A recent paper in Nature, Speciation by hybridization in Heliconius butterflies is getting a fair bit of comment on the internet. This is a case where the researchers, wondering if an Andean butterfly species was a hybrid of two others, decided to test the hypothesis by re-evolving it deliberately.
The process is called "homoploid" because there was no change in chromosome number. Previous studies in various organisms, particularly in plants, but not restricted to them, had shown cases of speciation by hybridisation where disparate numbers of chromosomes in the hybrid were equalised by a secondary duplication of the fertilised cell's chromosomes.
Standard theory, for a long time, has been that homoploid speciation is unlikely to occur because without the difference in chromosomal number to inhibit backbreeding, the hybrid would eventual remerge into one or both of the parent species.
However, the Heliconius species use colour variation as a sexual signal as well as a warning signal to predators. Hence the new phenotype is a direct cause of reproductive isolation, or at any rate lowered reproductive traffic, with the parental forms.
Speciation is, it turns out, a process that philosophers technically call "supervenient" - which is to say that it is multiply realisable, and many distinct physical processes can generate it. In the early days of speciation research, around the 1940s, apart from polyploid speciation (where the progeny species has a different number of chromosomes inhibiting backbreeding) the only conceivable (to the researchers) form of speciation was where a population is isolated and local adaptation and drift incidentally caused reproductive isolation (allopatric speciation). Darwin's case for sympatric, or local, speciation was rejected on the current understanding of genetics. But we've learned a lot about genetics, and the math required to model it, since then and as I have previously blogged at my old site, a lot of work has made it feasible this occurs.
This mode of speciation, however, has slipped under the radar, at least for me. What it shows is that when the phenotypic changes affect not only ecological adaptation but reproductive isolation, you don't need developmental or karyotypic (chromosomal structure) isolation to form new species, and that it can happen rather quickly. And even more interesting is that it shows evolutionary biology can be an experimental science. Well, we knew that already, but cladogenesis (splitting of species) hasn't previously been the subject of experimentation, apart from Dobzhansky's laboratory experiments on reproductive isolation of Drosophila back in the 1940s.
And it suggests what has become ever clearer over the past 60 years - that hybridisation is a major force in evolution. This is ironic, as some have suggested that Mendel was inspired to do his work to bolster exactly that idea. Of course the hybrids he was working with were within one species. But there has been a lot of cases where hybrids between species form new species, just as Aristotle suggested.
I doubt that hybridisation is all that radical, though. In one sense, if two species evolved in different areas and came together and formed a hybrid, that is, in fact, a kind of allopatry. What it means to me is that the standard categories of speciation are miscast. If reproductive isolation is the sine qua non of sexual speciation, then allopatry, along with hybridisation, karyotypic differentiation, and local adaptation, are the incidental causes of what really is the driver of sexual evolution - lowered hybrid fitness. And that can be generated many ways.
Late note: See also Carl Zimmer's entry on this.
The mating preference is what surprised me. I never would have thought that it would be so tightly coupled to the wing pattern. In retrospect it sounds reasonable; mating preference could be tied to self-recognition, or maybe there's a connection between mating pheromones and wing markings.
Late to the party as ever,
Jackd - There was another paper on Heliconius cydno (the white stripped parent) in PNAS this year that showed the genes for wing pattern and mate preference are actually very tightly linked (or the two phenotypes are pleiotropic effects of the same gene)
The researchers also posit a fairly plausable route as to how a geographically isolated population might generate a large number of two stripped hybrids which might make mate prefference follow suit, actually I should really blog about this myself...