Monday, March 26, 2012

On the antagonistic relationship between sexual selection and assortative mating

Posted by Erik Svensson

It is time for lab-meeting again, and this week we are happy to welcome Machteld Verzijden back from Rutgers University, where she has been working with Jessica Ware. Let's start the lab-meeting with her telling us about her work and the progress made during the visit.

Note the new time: Wednesday March 28 at 13.30 (not 13.00!).

After this, I was thinking we should discuss the relationship between sexual selection and assortative mating, two processes that are often confused and mixed up, particularly in the field of sympatric speciation. Although these processes are by no means totally independent, they are not identical and their population genetic consequences are very different. Moreover, they can counteract each other and hence could be antagonistic.

To understand the finer details of the complex relationship between assortative mating and sexual selection, we'll have to leave the murky shallow waters of "Adaptive Dynamics" and instead to turn to a clear thinker and a population genetic theoretician who knows what he is talking about: Mark Kirkpatrick från Austin (Texas). I was thinking we should read a much-cited papers from Proceedings of the Royal Society, entitled "Sexual selection can constrain sympatric speciation". Below, you will find the Abstract and here is a downloadable link to the PDF:

Sexual selection can constrain sympatric speciation
Source: PROCEEDINGS OF THE ROYAL SOCIETY OF LONDON SERIES B-BIOLOGICAL SCIENCES  Volume: 271   Issue: 1540   Pages: 687-693   DOI:10.1098/rspb.2003.2645   Published: APR 7 2004


Recent theory has suggested that sympatric speciation can occur quite easily when individuals that are ecologically similar mate assortatively. Although many of these models have assumed that individuals have equal mating success, in nature rare phenotypes may often suffer decreased mating success. Consequently, assortative mating may often generate stabilizing sexual selection. We show that this effect can substantially impede sympatric speciation. Our results emphasize the need for data on the strength of the stabilizing component of selection generated by mating in natural populations.


  1. Erik,
    Don't be so dogmatic.
    If you look into the finer details of the Kirkpatrick & Nuismer paper (and I take it, you will), you will realize they base their analysis on a quite unrealistic initial condition, degrading the value of their study (although I agree with many of their conclusions).
    Had you bothered to look into the finer details of Adaptive Dynamics, you would realize that genetic variation in fact is a common theme, that the effects of assortative mating have been discussed in detail, and that it is not all about sympatric speciation. In fact, it's mostly about general evolutionary ecology.
    Finally, I would humbly like to recommend a paper of my own:
    The topic is almost identical to that of Kirkpatrick & Nuismer, but the analysis and conclusions reach much further. Unfortunately, however, it swims deep in the murky waters of Adaptive Dynamics, so you'll have to dismiss it. Besides, I probably don't know what I'm talking about.
    Jörgen Ripa

  2. Hmmm...Don't be so sensitive, it was only a provocative comment (fraction of a sentence!) and I do personally not think that one should be too defensive of ANY approach, neither AD nor traditional population genetics, as both have their weaknesses.

    One thing that I think was interesting with this paper by Kirkpatrick and Nuismer (which has questionable assumptions as well) is that they clarified some of the assumptions behind models of sympatric speciation (whether AD or other models). For instance, it is apparently often assumed that allele frequencies for the trait under selection are 0.5, which is a reasonable assumption if two partly diverged populations/ecotypes come together and speciation is completed in sympatry (but was not initiated there). This could very well happen, but it seems that one could discuss whether it is "true" sympatric speciation.

    Another interesting thing is that sympatric speciation requires both disruptive selection AND assortative mating (including your model), which seems perfectly reasonable. The issue here is not this rather logical and stringent requirement, perhaps, but rather the strength of disruptive selection and assortative mating. In other words: how strong must these two processes be, at a minimum, for speciation to be accomplished? Documenting either or both of them is thus not sufficient, their combined strength must exceed some critical threshold.

    You might remember Daniel Bolnick at the Vienna-meeting (who by the way is quite critical of AD) who presented empirical evidence from stickleback populations showing both disruptive selection and assortative mating in operation, but their combined strength was far below the critical "speciation threshold". I strongly suspect that this is the case in many, perhaps most, systems in nature.

    Finally, I agree that AD is a useful tool which has been extremely succesful, particularly in the field of life-history evolution. Also, "evolutionary branching" is not necessarily the same as "speciation", but unfortunately this important distinction is sometimes forgotten as some of the most prominent advocates of AD tends to overemphasize the importance of AD in speciation (no names mentioned).