In yesterday’s post I outlined what I see as “the species problem” (the existence of discontinuities in nature in one area), outlined the “biological species concept” (BSC), and suggested that the BSC was the best species concept to use when studying that species problem.  Let me reiterate that I don’t think there’s a single “right” or “best” species concept—each has its weaknesses (some more than others)—nor do I think there’s a single species problem, either. On page 26 of Speciation, Allen Orr and I list five different species “problems,” though the one that I find most intriguing is the one I mention above.  I favor the BSC because I think it is the best one for solving that species problem.

I’ll try briefly today to show why the BSC, which conceptualizes species as those entities whose members are interbreeding among themselves but genetically separated from members of other species by reproductive isolating barriers (RIBs), is the species concept that not only explains the species problem (i.e. the discreteness of nature) but offers a solution to how that discreteness evolves.

It’s actually pretty simple.  Two sexually-reproducing and related taxa living in the same area cannot maintain their distinctness unless they have evolved substantial barriers to gene exchange. If all species exchanged genes with their relatives, and did so commonly, then nature would form a continuum: you would not be able to instantly discriminate between a sparrow and a starling, or a gingko and an oak.  (Some “species” do have limited gene exchange; more on that later.)  The mere observation that related and distinct groups in one location maintain their distinctness means perforce that there are reproductive barriers between them, even if those barriers involve the ecological or developmental inferiority of hybrids that do form between them.

Thus, the discreteness of nature must have some connection to impediments of gene flow between the discrete groups.  This may be why Dobzhansky, a geneticist, was the first person to draw this connection.  In a paper in Revista di Scienza in 1937, he made this explicit:

Any discussion of these problems [of discontinuities in the living world] should have as its logical starting point a consideration of the fact that no discrete groups of organisms differing in more than a single gene can maintain their identity unless they are prevented from interbreeding with other groups .  . Hence the existence of discrete groups of any size constitutes evidence that some mechanisms prevent their interbreeding, and thus isolate them.

For the professionals reading this, let me note that disruptive selection between two types of plants or animals in one area can also maintain them as discrete entities, but this is really a form of reproductive isolation as well, since by definition the intermediate forms are maladaptive under disruptive selection. Maladaptive hybrids constitute reproductive barriers known as postzygotic isolation.

This hybrid inferiority, for example, is the case in benthic and limnetic sticklebacks (Gasterosteus aculeatus) in the lakes of British Columbia, two groups with ambiguous species status. Some consider them different species, but most, like my friend Dolph Schluter, consider them “incipient species” or morphs. They forage in different places in the lake, and eat different food.

Here are females of the two types, showing how different they are (benthic on top, limnetic on bottom: photo by Todd Hatfield):

Regardless of whether we call these forms “species,” though, the point is that these two forms do interbreed in the wild.  Why, then, do they still remain distinct? Because, as Dolph and others have shown, the hybrids are at an ecological disadvantage.  They cannot forage as efficiently as either the pure benthic or limnetic forms and so leave few offspring.  This is a type of reproductive isolating barrier (we call it “extrinsic postzygotic isolation”) that helps keep the benthic and limnetic forms distinct.

This kind of barrier also operates in plants.  Several readers have vociferously asserted that related species of plants hybridize profusely in sympatry.  And, indeed, plants do seem to hybridize more often than animals (but not as often as everyone thinks; see pp. 40-45 of Speciation).  What people don’t consider when asserting this is that if plants really did hybridize profusely, or even moderately, with their relatives in the same area, the groups would no longer remain distinct.  Those plants would blur into a single variable taxon.  The fact that they do remain distinct despite hybridization indicates pretty clearly that there is some problem with the hybrids, which is, after all, demonstrates reproductive isolation.  And indeed, when you investigate these situations, you do find hybrid inferiority. We outlines some cases in Speciation.

One of the groups that’s often said to flagrantly violate the BSC is the oaks (genus Quercus; this is often said of cottonwoods, Populus, as well). On pp. 43-45 of Speciation I dissected this story and found it grossly exaggerated.  Boundaries between oaks are not nearly as porous as commonly thought.  Further, some of the evidence of “hybridization” between plant species is based on exchange of mitochondrial DNA (mtDNA) or chloroplast DNA (cpDNA) Biologists love to use these types of DNA because they’re easier to obtain than nuclear DNA, and also evolve more rapidly.  But we now know that both mtDNA and cpDNA move between species a lot more readily than the vastly greater amount of nuclear DNA, so claiming pervasive gene exchange based on observations of mtDNA or cpDNA alone is a perilous claim, and is known to be wrong in oaks.

The main point is that reproductive isolation is what keeps species distinct in sympatry (i.e., in the areas where they encounter each other and could potentially exchange genes). Without RIBs, nature would be a continuum.  Therefore, understanding the origin and evolution of RIBs is equivalent to understanding why nature is discrete in one area.  That is why the BSC is the best species concept to use for addressing this particular species problem. What is the origin of species? Under the BSC, that question becomes equivalent to “What is the origin of reproductive isolating barriers between closely related species?”.  And that is a much more tractable question.  Reducing the problem of speciation—or at least of discrete groups of sexually-reproducing organisms in sympatry—to the problem of the origin of RIBs is perhaps the greatest achievement of the modern synthesis, an achievement that, unlike many others, wasn’t foreshadowed by Darwin.

It’s telling that when evolutionists are studying the origin of species—how new species come to be—nearly all of them adhere to the BSC.  That is, they study the origin of reproductive barriers between incipient species.  This holds even if those workers adhere to other species concepts.  We give an example in Speciation: my friend Kerry Shaw at Cornell is a strong advocate of the phylogenetic species concept (PSC), and has written papers defending it and asserting its superiority over the BSC.  Yet in her own research on speciation in Hawaiian wingless crickets (Laupala), Shaw studies the origin of differences in mating song and cuticular hydrocarbons, factors that keep different cricket species from hybridizing in sympatry. This is an implicit admission of the value of the BSC in studying the origin of species in nature.

In fact, I don’t know of a single paper studying the process of speciation in real animals and plants in nature that doesn’t implicitly adhere to the BSC, concentrating on factors that impede gene flow.  There may be a few I have missed, but I try to keep up with the literature.  This, again, is a tacit admission of the usefulness of the BSC in understanding the origin of those entities we call “species.”

Finally, some species concepts that claim to differ from, and be superior to, the BSC really incorporate the BSC into their assumptions.  One of these is the “evolutionary species concept” (ESC), currently the most popular alternative to the BSC and one much beloved by systematists.  The ESC considers a species to be a lineage that is evolving “independently of other lineages.”

This concept has even more problems than the BSC.  First of all, its adherents never define what they mean by “independent evolution.”  If there is even a tiny bit of gene exchange, are lineages still “independent”  What about the very common situation within species (an example is different subspecies of birds) when some genes for morphology or behavior are evolving independently in different places, because the birds live in different habitats that select for different morphology or behavior, but the rest of the genome is not subject to divergent selection and so is not evolving independently in those two lineages.

A bigger problem is that of allopatry: when groups can’t exchange genes simply because they are geographically isolated (note: that is not necessarily the same thing as reproductively isolated). According to the ESC, if a few lizards travel to a distant oceanic island on a vegetation raft, and start breeding there, they instantly constitute a new species the minute they set foot (feet?) on the island, because from that moment they are “evolving independently.”  Would anyone really want to consider these a new species, though, simply because they land on an island that makes gene exchange with their ancestors impossible? Consider this: nothing biologically important has changed at the moment when those lizards land on the island.  Is that really the moment of “speciation”, then?

So one of the main weaknesses of the ESC is that it conflates geographic isolation with genetic isolation. And when the distinctness of taxa (and lineages) really is discernible—in areas where they coexist—lineages evolve independently precisely because they are reproductively isolated! After all, it’s the absence of gene exchange that constitutes “independent evolution.”  In that sense, the ESC really puts the phylogenetic cart before the genetic horse.  To explain why lineages evolve independently in sympatry, you must explain why there are barriers to gene exchange between those lineages—and that’s the species problem addressed by the BSC.

The common assertion of systematists that they alone—and not those pesky population geneticists—have successfully divorced pattern from process (see, for example, any paper by Quentin Wheeler), is simply wrong.  The ESC implicitly requires an evocation of process—the evolutionary process that creates reproductive barriers between lineages.

(A mini-rant about systematics here. I love systematics and think it’s ground zero for nearly all evolutionary studies. You can’t understand much about evolution, and certainly nothing about speciation, unless you understand the evolutionary relationships between species.  That’s why I’ve spent a lot of time making phylogenies, and why my most cited paper [Coyne and Orr 1989, Evolution] involves combining phylogenies of Drosophila species with data on reproductive isolation.

Nevertheless, many systematists, especially cladists, seem to have gone off the rails, making dumb and insupportable statements about evolution.  I don’t know why this is, since cladism has as its purview one of the most important insights in evolutionary biology: Hennig’s recognition that one could help reconstruct the history of life using shared derived traits (these traits are called “synapomorphies”).  Nevertheless, cladists like Wheeler constantly impugn evolutionary genetics and assert that only systematists have the right to construct species concepts. They also claim that their own systematically-based species concepts divorce evolutionary pattern from evolutionary process.  As I’ve just shown, they are wrong.  The assertion that synapomorphies are the traits to use when constructing phylogenies is a statement about process: those synapomorphies are shared and derived because they’re passed on from ancestors to descendants during speciation! If that’s not a process, I’ll eat my boots.)

I have reached the end of my two-part rant.  I won’t convince everyone, of course, but I hope I’ve convinced at least some readers of the utility of the BSC for answering a most important problem of evolutionary biology: why is nature discontinuous rather than continuous?  For those who wish to go further, have a look at Speciation by Coyne and Orr, especially chapter 1 and the appendix.

Enough about species. Now on to something really important: boots!

We had some good discussion on the thread about elephants yesterday, and by “good” I don’t mean “everyone agreed,” because they didn’t.  I was pleased to see people exchanging and defending their divergent views.

The quarter starts today, teaching looms, and I don’t have a lot of time to write this morning, but I’d like to mount a brief defense of the biological species concept (BSC).  Today I’ll just explain what the BSC is and state what I consider to be the “problem of speciation,” recognizing that not everyone will agree. Later today, or tomorrow, I’ll explain why I think the BSC is the best concept to address that problem.

The BSC defines (or rather conceptualizes) species as groups of interbreeding individuals that are separated from other species by reproductive isolating barriers.  By “reproductive isolating barriers,” I mean genetically based traits of organisms that prevent them from exchanging genes with members of other species. (Gene exchange can occur only when members of two different groups can form fertile and viable hybrids).

These barriers can include genetically based ecological differences that limit species to different habitats, so that they never meet and thus can’t mate; “sexual isolation”, the very common tendency for individuals to court and mate preferentially with members of their own species; the inability of pollen from one species of plant to grow down the stigma of members of another; the preference, when a female is multiply inseminated by a member of her species as well as a member of another, for using the same-species sperm to fertilize eggs; and the various forms of “postzygotic isolation” that prevent gene flow after members of different species have already mated and formed fertilized eggs.  These last forms include the inviability of hybrids or their sterility (the mule, a hybrid between a donkey and a horse, is a classic example of hybrid sterility).

All of these reproductive barriers keep members of different species from exchanging genes, that is, the barriers maintain the integrity of species.  That doesn’t mean that they evolved to maintain the integrity of species.  In most cases they didn’t.  As Allen Orr and I show in our book Speciation (from which all of these ideas are taken), in nearly all cases reproductive isolating barriers (RIBs, a good Chicago acronym), are the accidental byproduct of divergent evolution between populations.

For example, two geographically isolated populations (don’t mix up geographic isolation with reproductive isolation!), may evolve by natural selection to adapt to different habitats.  When that divergent evolution has proceeded sufficiently far, the genomes of the different “populations” may have diverged so much that they can’t work well when combined in a hybrid individual. The hybrid could then be inviable or sterile, a form of reproductive isolation that prevents gene exchange.  As this example shows, such barriers can simply be “accidents”: the byproduct of what happens when populations evolve along different paths.

I think the existence of species, in most cases, simply reflects these accidents of evolution. These discrete groups of organisms (see below) are simply what transpires when geographically isolated populations evolve away from each other. And their discreteness then becomes evident when those newly-evolved species come back to coexist in the same area. (For truly, one can see the discreteness of species only when they’re living in the same place.  That gives a clue to the connection between the species problem and the BSC.)

In a few cases, however, natural selection can directly favor the production of discrete groups.  One of them is “reinforcement”, which we’ve discussed before, in connection with my student Daniel Matute’s recent paper. The other, of interest mostly to evolutionists, is sympatric speciation.

If you read Speciation—and I realize that most readers haven’t—you’ll see that we have extensive discussion about alternative species concepts, and of the problems of both those concepts and the BSC. (By the way, if you’re at all interested in speciation you should read the book.  I don’t like sounding like Chris Mooney with incessant repetitions of “read my book,” but I’m quite proud of it. It took Allen and me six years to write, and involved reading hundreds and hundreds of papers.)

I know that the BSC can’t extend to asexual organisms, and in many cases is somewhat subjective.  What do you do, for example, if two groups have just a limited amount of gene exchange?  What about if two populations occur in different areas, like the elk and the red deer, so there’s no possibility of them encountering each other—something that is almost required to determine whether they can exchange genes? (You can force-mate them in zoos, of course, but that’s a one-way test.  If they do produce fertile hybrids in a zoo, that doesn’t tell you that they’d do so in nature, for the enforced confinement of a zoo may break down reproductive barriers that would operate in the wild. But if they do produce inviable or sterile hybrids in a zoo, like the Indian and African elephants, that tells you that they’re almost certainly biological species.)

But let’s put this aside for the nonce.  I’d recommend reading Chapter 1 and the Appendix of Speciation if you’re interested in species concepts. (You don’t have to buy the book: many libraries have it.)  Many of you might not agree with our take, but at least read it before you belabor me for not considering the problems of the BSC and the “advantages” of other species concepts.

Why I like the BSC is because it has a natural connection to what I—and many early evolutionists at the beginning of the “Modern Synthesis“—consider the “big species problem.”

The problem is to explain why nature is discontinuous rather than continuous.  Why, in one patch of forest, do all the birds, insects, and mammals (and yes, most of the plants, too!) fall into a limited number of discrete, objectively recognized categories?  Nature is not a continuum, with blackbirds blurring into starlings blurring into sparrows and so on.  If that were true, what good would Peterson’s bird guide be?  Nature—at least that moiety of nature that reproduces sexually—is discrete.

This problem was stated at the very beginning of the book that is widely regarded as having launched the modern synthetic theory of evolution, Genetics and the Origin of Species (1937), written by my academic grandfather, Theodosius Dobzhansky:

Theodosius Dobzhansky (1900-1975) at the microscope. He looked at flies until the day he died.

Right at the beginning of this book (and I recommend the first edition to budding evolutionists) Dobzhansky sets out the “species problem”.  On the first page he describes how many species there are on earth (he estimated 822,765 at that time!), and then, on the second page, is a section called “Discontinuity.”  I quote at length, because this is perhaps the best existing statement of the problem of speciation.  (Note, by the way, what a splendid writer Dobzhansky was.  And his native language was not English, but Russian!)

Organic diversity is an observational fact more or less familiar to everyone. It is perceived by us as something apart from ourselves, a phenomenon given in experience but independent of the working of our minds.  A more intimate acquaintance with the living world discloses another fact almost as striking as the diversity itself.  This is the discontinuity of organic variation.

If we assemble as many individuals living at a given time as we can, we notice at once that the observed variation does not form a single probability distribution or any other kind of continuous distribution. Instead, a multitude of separate, discrete, distributions are found. In other words, the living world is not a single array of individuals in which any two variants are connected by unbroken series of intergrades, but an array of more or less distinctly separate arrays, intermediates between which are absent or at least rare. Each array is a cluster of individuals, usually possessing some common characteristics and gravitating to a definite modal point in their variations. . .

Dobzhansky then goes on to talk about the discreteness of cat species (bless his heart), and shows that there’s no continuum between house cats, lions, and other felid species. Cat species are discrete, and any cat from nature is easily recognizable (that’s why field guides work!).  He continues:

What has been said above with respect to the species Felis domestica and Felis leo holds for innumerable pairs of species, genera, and other groups. [JAC: I’d maintain that Doby was wrong about groups above the species level: genera, families, etc. are not as discrete, or as objectively recognized, as species.] Discrete groups are encountered among animals as well as plants, in those that are structurally simple as well as in those that are very complex. Formation of discrete groups is so nearly universal that it must be regarded as a fundamental characteristic of organic diversity.  An adequate solution of the problem of organic diversity must consequently include, first, a description of the extent, nature, and origin of the differences between living beings, and second, an analysis of the nature and the origin of the discrete groups into which the living world is differentiated.

(My emphasis.)  Although students no longer seem to be interested, or even aware of, this fundamental problem, it was something that preoccupied the founders of the modern synthesis, including Dobzhansky and Ernst Mayr.

As we point out in Speciation, there is no “best” species concept.  Different species concepts are useful for different purposes. But I contend that the biological species concept is the only one that enables you to tackle, and ultimately to solve, the “species problem” described by Dobzhansky.  It’s the only one you can use to address the problem of why animals and plants in one area fall into a finite number of discrete groups.

In the next (and final) post on this, I’ll explain why.