The limits of male attractiveness
March 3, 2011 § Leave a comment
Why aren’t males more attractive to females? Because what females want simply isn’t attainable — alas. That, at least, is the message of a recent paper (Hine et al. 2011 Natural selection stops the evolution of male attractiveness. PNAS 108 3659-3664) that examined the evolution of sexually selected traits in male Drosophila.
The authors were puzzled by the fact that although sexually selected traits — such as the wild and unnecessary tail of the peacock — are found in a lot of different species, it’s rare to spot a trait evolving (or diverging) in present-day populations. Why is this? One possibility is that many of these traits have gone as far as is physically possible to go already — for example, it’s often been suggested that the enormous antlers of the Irish Elk were so large that they contributed to the species’ demise. In this paper, Hine et al. set out to ask just how far they could push a sexually-selected trait. If they found that the evolution of this trait towards further attractiveness was blocked at some point, they wanted to know why.
The trait they chose to focus on was the production of sex pheromones called cuticular hydrocarbons in Drosophila serrata (I wrote about the possibility that the production of such pheromones may be affected by the bacteria the flies carry a few months ago, though this is not especially relevant to the current discussion). The authors set up an artificial selection scheme that involved taking 60 mating pairs, allowing them to mate, then killing the males and running their exudates on a gas chromatograph. This allowed them to detect the 8 cuticular hydrocarbons they were interested in, and determine how attractive each male ought to be, in theory, to females. They then ranked the males based on their level of expected attractiveness, based on a fit to the vector of previously estimated sexual selection gradients for these cuticular hydrocarbons. For the top 30 males, they took two male and two female progeny onto the next round. The progeny of the bottom 30 males were toast. They used two types of controls, one in which the males selected to go forward to the next round were chosen randomly, and another in which the selection was along a vector that equated to weak sexual selection.
Over the first 7 generations, this strong artificial selection indeed caused a smooth improvement in the pattern of sex pheromones produced. But after that point, they hit a plateau: they couldn’t get any further increases. After 10 generations, they stopped selection and bred the flies normally for a while. Four or five generations later, the sex pheromone pattern was almost back at the starting point. The males from generation 10 did show increased female attractiveness, as measured by greater reproductive success when they were competed with ordinary males. So here’s the puzzle: why would a male trait that offers greater girl-getting power disappear so quickly?
To answer this question, Hine et al. first wanted to know why they got a plateau after 7 generations. It could be that the generation 7 flies had simply run out of genetic variation in the expression of these pheromones, making it impossible to improve the pattern of expression further. Or it could be that the major gene affecting expression patterns is overdominant, which means that heterozygotes are much more fit than the homozygotes (as is the case, for example, with sickle cell anemia). This would mean that selection hits a wall when everyone is a heterozygote. A third explanation is that natural selection of some kind opposes the effect of the sexual selection. This would not only explain the plateauing effect, it would also explain the backsliding the authors observed at generations 14 and 15.
To distinguish among these possibilities, Hines et al. measured genetic variance in their selected lines over the period of the selection (a previous paper with an overlapping set of authors describes their method of measuring changes in genetic variance, which uses tensors to allow them to consider variation in multiple functionally related traits at once). They found that variance actually increased by at least 150% between generations 1 and 10. Now this is interesting, because it gives you a clue about the underlying genetic effects that are happening during selection. If the trait you’re selecting for is controlled by many genes, all of which make small contributions to the overall result, it’s hard to get a large increase in variance: too many things have to happen at once to make a big change in the spread of the phenotypic measurement in the population. If you have one or two genes that make a major contribution, though, increasing the frequency of these genes in the population could produce a large change in variance. Using complex segregation analysis, the authors found evidence for at least one gene with a major effect, which was present at a frequency of about 12% in control lines and increased in frequency to 30% in the artificially selected lines. This gene hits its maximum frequency at generation 7, and then plateaus — just as the pattern of attraction through cuticular hydrocarbons does. It also turns out to be partially recessive, which means that the plateauing effect cannot be solely due to overdominance.
So there seems to be good evidence against our first two possible explanations: the plateauing isn’t due to lack of genetic variance, and it isn’t entirely due to a sickle-cell-like disadvantage of being homozygous, either. What’s left is the possibility that the gene(s) that confer greater attractiveness also cause a reduction in fitness; hence the decline in attractiveness after the selection stops. It’s plausible that this might be generally true for sexually selected traits. Perhaps, because sexual selection is so strong, it quickly reaches its practical limit in relatively short evolutionary times, and gets stuck there unless something changes. For example, the environment might change (changing the male’s ability to express the trait) or the preferences of the fickle females might change. This would explain why you rarely see sexual selection happening today: it’s important, but it’s fast. You blink (evolutionarily speaking) and you miss it.
The other conclusion we are forced to is that George Clooney is an evolutionary dead end, alas. But those who are au fait with celebrity gossip will know this already.
Hine E, McGuigan K, & Blows MW (2011). Natural selection stops the evolution of male attractiveness. Proceedings of the National Academy of Sciences of the United States of America PMID: 21321197