichhaney et al. describes what seems to be another case, this time occuring in the Galápagos finch genus Geospiza. The paper, which is a good one, has received a lot of press, some of it misleading, implying that this process could be common or that the concept of “biological species” is worthless. But the Science paper itself doesn’t say that.  The reference to the paper is at the bottom, but it’s behind a paywall. Judicious inquiry might get you a copy, though.

What it shows is that a new and very small species of finches arose on the Galápagos island of Daphne Major after a stray finch from another island made it to Daphne and mated with a local, resident species. This mating gave rise to a population that appears to be reproductively isolated from at least one parental species, and perhaps from the other. This isolation evolved in three generations or so, and thus the speciation event was very quick.

The story. A juvenile male of the large cactus finch G. conirostris, resident on the island of Española (and a small satellite island), flew more than 100 km to land on the tiny island of Daphne Major. Here’s a large cactus finch and the proposed journey that male took, bypassing at least two other islands to get to Daphne Major:

The hypothetical flight path, taken from the paper, which explains the route.

Once on Daphne, the male mated with a female of the local species: G. fortis, the medium ground finch. Here’s a female. Note that her bill (and also her body, which you can’t tell) is considerably smaller than that of the cactus finch, which has a massive, deep bill.

One of the big findings of this paper, achieved through genetic analysis, was that the errant male parent was a G. conirostris rather than a G. scandens (common cactus finch), which initially seemed more likely because G. scandens lives on the much closer island of Santa Cruz. What the newspapers that described the research usually failed to add was that this hybrid population and its isolation has been known for some time and the observations have been published (there have been no observations since 2012). The real novelty of the Science paper is the fact that the identity of one founding parent was a surprise, and that it flew over 100 km to get to Daphne Major.

The G. fortis X G. conirostris mating produced one female and four male hybrids. One male mated with another local G. fortis female, while another male mated with his hybrid sister. From then on, every bird descending from that first hybrid coupling mated incestuously, within the lineage, and this has been going on for six generations. As of 2012, this inbred population has formed a closed mating group that is tiny: eight breeding pairs and 23 individuals. Here’s the lineage showing the immigrant male (right), his female mate, and the offspring and who they mated with. After the first mating between hybrid 17870 and outsider G. fortis 15170, all matings have been incestuous—within the group:

We have then, a population whose members mate only with other members. That indicates some reproductive isolation from the local species (there are three, including G. fortis), but of what kind?  The authors posit three types of isolation:

1.) Sexual isolation. The physical appearance of the hybrid finches makes them undesirable as mates for the other G. fortis individuals, for their beaks and bodies are bigger. Finches choose mates partly based on their appearance, as they learn a “proper” appearance by imprinting on their parents. Although imprinting is based on “cultural” exposure and not specific genes that code for “mate with an individual having a big body and beak”, this produces mate discrimination based on genetic differences between the species, and thus can be considered true reproductive isolation.

2.) Another form of sexual isolation: song differences. We know that in the Galápagos finches the males learn their song by imprinting on or imitating their fathers (this is known from natural cross-fostering studies), and it’s likely that females, too, learn the “appropriate” song of a mate by hearing their father, and thus mate with a male having her dad’s song. Since these males sing a song different from that of either parental species, the hybrid females would tend to mate with hybrid males, perpetuating the incestuous lineage.

3.) There is ecological isolation. The deeper, stronger beaks of the hybrid population enables them to open the tough, woody fruits of Tribulus cistoides  (the “fever plant” or “puncture vine”), especially in the dry season when food is scarce. This allows some ecological segregation due to differential resource use, but also allowed the hybrid population to at least hang on to a tenuous existence on Daphne Major.  Here are some T. cistoides plants and woody fruits:

But is this species reproductively isolated from the distant species G. conirostris on Española? If it isn’t, it’s not a true biological species. We don’t know the answer because the hybrid “species” doesn’t encounter the population on Española. This is crucial, because it’s not a new species unless the hybrid species is reproductively isolated from both parental species. The authors admit that this is “unknown” but guess that isolation from G. conirostris is “likely” because of the difference in bill size, body size, and song. But this is speculative.

There are other interesting data in the paper, too, but they need not detain us here. The important result is that we probably have a very sparse hybrid species of bird, that it formed in about three generations, and we now know exactly which species were the parents.

It’s another question whether this new species will persist. I suspect it won’t because it’s very small and may be wiped out either by demographic stochasticity (fluctuations in population size) or by another cross-breeding event with G. fortis that could cause the new species to be “mated to death”.  But species concepts aren’t prospective: we don’t say a reproductively isolated population isn’t a species because it’s liable to go extinct. For, in the end, virtually all species go extinct without leaving descendants.

As for what this paper says about the ubiquity of homoploid hybrid speciation, well, not much. We had four good cases and now we have five, though you wouldn’t know of this paucity by reading the popular press. The only statement I object to in this very nice paper is its very last sentence, “Joint occurrence of rare and extreme events such as these may be especially potent in ecology and evolution.” But if they’re rare and extreme, why are they “potent”? They’re surely interesting, but don’t suggest for a second that these speciation events somehow change our view of how new species form.

h/t: Dom, j.j.

________

Lamichhaney, S., F. Han, M. T. Webster, L. Andersson, B. R. Grant, and P. R. Grant. 2017. Rapid hybrid speciation in Darwin’s finches. Science, online  23 Nov 2017. DOI: 10.1126/science.aao4593
eaao4593. DOI: 10.1126/science.aao4593