Understanding the sexual selection we cannot see

So I’ve been considering starting a blog for some time now, mostly to share cool (and not so cool) research I happen to stumble upon and feel like talking about. But between trying to survive my PhD, organizing the Evolutionary Biology Online Journal Club (join us! :D) and participating in the Breaking Bio podcast, I never really got to doing it. Rambling in 140 characters has usually been enough, and I rarely feel like I have much to add to the conversation. But today this paper was published, and I felt like talking about it for a little longer.

But first things first: my name is Rafael Maia and I’m a 4th year (ohmygodit’sgoingsofast) PhD student at the University of Akron Integrated Bioscience program. I study structural colors in feathers – colors produced by structures organized at the nanoscale to interact with light in a predictable way across the visible wavelength range -, in particular iridescent colors, which change their properties depending on the angle between the incident light and the observer, a classic example being the bright and colorful feathers of hummingbirds. My overarching interest lies in understanding how these colors are produced, how the nanostructures get organized during feather development, and how this influences how these colors can diversify under natural and sexual selection.

You can find more information about my research on my website, but essentially I am interested in how sexual selection interplays with developmental constraints to promote the diversification of ornaments – traits that can’t be explained by their benefits to survival alone.

So, on to the paper.

The paper by Padian & Horner published in TREE a couple of days ago is a response to two papers published in the same journal last year, so I guess I should talk a little bit about them before talking about the comment.

Sexual dimorphism in Anchiornis huxleyi? (source: http://ewilloughby.deviantart.com)

The first paper by Robert Knell and collaborators tackles a very interesting (and important) topic: sexual selection in prehistoric (and thus extinct) animals. Many fossils have horns, crests, and exaggerated traits analogous to those found in present day animals, which commonly use these to attract mates or compete with individuals of the same sex for mates – that is, traits that are sexually selected. However, such traits can also evolve for other, naturally-selected reasons as well. So we have a problem here: we know sexual selection is important in extant species, and given that the Doctor hasn’t visited yet and we can’t go back in time and actually measure how these traits function in extinct animals, how can we use the evidence we have to infer sexual selection in extinct animals?

Knell et al. do an excellent job at describing the problem and a potential solution to it: while we can’t measure sexual selection in species long dead, sexual selection is a very robust theory that makes clear predictions, which may allow inference from these fossilized structures. While definite answers may be hard to get, we don’t need to cross our arms and say there’s nothing that can be done. This paper was one of my favorite papers published last year, and I strongly recommend reading it even if you’re not a paleontologist nor interested in sexual selection, because the problem presented (and the potential avenues for progress suggested) are shared by many areas. For example, I’m interested in testing sexual selection hypotheses in a comparative framework, but comprehensive data on mate preference and selection on ornamental traits is usually restricted to a handful of species, meaning that for a broad phylogenetic comparison to be made, I face some of the same problems that Knell and collaborators discuss in their paper.

The second paper, by Tamra Mendelson & Kerry Shaw, focuses on species recognition – how individuals identify other individuals as being from the same species, which has important evolutionary consequences to hybridization and the maintenance of reproductive isolation, and thus ultimately to speciation. Mendelson & Shaw propose that species recognition is not fundamentally different from mate choice, representing one extreme in a continuum of the fitness consequences of mate choice: choosing a poor-quality mate from your own species reduces your fitness, much in the same way that (but to a lower degree then) choosing a mate from the wrong species does. They propose that treating species recognition as an extreme case of mate recognition and mate choice can move the field forward by integrating the advances from these fields in a common framework, while also presenting interesting consequences to how we define or delimit species. It is also an excellent paper and I also recommend reading it.

So, on to the paper (for realsies this time).

As I say above, I really liked both papers a lot, which is why the response by Padian & Horner immediately got my attention. It’s a short 2-page response that attacks both papers with a single blow. The gist of the paper, as outlined by the authors themselves, is as follows:

► Sexual selection requires sexual dimorphism; this was integral to Darwin’s original definition.

► Claims of sexual selection in extinct animals often rely on vague or untested assertions.

► There is at present no documented case of sexual dimorphism or sexual selection among dinosaurs and their relatives.

► A test of sexual dimorphism requires an independent understanding of the chronological age of the specimens in question; this has never been accomplished in an extinct animal.

I woke up, opened my computer, and read that as I was going through the Table of Contents I subscribe to. The first sentence alone got on my nerves, so I took a deep breath, brewed some coffee and downloaded the paper to read.

Padian & Horner’s critique of Knell et al.

Padian & Horner define sexual selection as (emphasis mine):

the process of selection for traits possessed by one sex but not the other, or used by one sex and not the other, that increase access to mates by attracting them or by repelling rivals for mates.

Therefore, sexual selection generates sexual dimorphism, and it becomes necessary to establish the sex, and a sex-dependent expression of the ornament, of the fossils. According to them, this has never been done.

In addition, to show that an exaggerated trait is an ornament, it is necessary to show that it does not merely represent differences in developmental stage, and thus it is also necessary to assess the age of the animal being investigated. They show the example of elephants, where after their adult mass is reached, some bones still continue to develop and ossify. So individuals can have more or less the same overall size but very different sizes of particular structures which actually represent differences in age, not sex – thus not representing dimorphism, and consequently not being sexually-selected traits.

Padian & Horner’s critique of Mendelson & Shaw

The criticism of the second paper is a lot more murky and harder to “really get”, but essentially boils down to Padian & Horner not being too keen on the idea of treating species recognition at the same level as mate recognition (the latter, they argue, being a subcategory of the former). In their words (again, emphasis mine):

Notions about species concepts, which are particular to limited clades at best, and whether mate recognition systems precede or follow species divergences, can be discussed endlessly, but are irrelevant to a basic fact: individuals of a species can interact with others of a species for many purposes besides mating.  

In a nutshell: the unification proposed muddles our understanding of sexual selection, which should focus on the interactions of individuals of the same species.

My beef with the Padian & Horner paper

I’m not really a paleontologist, so I can’t really comment on what can and cannot be inferred from the fossil record. But as someone who really liked both papers that this one responds to, I tried to read the response as carefully as I could. I often really enjoy response papers and the discussion they start, for they often bring to light things that I ignored when first reading the original paper. So a double-whammy take at both those papers should provide a really interesting perspective and open up some nice conversation, amirite?

Well, not so much. I can’t really say I like the perspective Padian & Horner bring to the table, for a couple reasons. So here are my two cents:

Sexual selection does not require, nor does it equate to, sexual dimorphism. 

This is kind of a pet peeve of mine, and a statement that weirdly I’ve been seeing around a lot lately. So as soon as I read the first line of the paper highlights, boy I knew I was in for a treat.

Yes, sexual selection was pretty much proposed as an explanation for the gaudy ornaments that males have and females don’t. Why does the peacock have a huge tail and the peahen doesn’t? Sperm is cheap, eggs are scarce, so females are choosy and males need to show off to get the prize.

But our understanding of sexual selection has grown a lot since Darwin first proposed it about 150 years ago. We now know that while that simple idea was a breakthrough, things are much more complicated than that. Sexual selection theory has matured a lot, and in several different ways, but I’ll focus on a couple that are relevant to the discussion.

There are many ways by which you can have strong sexual selection without sexual dimorphism ever evolving. Knell and collaborators actually do a good job explaining this in their paper. First, traits may never become sex-linked, such that they are expressed in both sexes even though they only have fitness benefits to one of the sexes. This is theoretically possible, but for different reasons not considered very likely. Second, not all monomorphism is the same: you can have a species where both sexes are very cryptic and inconspicuous, but you can also have both species as showy and ornamented as you can imagine. This is particularly common in birds: if males and females are both super colorful and sing very elaborate songs, does this mean there is no sexual selection driving the evolution of these songs and colors? Probably not! It means that sexual selection is important in both sexes. Dimorphism does not indicate sexual selection; it indicates the differential balance of natural and/or sexual selection among the sexes.

“What do you mean, we have to be dimorphic?” In the cooperatively-breeding superb starling (Lamprotornis superbus), both males and females face intense competition for mating opportunities; thus, as expected from sexual selection theory, both sexes are very ornamented. (photo: flickr.com/photos/bamberry)

Ornamental traits can be used by both sexes in intrasexual competition for mating. For example, cooperative breeding systems may result in intense competition between females for breeding opportunity, such that some traits are favored in competition for mating opportunities (this is usually referred to as social selection, of which sexual selection is a subcategory). This would favor the evolution of monomorphism due to the selection for elaborately ornamented females. In monogamous species where the females have to be cryptic to incubate their eggs, the balance between natural and sexual selection for each sex is different, leading to the evolution of showy males and cryptic females. Similarly, both sexes may be choosy and use ornamental traits to select their mates.

“OK then, you can have sexual selection without sexual dimorphism, but if you have sexual dimorphism that means there’s sexual selection right?” Well… not really.  Sexual dimorphism can evolve entirely due to natural selection, if males and females have different ecological needs. Females may need specific nutrients to make eggs, or specific traits to find and occupy the best nesting sites. This idea, advocated by Wallace, has not received as much attention as sexual selection has, but can be very important in maintaining sexual dimorphism and even promoting diversification. Anoles, for example, show differences in their morphology between sexes that is unrelated to sexual selection, and maintained by ecological differences between the sexes. Hummingbirds are another potential example where the needs of females (who lay and incubate the eggs) can be quite different than males, who can become much more specialized.

So there you have it. You can have sexual selection without sexual dimorphism, and you can have sexual dimorphism without sexual selection. Can we please move along from considering these two things synonymous?

There are a couple other things that bother me in the response, but this blog post is becoming pretty much as long as their original paper, so I’m going to stop here. I just want to bring up one last thing:

On arguments from authority and ignoring discipline progress

One of the main criticism brought by Padian & Horner is that both Knell and collaborators and Mendelson & Shaw distort original – and therefore correct – definitions in their papers. Again, from their response:

Because Darwin invented sexual selection, and based it on copious observations that have never been falsified, his definition cannot be wrong.

I agree that there is abundant, unfalsified evidence for sexual selection. But there has also been a humongous body of evidence accumulating that has led us to revisit the concept and even alter it accordingly. When I read The Descent of Man and Selection in Relation to Sex, I understood that the main issue that sexual selection was meant to deal with was sexual dimorphism. But I did not at all get the impression that they were to be treated as synonymous or so strongly entwined that one could not make sense without the other. But even if it did, we have learned a lot since then. In a beautiful 1983 paper, Mary Jane West-Eberhard revisited and expanded these ideas, and developed much of what I present here. Please go ahead and read it – if I had to choose a single paper that dramatically influenced how I think about sexual selection, that would be the one.

Take natural selection for example. Though we understand the idea in the same overarching, conceptual framework as Darwin coined it, our understanding of it has changed a lot. We now have a much greater understanding of genetic inheritance, mutations, epistasis, population-level effects and so on, which have fundamentally changed our view of evolution. We study the natural selection that Darwin defined, but at the same time we don’t – much has changed. The same holds true for sexual selection. Ignoring about 150 years of research by using an interpretation of an outdated definition as the only acceptable one is not the way to move science forward.

As customary, the original authors of both papers have replies to the response by Padian & Horner in the works. I’m sure they’ll cover this and more, so I am vey much looking forward to reading those.

Further Reading

Butler, M., & Losos, J. (2002). Multivariate sexual dimorphism, sexual selection, and adaptation in Greater Antillean Anolis lizards. Ecological Monographs, 72(4), 541–559.

Butler, M. A., Sawyer, S. A., & Losos, J. B. (2007). Sexual dimorphism and adaptive radiation in Anolis lizards. Nature, 447(7141), 202–205.

Knell, R. J., Naish, D., Tomkins, J. L., & Hone, D. W. E. (2013). Sexual selection in prehistoric animals: detection and implications. Trends in Ecology & Evolution, 28(1), 38–47.

Mendelson, T. C., & Shaw, K. L. (2012). The (mis)concept of species recognition. Trends in Ecology & Evolution, 27(8), 421–427.

Padian, K., & Horner, J. R. (2013). Misconceptions of sexual selection and species recognition: a response to Knell et al. and to Mendelson and Shaw. Trends in Ecology & Evolution, Online Early.

Rubenstein, D. R., & Lovette, I. J. (2009). Reproductive skew and selection on female ornamentation in social species. Nature, 462(7274), 786–789.

West-Eberhard, M. J. (1983). Sexual selection, social competition, and speciation. Quarterly Review of Biology, 58(2), 155–183.