Hybrid speciation might be rare

October 23, 2016 • 10:49 am

Data show that the “normal” mode of speciation—the process in which one lineage divides into two or more species—involves the geographic isolation of populations of a single species. Over time, natural selection (and genetic drift) causes those populations to become more and more genetically different. When the genetic differentiation has gone to the extent that the separate populations evolved features that make them unable to produce fertile hybrids when they come back together in the same area (i.e. regain “sympatry”), then these populations have become separate species. They are now groups on distinct evolutionary trajectories, and their inability to exchange genes because of the evolved “reproductive isolating barriers” between them (e.g., behavioral differences in mating, preferences for different host plants or microhabitats, different times of mating, different pheromones, or the sterility or inviability of hybrids), is what makes nature “lumpy” rather than a continuum. The lumpiness of nature—the fact that, in a single geographic locality, in most groups you readily see distinct clusters of plants or animals (look at the birds outside your window, or look at a field guide)—is an important fact that can only be explained by connecting the formation of those “lumps” with the reproductive barriers that keep them from forming a continuum.

Geographic isolation is thought to be important because gene flow between diverging species tends to keep them from diverging. In our own species, humans in different places began the process of genetic divergence, as witnessed by the traits that distinguish human populations (these are correlated with geographic isolation, as the theory predicts), but this process was nipped in the bud by both population growth and the invention of forms of transportation that allow people to move much farther than they used to. There is now gene flow between many populations, and Homo sapiens is an example of a “polytypic” (variable) species that, if the populations had remained isolated for a million years or so, might have become more than one species of Homo.

One thing that biologists have discovered since the advent of DNA sequencing, though, is that gene flow between species is more common than previously thought. Reproductive barriers aren’t always complete (although they are now between our species and all other living species), and so sometimes hybrids are formed and genes can sneak between different species. In the group I used to work on, the closely related species Drosophila santomea and Drosophila yakuba, we and others discovered that the entire mitochondrion, with all of its own DNA, invaded D. santomea from D. yakuba, and there’s been a bit of other gene flow as well. (In most of the genome, however, the species remain distinct.) This could only have been due to hybridization, and it happened because although the species tend to live at different altitudes, there are areas of overlap where they can meet and hybridize, and the female hybrids (but not the males) are fertile.

So we know that genes sneak between species more often than we used to think.

Some biologists, however, have gone farther, and postulated that hybridization between two species can itself cause the formation of a third species, a process called “hybrid speciation.” This is somewhat common in plants, occurring through a special genetic mechanism called polyploidy. There are two forms. Allopolyploidy involves the hybridization of two species having different chromosome numbers, and since the different chromosomes can’t pair in the hybrids, those hybrids are sterile. However, if the chromosome number doubles in the hybrids, so that a new individual is formed with a chromosome number equal to the sum of the numbers in both parental species, one can get an “allotetraploid” populations whose members are fertile among itself but sterile when they mate with either parental species. (See any evolution textbook for an explanation.). This would, then, be a new biological species.  A similar process can occur if chromosome number doubles within a single species, producing an autotetraploid. Further hybridization and chromosome doubling can lead to entire polyploid series of plants with hundreds of chromosomes, as in ferns.

As I said, polyploidy, both auto- and allo-, is a fairly common mode of speciation in plants. As Allen Orr and I noted in our book Speciation (read chapter 9), roughly 2-4% of speciation events in flowering plants involved polyploidy of one sort or another, and maybe as many as 7% of speciation events in ferns. This is a rough estimate, and the real frequency could be higher. But polyploid speciation in animals is much rarer, and I won’t go into the suggested reasons for it (see pp. 333-337).

There’s another form of hybrid speciation called “homoploid hybrid speciation” or “recombinational speciation.” In that process, a hybrid is formed between two species, and then, if it is at least partly fertile, the genes from the different parental species can sort themselves out into new combinations of genes or chromosome arrangements from the parental species. If the new sorted-out population is reproductively isolated from the two parental species that produced it, we have a new homoploid hybrid species.

Many biologists (I won’t name them) have posited that this kind of speciation is rampant in nature, so that it’s not just the occasional sneaking of genes between species that’s important, but also the wholesale formation of new species after hybrid formation. Lots of suggested examples of such species have been given.

However, it appears that most of the evidence for non-polyploid hybrid speciation is weak. That, at least, is the conclusion of Molly Schumer, Gil Rosenthal, and Peter Andolfatto in a 2014 paper in Evolution (link and free access below), a paper that I only learned about at CoyneFest. Schumer et al. argue that good evidence for a non-polyploid hybrid speciation event requires satisfying three conditions, and I quote:

To demonstrate that hybrid speciation has occurred given this definition, three criteria must be satisfied: (1) reproductive isolation of hybrid lineages from the parental species, (2) evidence of hybridization in the genome, and (3) evidence that this reproductive isolation is a consequence of hybridization. By contrast, a large number of empirical studies have simply used genetic evidence of hybridization (Criterion 2) as support for hybrid speciation. . .

The authors argue that there are many ways that a species can look as if it’s a hybrid without actually being the result of full-scale hybridization (or any hybridization); that in some cases a hybrid lineage hasn’t been tested to show that it’s interfertile with other members of that lineage and reproductively isolated from the parental species, and, especially, there are almost no demonstrations that the genes or chromosome arrangements of parental species have sorted themselves out in a way that has created a reproductively isolated homoploid hybrid. That is, few people have shown that the reproductive isolation of a putative hybrid species involves genes that came from the parental species rather than, say, genes that evolved via natural selection after hybridization.

You can read the paper for details, but Schumer et al. conclude that despite the big noise from some biologists, there are only four cases of homoploid hybrid speciation that meet their criteria. Three of them are in one genus: the wild sunflower Helianthus, which has formed three diploid species—all adapted to novel environments—by hybridization of pre-existing species and the sorting out of chromosome arrangements that, with their divergent genes, reproductively isolate the hybrid population from the parents. That superb work was done by Loren Rieseberg and his colleagues.

The other case is the butterfly Heliconius heurippa, which genetic evidence shows almost certainly resulted from hybridization between the species Heliconius cydno and Heliconius melpomene, H. heurippa has a hybrid wing pattern, which you can see below, and it’s been shown that each species, as well the “hybrid”, are reproductively isolated from the others because males mate almost entirely with females who have their own wing patterns. Thus H. heurippa (shown below with its parents) satisfies all three of the authors’ criteria, for the genes causing reproductive isolation are precisely the color-pattern genes derived from the two parental species.

H. heurippa (the hybrid species)
H. melpomene
H. cydno

The upshot is that while the movement of individual genes between both plant and animal species is more common than evolutionists assumed before the gene-sequencing era, there is still scant evidence that entire new species of animals form via hybridization. Hybrid speciation is more common in plants, but only through the unusual mechanism of polyploidy, and homopoloid hybrid speciation (without an increase in chromosome number) doesn’t appear common in either plants or animals.


Schumer, M., G. G. Rosenthal, and P. Andolfatto. 2014. How common is homoploid hybrid speciation? Evolution 68:1553-1560.

37 thoughts on “Hybrid speciation might be rare

  1. Excellent post. I just saw a paper on allopolyploidy by hybridization in Xenopus laevis. Now I have things to read while getting my vehicle worked on.

  2. Why “evidence that this reproductive isolation is a consequence of hybridization”?

    If closely related species A and B occasionally come in contact due to migrants or to climate changes resulting in range changes, AB hybrids may form. If they’re +/- fertile, they may form a population. If they don’t have much contact with A or B, they can form a new species, even if the cause of speciation isn’t specifically the hybridization.

    However, as I write this, I see that speciation will be much more likely if hybridization made the F1’s unable, or very unlikely, to breed with at least one of the parents. Hmmm.

    1. This was going to be my question as well. If you have evidence of reproductive isolation from both parent species, why is it necessary to also show that the isolation is due to hybridization? In what sense does the evidence for condition 3 differ from the evidence for condition 1?

      1. The idea is that you might have formed a hybrid population which only later evolved reproductive isolation from the two parental populations, and that in this case Schumer et al believe (and I agree) that this is just hybridization followed by regular old speciation.

        i.e. if “hybrid speciation” is distinct from other kinds of speciation and interesting for that reason, then it has to be the case that speciation happens BECAUSE OF hybridization, not merely after it.

        1. I see your point, but how often does it happen that an existing hybrid population becomes isolated for reasons that have nothing to do with the hybridization?

          Perhaps they’re also thinking of the case when isolation precedes hybridization. That is, portions of two parental populations could become isolated together, and only then begin to interbreed and hybridize.

          1. >I see your point, but how often does it happen that an existing hybrid population becomes isolated for reasons that have nothing to do with the hybridization?

            We don’t know, but I think this is an “extraordinary claims require extraordinary evidence” kind of thing.

            In that it requires something fairly remarkable to happen in order for the offspring of two inter-fertile species to be non-fertile with either parent AND to be inter-fertile with other hybrids. Whereas hybridization followed by the evolution of reproductive incompatibilities via more normal means doesn’t require me to believe in anything that we don’t already know happens all the time.

            >Perhaps they’re also thinking of the case when isolation precedes hybridization. That is, portions of two parental populations could become isolated together, and only then begin to interbreed and hybridize.

            I’m not totally sure I follow. What do you mean that portions of the two parental populations become isolated together?

            1. I’m imagining a scenario in which two closely related species occupy the same territory but are disinclined to interbreed due to behavioral or pheromonal differences or whatever. A migration route opens up into an adjacent territory, and a few members of each species pass through before it closes again. Now we have a small mixed population isolated from both parent species, with a shortage of potential mates. Under these circumstances they might overcome their aversion to interbreeding and form a new hybrid species. But if I understand you right, this should not be counted as an instance of “hybrid speciation”.

  3. What struck me here was the background info that humans show some divergence of genomic traits as a result of geography but stopped short of speciation because of growing population and effective transportation. This is an interesting foundation for the ways in which our physical differences within the single species have played a role in human social structure, war, and philosophy for a long time. Perhaps at both the personal and political levels, the knowledge that people are different and yet the same has heightened not only our conflicts but also the emergence of the concept of a single humanity. So our “polytypic” status, an ambivalence of sorts, has been important.


    1. I must admit I am a but confused, because I seem to remember reading that Jerry describing races (polytypes) as real but contextual. I found a pay-walled article claiming human polytypes however. [ https://www.ncbi.nlm.nih.gov/pubmed/19695787 ]

      By the way, maybe your handle should be “fourpointthreebillionyears”? As I am interested in astrobiology I wrote a short opinion piece on the time scale of the universal tree when my new studies in bioinformatics [going for a 2nd MSc so far] demanded a short presentation. I apologize for the format, but at least the ideas and references are there.

      [When and Where Did Life Emerge?]

      The idea: The scant record rejects the hypothesized time constraint of a late heavy bombardment – the 3.8 Ga limit – as best it can. Zircons > 3.8 Ga lack the telltale shock fractures of impacts, and molecular clocks and organic carbon > 4.1 Ga seem to agree that life emerged as soon as the oceans did > 4.3 Ga.

      Harrison, who was co-discoverer of the oldest organic carbon, looked over the evidence for the late heavy bombardment. He found that it was consistent with bad statistics in coverage and methods, something that had not been controlled for. [ Boehnke P, Harrison TM. Illusory Late Heavy Bombardments. 2016. PNAS 113, 10802–10806. ]

  4. This is good to know. I would not have guessed, however, that H. heurippa was a hybrid of the other two species by looking at them.

  5. Fascinating! I am constantly amazed of how biologists (and biochemists) make so much out of so little.

    This was interesting: “males mate almost entirely with females who have their own wing patterns”. Is there an underlying selective process that latch on to the perceived difference in patterns, to adapt to the new mates? If it is so specific that it puts up a species barrier, the successful hybridization process must indeed be rare in such cases?

    1. I find that interesting too, though for different reasons. What are the genetics of this phenomenon? Do males learn their own wing patterns and look for them in females, and if so how? Or is wing pattern genetically linked to mate preference or to some other character (pheromone or whatever) that influences mate preference? Do females display any mate preference?

      1. Yes, those are good questions, John, and I don’t have the answers. But I bet the butterfly geneticists do. You could, for example test this easily simply by painting over the wings of newly hatched males, something that butterfly people do a lot. I’m not sure about female mate preference, but that’s probably in the literature, too.

          1. Interesting paper. It doesn’t offer much on the genetics, though. Mate preference could still be due to linked genes or self-assessment. If the former, I’d like to know how that works.

            If the latter, I’d like to know too. Their experiments rule out only imprinting at eclosion, as far as I can tell. They don’t rule out comparison of self with potential mates at mating time. Can these butterflies even see their own wing patterns?

  6. I just noticed a paper in Nature Communications showing that the European bison (Bison bonasus) is the product of hybridization between the steppe bison (B. priscus, an extinct species closer to the American bison), and the wild cattle, the aurochs (Bos primigenius) which went extinct in the early middle age. According to the authors, based on mitochondrial and nuclear DNA, the hybridization occured earlier than 120’000 years ago and indeed two morphs of bisons (corresponding to B. priscus and B. bonasus) can be recognized in prehistoric cave paintings.
    Soubrier S. et al, 2016: Early cave art and ancien DNA record the origin of European bison. Natzure communications 7, article 13158 (2016). doi: 10.1038/ncoms13158. Open access.
    I was lucky enough to see wild European bisons in Bialowieza forest (Poland) several times (and I was also invited to taste them! Excellent!)

      1. Difficult to test, as the two parent species have now vanished…
        According to the paper (I didn’t have time to read it in details) the wisent and the steppe bison occupied separate ecological niches (they don’t speak of the aurochs any more) and B. bonasus “alternated ecological dominance with steppe bison in association with major environmental shifts since at
        least 55 kya”. But that was at leat 70’000 years after the hybridization occured.
        What remains is that a species of hybrid origin did stay in parallel with both parent species for several thousands of years-

        1. Because most auroch genes must be present in their descendants, the domestic cattle, the hypothesis should be verifiable.

          I hope that, by digging enough into the genomes of different cattle breeds, and possibly by reactivating a few pseudogenes, the auroch may be brought back to life.

  7. Based on what I’ve read about 20 years ago, seems that prior to the arrival of Europeans in the 1700s, the Tasmanians had been the most genetically isolated people on Earth. Since then, alas, Anglo-Australians exterminated them as a distinct sub-set of Homo sapiens by the late 1800s so that the surviving modern Tasmanians are the descendants of offspring of multi-great grandfather rapists and multi-great grandmother victims. Before all that, however, they had not even had contact with mainland native Australians in thousands of years. I wonder if they had somehow remained isolated and survived for another million years if their outward appearance would have changed very much even as they changed genetically enough so as to no longer be able to produce viable offspring with other humans.

  8. Biologists are just obsessed with sex.

    While sex unleashes combinatorial explosions of possibilities for evolution to work on, and speeds evolution along handily, it’s not all that. If there are things to fix in the fitness of a common ancestor, those innovations will be made as mutations are introduced to repair them, and sex hasn’t so much to do with it. If the innovation is really selective, carriers will come to dominate even absent recombination with competitors. And if the innovations are relatively easy to reach, they will be reached in parallel on a comparable schedule, as we have so many examples of.

    The interesting thing with speciation is more about differential adaptation, or even arbitrary drift, rather than who’s banging who. Sexual compatibility itself is an adaptive spectrum, as we see with obligate outbreeders vs. anything-goes types. Sex can interfere with an adaptive program, and as such we would find incompatibility as a evolved defense against marauding rapists among others. Whereas if mates are hard to come by, or there is a cost of inbreeding depression, broad compatibility would be selective.

    I think the interest in speciation is an outgrowth of the old Linnaean program of tirelessly cataloging biodiversity, given new importance due to the ongoing loss of such diversity. But it’s just another collection of fuzzy demarcation points across biological space, one axis among countless.

  9. Heliconius heurippa has some biology that may make its origin easier to interpret. The species has a very restricted range at higher elevations (high for Heliconius – 1200 -1900 m) in the eastern Cordillera of the Colombian Andes. Heliconius melpomene is widespread (central America to s. Brazil) and usually occurs at lower altitudes – sea level to 1400 m. Heliconius cydno ranges from the northern Andes to s. Mexico, is mostly lowland, but unlike H. melpomene, does not abut or overlap with H. heurippa (although it overlaps widely with H. melpomene in Central America). How H. melpomene and H. cydno got together to make H. heurippa is a mystery.
    Many years ago I collected both H. melpomene and H. heurippa in the Rio Negro (a small Andes river) drainage near Villavicencio, Colombia. There H. melpomene occurred at lower altitudes and H. heurippa higher up. I don’t remember them overlapping, but I was only there for a couple of days. I suspect that melpomene-cydno hybridization resulted in cold-hardiness and enabled some hybrid butterflies (proto-heurippa) to perform better at cooler temperatures and invade higher altitudes. This effect (and the fact that melpomene and heurippa compete for the same host plant: a species of Passiflora) may have reduced sympatry with H. melpomene (and along with reinforcement – backcrosses probably were not so fit at low temperatures) reduced gene flow enough for (or to help) reproductive isolation.
    There are a couple of other “high-altitude” Andean Heliconius that are postulated to have hybrid origins.

  10. So we know that genes sneak between species more often than we used to think.

    My two cents: This fact seems much more significant to me in terms of producing diversity, and in fact points to reproductive isolation being far less important as a driver of evolution.

    After all, in the absence of reproductive isolation and in the presence of increased successful hybridization between different species, the exchange of diverse genes seems likely to increase both the rate and opportunity for diversity among fertile offspring. That diversity is what leads to the opportunity to branch into new survival niches, not the later isolation — which by definition means a reduction in new genes from further hybridization — which later leads to the loss of cross-fertility, and the awarding of a new binomial.

Leave a Reply