Field of Science

The peacock's tail is not only a product of sexual selection

Oh.. dear, I'm becoming one of those annoying obsessed types - with the peacock tail. For the I don't know what time, someone is stating as fact that the peacock's tail evolved via sexual selection. Here's one Denis Dutton on TED:
The peacocks magnificent tail is the most famous example of this [sexual selection]. It did not evolve for natural survival. In fact, it goes against natural survival. No, the peacock's tail results from the mating choices made by peahens.
I've said it somewhere before, but I'll just go on and say it as many times as the editor* of this blog allows me: THE PEACOCK'S TAIL SCARES AWAY PREDATORS! (I'm told all-caps gets the message across).

Now, I admit immediately that this is not a well tested hypothesis. Here are the two pieces of evidence that I have: I have myself seen a peacock raise its tail** in San Diego Zoo when it was approached by... children. They were going too close, and as a result it raised its tail feathers (yeah, I know, correlation/causation - maybe the peacock was horny at the sight of humans its own size).

Another is a story I've heard about a peacock in a garden and two golden retrievers. The first time the dogs saw the bird, they ran barking towards it, and in response (or, again, just maybe because it was at that very moment getting all horny thinking about a certain peahen it had met earlier that morning) it raised its feathers. And the dogs cowered and retreated, and never bothered the bird again.

In addition, I have also once heard that eyes generally confuse animals, which might be why there are "eyes" on the feathers. And of course, if you can fool the predator into thinking you're the bigger one, they may just forget about the peacock meal.

The story for how the tails would increase the likelihood that a predator catches the peacock usually describes a peacock on the run, with the tail feather in the "neutral" position, and the predator grabbing on to the long feathers. But that of course only works if the peacock retreats in the first place.

Now for some testing. Next time you see a peacock somewhere (zoos, public gardens - there are many in California, but I haven't seen any here in Michigan - or did I see one in Detroit Zoo?), try to ask a child to approach it too closely and see what happens. If you're against involving children in research (why would that be?), get down on all fours yourself. Tell me what happens.

And here we damn well go:

And here is a great dramatization of a peacock scaring off a pesky dog.

And here's one interacting with humans.

Lastly, I do not at all contest that sexual selection is also involved. It is unlikely the the peacocks fan their tails at peahens in order to scare them away.

Dutton's talk, by the way, is about a Darwinian theory of beauty. Watch it here, if you must.

* Fortunately, that's also me.
** Okay, so it's also called a "train", and those feathers are not on the tail at all, but "highly elongated upper tail coverts." [Source.]


  1. Yet another demonstration of where Fodor et al almost got it right when they argued that it is inaccurate to say that natural selection "selects for" anything. It's often a rather convenient way of speaking, but natural selection operates without intention, without plan... and therefore, as we might expect, without any adherence to design principles such as modularity, etcetera.

    It's not like a human-engineered system, where part A has function X, and part B has function Y, and together maybe they have function Z. It's all just a big jumble. We can tease out some of that jumble... but it may turn out to not even be meaningful to talk about whether the peacock's tail was "for" scaring off predators vs. "for" attracting females. It might do both, and it might not even be clearly discernible which function "came first".

  2. While I don't disagree with you directly in terms of selection of/for and pleiotropy, you did break an unwritten rule on this site, which says that no good things may be said about Fodor and PP's book.

    But seriously, selection of is trivial (but must be shown, of course). Selection for can be quantified. E.g., "sexual selection increased the probability of fixation of the peacock's tail by 24%, while natural selection increased it by 17%." The two may not be independent - in fact, sexual selection for this trait may very well exist because it is an honest signal.

  3. Your final sentence is sort of what I was getting at... I suppose in principle one could quantify every single variable that was entering into the selection, and in that sense there is a "selecting for" going on... but I guess my point is that since natural selection has no agency (and even sexual selection has no overarching plan), and since the grist for non-random selection is entirely random, we would expect that the correlation between alleles and various manifestations of fitness would be terribly haphazard. Obviously I don't need to tell you that, Mr. My-blog-is-called-Pleiotropy...

    Let's say in the year 5673 BC, one peacock was saved from certain death when his tail got caught in some branches, and that slowed him down enough that an impending avalanche that would otherwise have hit him now missed him -- whereas his brother with a slightly smaller tail continued unimpeded and was killed in the avalanche. Do we now have to include in our analysis a 10^-10% contribution to the selection of the tail by "stops peacocks from walking into the path of avalanches"? If not, why not? (Without invoking agency, that is)

    I have not read Fodor et al's book, but it seems to me their mistake was in misunderstanding the very nature of abstraction... Sure, it's fine to point out that whenever we say, "Trait X was selected for because..." we are necessarily making an abstraction. The problem is in objecting to the act of abstraction, because without abstraction all human inquiry must cease.

  4. Hmm. I think the deal is that if this avalanche event where the large tail helps the peacock escape is one that occurs often enough, then we can definitely say that peacock tails are selected for avalanche avoidance.

    But if it's just one time, then we attribute it to a random event. I guess there is then a gray area where I would not readily be able to decide which it is.

    And thanks for sharing the example. Very thought-provoking.

  5. So just continuing the line of thought (I'm making this up as I go along, of course) I think there is an analogy here between the "sum over histories" that is used in quantum mechanics. In principle, to make an "accurate" prediction in QM you need to sum the waveform magnitude of all possible paths that the particles in question could take through the universe -- and really, all possible paths that all particles in the universe could take, I suppose. But in practice, except for a few particular paths, all the rest of them cancel each other out, so you don't actually have to sum them (thank goodness!)

    In much the same way, we can probably infer that for every peacock that was saved from an avalanche by getting stuck in the bushes, an approximately equal number were killed in avalanches because they got stuck in the bushes. Even though in reality the fitness contribution of a particular phenotype is the sum of all reproductively significant events that ever had a causal relationship with the phenotype, even apparently freak events like the one in my thought experiment, it is (usually) a safe assumption that the vast majority of them cancel each other it.

    Of course, those new to QM sometimes fail to sum all of the relevant paths, and in the same way of course evolutionary biologists may sometimes assume some random events (like tail-stuck avalanche protection) get cancelled out when in fact they don't.

    So there's the lesson, I guess: Whenever biologists say something is "selected for", they're doing something kind of like quantum physicists do when they simplify their sum-over-histories to only include the relevant paths a particle can take. It's an abstraction where countless terms in the equation are simply discarded on the assumption that they cancel out. This I think is a fair abstraction (as I said, without abstraction all human inquiry must cease) but I suppose it's good to do so with eyes wide open.

    Probably evolutionary biologists have already been pretty aware of this, but for me as a layperson, I did get some value from reading all the negative reviews of Fodor et al's book, in that it was sort of thought-provoking in terms of the disconnect between the "selected for" abstraction and the messy way in which selection actually works.

    I think I'm just rambling at this point :)

  6. Without commenting on the QM analogy, I think it is a fair assumption to make that events like these cancel out. I don't think many biologists ever think about this, though.

  7. if the peacock tail is indeed a means to both ward away threats, by perceived size or the "eyes" - the very big-ness of the tail could easily correlate to attraction via the perceived impressiveness by appearing to the peahen as a more effective warder and therefore more "fit" in the scheme of survival. Simple extrapolation.

    Why to humans desire features of the other gender? Like large breasts or big muscles, or a towering intellect? breasts = fertility, muscles = a capable protector. Its easy to forget these fundamental things in the social fad or style of appearance and sexuality or just desire to have sex - but true all the same.

  8. So why don't peahens have big tails? Don't they have predators too?

    1. That's a good point. Perhaps the females respond differently to predators. I consider it most likely that the large trains of the males work both to scare off predators and to attract females.


Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="">FoS</a> = FoS