21 February 2007

The moment one species becomes two

A diagrammatic representation of speciation (metapopulation
lineage divergence). Even if every taxonomist using a different species definition based on different properties (SC1 thru SC8) agreed that originally there was one species that subsequently evolved to become two, there would necessarily be disagreement about the number of species present in the gray zones. Drawing from de Queiroz1.

According to Kevin de Queiroz1, during the period of the Modern Evolutionary Synthesis species (including those that are asexual) "were equated with groups of interconnected populations [metapopulations] that form an extended reproductive community and an unevenly distributed but unitary gene pool or field for gene recombination". Moreover, as de Queiroz points out, even in descriptions of species as lineages, the general metapopulation concept of species was implied.

de Queiroz adopts this notion and proposes that the only necessary defining property of species is that they are separately evolving metapopulation lineages. All other properties that have previously been attributed to species (SC1 thru SC8 in the drawing above), for example, reproductive isolation, or the occupancy of a distinct niche, become properties that species as metapopulation lineages may or may not acquire. In de Queiroz’s words:

...metapopulation lineages do not have to be phenetically distinguishable, or diagnosable, or monophyletic, or reproductively isolated, or ecologically divergent, to be species. They only have to be evolving separately from other such lineages.

It is not clear to me, however, how de Queiroz’s proposal, which otherwise makes perfect sense, would pinpoint the moment of appearance of two species from one if we really cared to do so (see the drawing above). This is, of course, not a new issue and was discussed a while ago by Ernst Mayr2:

...the two daughter species are virtually identical at the moment of the split, and if any species differences evolve in the two separated lines, it is by gradual transformation. This makes it impossible to designate a precise point of origin of the new daughter species.

1Kevin de Queiroz. Ernst Mayr and the modern concept of species. PNAS 2005 102:6600-6607. pdf
2 Ernst Mayr. 1989. Toward a New Philosophy of Biology. (p. 325).


Frank Anderson said...

It is not clear to me, however, how de Queiroz’s proposal, which otherwise makes perfect sense, would pinpoint the moment of appearance of two species from one if we really cared to do so.

Oh, you went and pushed my button!

You can't pinpoint the moment. I haven't read Kevin's paper (I will now), but it seems like the idea is fine in theory, but practically no better than any other view of species. I think the practical problem of species diagnosis can be as hard as trying to find the exact point where the Ohio River "becomes" the Mississippi River. In fact, it's worse than that. It would be more like trying to figure out where the Ohio becomes the Mississippi given only a bunch of vials containing samples of river water from various points in the Ohio, Upper Mississippi and Lower Mississippi. With species, you can usually only sample some characters from some populations at a particular point (or set of closely clustered points) in time.

An aside: I should note here that our inability to pinpoint the exact place where the Ohio becomes the Mississippi doesn't mean the Ohio and Mississippi don't exist. You can infer from this statement what you will about my view of species...

I think the best we can do is attempt to determine whether or not two or more groups are probably reproductively isolated right now. It's a hypothesis. We all need to stop worrying and love the reality of biodiversity generation!


Andy, I agree with you that we can't pinpoint when one species becomes another one. But, otherwise, I seem to agree with de Queiroz that a species is a metapopulation lineage evolving separately from other lineages. And they don't have to be reproductively isolated!

Frank Anderson said...

And they don't have to be reproductively isolated!

This is technically true -- there are plenty of "good" species that hybridize with others -- but if there are no barriers to reproduction whatsoever, lineages probably will not remain separately evolving for long.

I think Kevin's definition is fine. I just think there's just no point in getting too hung up on where one species stops and another begins. In cases where we can make such distinctions easily, we're just getting lucky -- we're looking at lineages at a convenient (for species delination) slice of time. Move into the past or future to a speciation event and it gets messy again.