What are species?

Several years ago I was commissioned by Elsevier to revise Ernst Mayr’s entry on “Species” in their Encyclopedia of Genetics.  I was honored to be revising Mayr’s chapter, as he profoundly influenced me (and many thousands of others) on the nature of species. Because I have come to disagree with Mayr on many issues of species and speciation, my encyclopedia entry was much more than a revision. I welcome you to read it.

Origin of ecological species in laboratory evolution, starting with one clone of Bacillus subtilis. Within one culture, my students found two ecological species that had diverged in their habitats: on the surface of the broth (left) versus within the water column (right).

In reading Mayr’s treatises on species when I was a graduate student in the 1970s, I was moved by his claims that the onset of reproductive isolation profoundly alters the dynamics of diversification. In particular, he claimed that populations that could exchange genes were cohesive while those that could not were irreversibly separate. Mayr took the cohesiveness of interbreeding populations as his motivation for defining species as interbreeding groups (his Biological Species Concept). Because he saw genetic exchange as such a powerful force hindering diversification, he developed the model of allopatric speciation, where one species could split into two only when populations were geographically isolated from one another.


As a student, I wondered if one aspect of the cohesion Mayr discussed might include a uniformity of evolutionary responses across a species’ populations. That is, when the various populations of a species become subjected to an identical selection pressure (e.g., through global warming), will they evolve in the same way? The strong version of this hypothesis is that the various populations would respond similarly even without exchanging genes during selection. I began testing this idea in experiments on Drosophila in my doctoral work and continued this into my postdoc. At some point during my thesis work, I discussed my project with Ernst Mayr (he was an emeritus professor in my department), and he told me I was wasting my time—of course the populations will respond in the same way; they’re members of the same species! However, the gist of my experiments was that populations of the same species were impressively diverse in the physiological and developmental processes by which they would respond to the same selection. I wondered for the next four decades just how cohesive a species might be.


Many evolutionary biologists, including me, have come to believe that Mayr overestimated the cohesive force of genetic exchange among populations. One counterpoint to Mayr’s arguments is the theory of ecological speciation, whereby two adjacent and interbreeding populations living in different microhabitats can diverge indefinitely as they adapt to their respective environments. Jim Mallet and others have argued that reproductive isolation is not a prerequisite for ecological divergence and that reproductive isolation often follows the evolution of ecological divergence, sometimes by millions of years. Here is the key to understanding the inefficacy of genetic exchange in preventing divergence: selection against maladaptive foreign alleles can keep those alleles at negligible frequencies, provided that genetic exchange between populations is infrequent enough. Consequently, the most significant milestone on the path toward irreversible divergence may be ecological rather than reproductive divergence.


In my work as a microbial ecologist, I have come to understand that if genetic exchange is incapable of preventing ecological speciation even in the obligately sexual animals, it is a fortiori incapable of preventing diversification in bacteria. That is, genetic exchange in bacteria is so rare even among closest relatives (usually occurring around the rate of mutation) that it cannot prevent adaptive divergence between ecologically distinct populations.


My Mayrian world view of species took its greatest hit when I came to realize that many bacterial species may not even be cohesive at all. I began developing models of bacterial speciation with the idea that all species, whether obligately sexual or rarely sexual, may be cohesive entities. While the highly sexual animal species may be subject to cohesion by genetic exchange and genetic drift, and perhaps by similar selection, I thought that the rarely sexual species of bacteria might also have their own peculiar form of cohesion. Following work by Bruce Levin and others, I argued that the rarity of recombination would cause a purging of diversity genome-wide within an ecologically homogeneous species every time there was selection for some new adaptive mutation, hence cohesion. Some have argued that these hypothesized genome-wide purges do not actually occur in nature, presumably because estimates of recombination rates have been wrong. I have countered their arguments in various venues. (I will have more on this later this year.)


Nevertheless, I now believe that genome-wide purges may not occur within some ecologically homogeneous species, but not because recombination rates are too high. Instead, the reason comes from the arguments of Ford Doolittle, who has claimed that species origin and extinction are so rapid that the membership within any given bacterial species may be limited in the extreme to a single cell and perhaps its daughter. Starting from Doolittle, I have developed what I call the species-less model of bacterial speciation. Here the diversity within a bacterial species is limited not by some force of cohesion but rather by the very short time from the founding of a species by a single mutant (or recombinant) cell until the species goes extinct. We are currently developing phylogenomic tests of this idea.


Mayr saw cohesion as the quintessential attribute of species. It is one thing to quibble with Mayr about what are the most significant forces of cohesion within this or that kind of species. It is quite another to claim that there is no cohesion whatsoever in many species. When we eschew the cohesion of species, we leave the Mayrian paradigm altogether.


I hope you enjoy my encyclopedia article. If you don’t have access to the articles I’ve linked, please email me at fcohan@wesleyan.edu .