When new species arise in sympatry, one has the opportunity to study natural selection while discounting the noise of environmental variation and evolutionary history (Mallet et al. 2009). Environmental variation and geography are ingredients that might lead to allopatric speciation but are not drivers of sympatric speciation. Rather than solely its rarity, controversy, and challenges, this is the real attraction of studying sympatric speciation: we can get most purely at the biological forces that lead to diversification.
It continues to be debated what evidence is needed to show that speciation happened under sympatric conditions. Some researchers emphasize that sympatric speciation must occur under conditions of divergence with gene flow while others place primary importance on the geographical setting, i.e. that speciation occurred without spatial barriers to gene flow (Bolnick & Fitzpatrick 2007; Fitzpatrick et al. 2008, 2009; Mallet et al. 2009). The ‘four criteria of Coyne & Orr (2004)’ have become the conservative gold standard to establish sympatric speciation: (i) sympatric contemporary distributions; (ii) monophyletic sister taxa not based on hybridization; (iii) substantial reproductive isolation; and (iv) a setting where a history of divergence in allopatry is unlikely. Generally, this has been taken to mean that crater lakes (Barluenga et al. 2006; Elmer et al. 2009), oceanic islands (Savolainen et al. 2006), and other dramatically isolated, homogeneous, and species depauperate locales are the habitats most amenable to sympatric speciation. The most plausible examples of sympatric speciation involve initial ecological divergence followed by the evolution of differential mate choice without spatial isolation, and only a handful of examples pass muster.
In this issue, Crow et al. (2010) argue to demote geography as an ingredient of sympatric speciation. Using a creative combination of population genetic, morphological, and in vitro methods, the authors focus on mechanistic criteria: the speed and pattern of reproductive isolation developing between sister taxa of fishes in an ocean setting where a history of divergence in allopatry would be entirely likely.
Greenling (Scorpaeniformes: Hexagrammidae) are marine fishes that live in the north Pacific continental shelf. This group of species is renowned for their brilliant colouration, especially in males (Fig. 1). Crow et al. (2010) studied three greenling sister species from a genus of six described species: Hexagrammos agrammus and Hexagrammos otakii, which have overlapping distributions off the east coast of Asia, and Hexagrammos octogrammus, which is distributed allopatrically to the east across the Aleutian islands. Sympatric species H. agrammus and H. otakii differ from each other in size, colour, body shape, meristics, nuptial display, location and timing of breeding, and also habitat use (H. agrammus inhabits seaweed beds of coastal waters and H. otakii, occurs on rocky coastal areas) (Crow et al. 2007, 2010 and references therein) (Fig. 2).
The greatest challenge of sympatric speciation is the development of reproductive isolation by natural selection in a group diverging with gene flow (Maynard Smith 1966; Felsenstein 1981; van Doorn et al. 2009). In contrast, in allopatric speciation spatial distance itself acts as a prezygotic reproductive barrier and no natural selection is needed since reproductive isolation may evolve in part as a by-product of adaptive differences between isolated populations. Prezygotic isolation can develop quickly by selection in sympatric conditions while postzygotic isolation (e.g. genetic incompatibility) tends to be acquired more slowly. This was first demonstrated empirically by the now famous Coyne & Orr (1989) study of sympatric and allopatric species of Drosophila.
Cleverly contrasting this differential rate that reproductive isolation will evolve is one of the most innovative aspects of the Crow et al. (2010) study. Theory predicts that if sympatric sister species show pre- but not post-zygotic isolation then they may have speciated by natural selection without geographic isolation. If, on the other hand, species pairs display post- but no pre-zygotic isolation, it suggests that they diverged in allopatry.
The authors use two approaches to assess the strength of these two types of reproductive isolation. First, they screened the morphology and genetics of hundreds of specimens and found no hybrids of the sympatric sister species H. agrammus and H. otakii. They take this as evidence for prezygotic reproductive isolation in nature. Second, Crow et al. (2010) did an experiment that few molecular ecologists studying non-model species even attempt: in vitro crosses between all three species to test the strength of genetic, or post-mating, reproductive isolation. They find that the sympatric species H. agrammus and H. otakii have reduced fertilization success (a post-mating but pre-zygotic test of isolation) but lack post-zygotic reproductive isolation. This fits with expectation for species that diverged without geographic isolation.
The situation for the allopatric species is more complicated. In nature, hybrids are commonly found between the two pairs of allopatric species, but only with H. octagammous as the maternal ancestor and all known hybrids are female (also see Crow et al. 2007). In vitro experiments generally again mirror the findings from nature (Crow et al. 2010). This suggests a complicated sex-linked incompatibility between species and hybrid inviability, which would be a very interesting focus of further research.
Whether all ‘required’ criteria for sympatric speciation are met with the Hexagammos example is debatable; certainly the geographical one is not. The fact that the sympatric greenling species differ in habitat and peak spawning times may argue for a role of spatio-temporal isolation, so purists who emphasize spatial settings (Mallet et al. 2009) over levels of gene flow (Fitzpatrick et al. 2008) might not be convinced. Crow et al. (2010) contend that the geographical criteria (sensuCoyne & Orr 2004) are unnecessary if we are truly interested in evolution and the mechanisms of sympatric speciation, i.e. a role for natural selection rather than geography in reproductive isolation. It remains to be seen how doubters of sympatric speciation will respond to these arguments. Regardless, this study reminds us to think critically about our foundations, definitions, criteria, and goals when we study (sympatric) speciation, be it with or without clear geographic barriers to gene flow.