Friday, February 10, 2012

How Did the Zebra Get Its Stripes?

We've been discussing the adaptationist approach to biology on another thread and this is a good example to illustrate the issues. If I were to ask you how the zebra got its stripes, what would you think?

Would you immediately assume that it could be an evolutionary accident with no adaptive significance then start to wonder if you could rule out such an explanation? Can random genetic drift of neutral alleles explain the zebra's stripes?

Or would you immediately start thinking of adaptive explanations for why all three extant species of zebras have stripes but no other large mammals in the same environment are striped. Most other horses don't have prominent stripes but many have faint stripes on some parts of their bodies (Darwin, 1859).

I argue that you have to rule out the null hypothesis (drift) before invoking adaptationist explanations. In other words, the first question you need to ask is whether zebra stripes are adaptive. But that's not the adaptationist approach. Adaptationists begin with the assumption that stripes are adaptive, then they start looking for adaptive explanations.

What if the favorite adaptive explanation is refuted? What does an adaptationist do next? Gould and Lewontin (1978) provide the answer ...
We would not object so strenuously to the adaptationist programme if its invocation, in any particular case, could lead in principle to its rejection for want of evidence. We might still view it as restrictive and object to its status as an argument of first choice. But if it could be dismissed after failing some explicit test, then alternatives would get their chance. Unfortunately, a common procedure among evolutionists does not allow such definable rejection for two reasons. First, the rejection of one adaptive story usually leads to its replacement by another, rather than to a suspicion that a different kind of explanation might be required. Since the range of adaptive stories is as wide as our minds are fertile, new stones can always be postulated. And if a story is not immediately available, one can always plead temporary ignorance and trust that it will be forthcoming .... Secondly, the criteria for acceptance of a story are so loose that many pass without proper confirmation. Often, evolutionists use consistency with natural selection as the sole criterion and consider their work done when they concoct a plausible story. But plausible stories can always be told. The key to historical research lies in devising criteria to identify proper explanations among the substantial set of plausible pathways to any modern result.
Let's look at the history of this question as summarized in a recent paper by Egri et al. (2012).
The most characteristic aspects of zebras are the bold black-and-white striped patterns on their body surface (Fig.1). Embryological evidence (Prothero and Schoch, 2003) has shown that the background colour of zebras is black, and the white stripes and bellies (where the production of dark pigmentation is inhibited) appear only in a later embryonic developmental stage. The reason for the striped coat pattern in zebras has long been debated, and Wallace suggested that zebras evolved striped coats as camouflage against carnivores in tall grass (Wallace, 1867; Wallace, 1879). Darwin, however, who had closely studied the inheritance of colours and stripes in horses and zebras, criticized this hypothesis as an explanation (Darwin, 1871), as zebras do not occur in areas with dense vegetation but rather prefer open savannah habitats with short grass.

Since the 19th century, a number of alternative hypotheses (Waage, 1981; Ruxton, 2002; Lehane, 2005; Caro, 2009) have been proposed to explain the striped pattern of zebras, including predator defence, social interaction, indication of physical condition, thermoregulation, and protection from tsetse flies (a more detailed account is given in the Appendix). These and more explanations have been thoroughly discussed and criticized by Ruxton (Ruxton, 2002) and Caro (Caro, 2009), who concluded that the majority of these hypotheses are experimentally unconfirmed, and thus the exact cause of stripes in zebras remains unknown. Nevertheless, the explanation of Waage (Waage, 1981) for the benefit of zebra stripes (i.e. protection from tsetse flies) has been the only hypothesis to be partially supported experimentally (Turner and Invest, 1973; Brady and Shereni, 1988; Gibson, 1992; Ruxton, 2002; Lehane, 2005; Caro, 2009).
Isn't that interesting? The question has been around for 145 years and nobody has suggested the null hypothesis. It certainly hasn't been ruled out. Why do adaptationist explanations continue to be so popular if most of them have been discredited?

Egri et al. propose that the stripes on zebras serve to protect them from horseflies (tabanid flies). They carried out several experiments to measure whether horseflies preferred black, white, or striped backgrounds. The results indicate that horseflies tend to avoid landing sites that have lots of stripes. In one experiment the authors measured the number of horseflies that were attracted to full-sized models of zebras that were black, brown, white, or striped. (The photo shows a spotted model but a white one was used in the reported experiment.)


The result was remarkable. The black and brown models attracted 562 and 334 horseflies over 60 days while the white models only attracted 22 horseflies. Only 8 horseflies landed on the striped model in 60 days! The authors conclude ...
We have shown here that the evolution of stripes with brightness and/or polarization modulations disrupting the homogeneous pattern of reflected light might be a selective advantage in avoiding attacks from polarotactic tabanids. We conclude that zebras have evolved a coat pattern in which the stripes are narrow enough to ensure a minimum attractiveness to tabanid flies.
There are a lot of steps between showing that flies avoid striped patterns in field experiments and concluding that stripes confer a selective advantage because there ain't so many flies on zebras. Egri et al. have constructed a plausible story that's consistent with adaptation but they have not demonstrated that stripes confer a selective advantage.

The author of the BBC news article (Victoria Gill) sought the opinion of Matthew Cobb who thinks that none of the adaptive explanations is satisfactory (but doesn't entertain any non-adaptive explanations).
Prof Matthew Cobb, an evolutionary biologist from the University of Manchester pointed out that the experiment was "rigorous and fascinating" but did not exclude the other hypotheses about the origin of zebras' stripes.

"Above all, for this explanation to be true, the authors would have to show that tabanid fly bites are a major selection pressure on zebras, but not on horses and donkeys found elsewhere in the world... none of which are stripy," he told BBC Nature.

"[They] recognise this in their study, and my hunch is that there is not a single explanation and that many factors are involved in the zebra's stripes.


[Hat Tip: RichardDawkins.net [Zebra stripes evolved to keep biting flies at bay] who copied a BBC news article.]

Darwin, C. (1859) On the origin of species (Chapter 5).

Egri, A., Blahó, M., Kriska, G., Farkas, R., Gyurkovszky, M., Åkesson, S., and Horváth, G. (2012) Polarotactic tabanids find striped patterns with brightness and/or polarization modulation least attractive: an advantage of zebra stripes. J. Exp. Biol. 215:736-745. [doi: 10.1242/jeb.065540]

Gould, S.J., and Lewontin, R.C. (1979) The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adaptationist Programme. Proc. R. Soc. Lon. B Biol. Sci. 205: 581-598. [PubMed] [doi: 10.1098/rspb.1979.0086]



25 comments :

  1. This post is interesting to me because like so many people I have a tendency to go for adaptationist ideas about this feature or that, but I know enough to know this is a bad idea. But how would you suggest going about proving -- as far as we prove -- any idea about a feature, for instance these stripes?

    While I'm fine with "we don't know" being preferable to "we know something which isn't true" or even "we know something which may not be true", I hope we're trying to work toward "we know". How can this happen?

    For instance, in this case there's a suggestion that flies elsewhere "tabanid fly bites are a major selection pressure on zebras, but not on horses and donkeys found elsewhere in the world" would do it, but would it? It's possible that's so, of course, and that could/should be tested, but would it answer the question? Isn't it so that we often find different species end up, via chance, following different paths to solve essentially the same problem, even if the correct answer is adaptationist?

    And do you ever "prove" the neutral, the null?

    ReplyDelete
  2. Just because nonadaptive hypotheses are often neglected where they should be considered, doesn't mean that this has happened in this specific case. The zebra stripes are a poor candidate for being the product of nonadaptive explanations -- e.g. genetic drift, or being a spandrel for some other adaptation.

    E.g.:

    1. The stripes are basically universal in zebras, if it was drift you would expect more diversity. This is not like, say, human sacrifice or homosexuality, where people attempt adaptive explanations even though the behaviors in question are very far from universal in the population.

    2. The phenotype is big, obvious to all (humans and lions alike), and on a feature (coat color) that is plausibly of adaptive significance in many, many mammals, for camouflage, display, etc. The weird thing about zebras is that their pattern is close to unique among mammals, and it seems to make them more visible, not less -- although, if you read the articles that Egri et al. cite, you will see that the stripes may actually serve as camouflage in twilight and at night, and maybe hiding is pointless when you are a fast herbivore living on huge short grass expanses.

    3. The third argument would be that stripes seem like they are a developmentally complex pattern to produce, which would be the product of a number of improbable mutations working together, which would presumably require selection to assemble. I am less sure of this one, because I just coauthored a paper on the evolution of patterns on cone snail shells, and in that system, just tweaking a few parameters of the developmental model can suddenly shift you from spots to stripes to blank to even more complex patterns. The 500 cone snail species have a huge diversity of shell patterns though, and it's not even clear if the pattern is visually obvious in many living specimens (many are covered by an opaque sheath, and/or algae etc.). So there is a case where drift really could be an explanation, for all we know.

    ReplyDelete
    Replies
    1. The stripes are basically universal in zebras, if it was drift you would expect more diversity.

      Why? Alleles can be fixed by random genetic drift. Humans have thousands of neutral alleles that have become fixed in the population since we diverged from our common ancestor with chimps.

      The phenotype is big, obvious to all (humans and lions alike), and on a feature (coat color) that is plausibly of adaptive significance in many, many mammals, for camouflage, display, etc.

      Some of us think that big obvious phenotypes can be neutral. What we need is data on how many are actually adaptive and how many are due to alleles fixed by random genetic drift. As long as there's a general bias out there that visible phenotypes must be under selection, we'll never get the data we need.

      The third argument would be that stripes seem like they are a developmentally complex pattern to produce, which would be the product of a number of improbable mutations working together, which would presumably require selection to assemble.

      Gould argues that the stripes can be produced by relatively small changes, Most developmental biologist would agree. Michael Lynch argues persuasively that complexity can just as easily arise by nonadaptive processes.

      "Where, then, is the direct supportive evidence for the assumption that complexity is entirely rooted in adaptive processes? No existing observations support such a claim, and given the massive global dominance of unicellular species over multicellular eukaryotes, in terms of both species richness and numbers of individuals, one can only marvel at the inability of natural selection to promote it." (p. 378)

      Delete
    2. Why? Alleles can be fixed by random genetic drift. Humans have thousands of neutral alleles that have become fixed in the population since we diverged from our common ancestor with chimps.

      But in the genetic case you can look at the frequency spectrum for "fixed" alleles at loci, and see that neutrally evolving regions have a much broader diversity and higher frequency of SNPs than regions under selective sweeps or stabilizing selection. The analogy to phenotype is inexact, but I think the fact that we see a lot more diversity in e.g. wild horses is indicative of what one might expect if neutrality and drift were the dominant processes.

      Gould argues that the stripes can be produced by relatively small changes, Most developmental biologist would agree. Michael Lynch argues persuasively that complexity can just as easily arise by nonadaptive processes.

      There is complexity and there is complexity. Complexity like two gene duplicates dividing up a function, or like introns dividing up exons, could plausibly be due to neutral drift/fixation of slightly deleterious alleles. Complexity like the Venus flytrap's trap, or eyes, or (not quite as strongly but pretty strongly) zebra stripes, not so much.

      Some of us think that big obvious phenotypes can be neutral. What we need is data on how many are actually adaptive and how many are due to alleles fixed by random genetic drift. As long as there's a general bias out there that visible phenotypes must be under selection, we'll never get the data we need.

      It would be nice to have popgen data on the alleles controlling coat color in horses, zebras, and relatives. It would also be nice to have a massive controlled experiment in the Serengeti where you take 5000 zebras and paint 1000 black, 1000 white, 1000 black and white spots, 1000 wildebeest colors, and 1000 you paint with the same black and white stripes as the originally had as controls. Then you put them all in 10,000 acre pens, each with a pride of lions, and see what happens.

      Delete
    3. But just because we don't have these data yet, doesn't mean that we have to pretend we have no information at all on which to base a judgment. We know that for many mammals coat color and camouflage are important for avoiding predation, at least at certain stages of life or in certain situations. We know that coat color can change adaptively, e.g. in Hopi Hoekstra's mice. We have innumerable other examples of camouflage, mimicry, disruptive coloration, predator-confusing strategies, etc., throughout the animal kingdom. We know that our vision systems are not wildly, radically different from those of e.g. lions. We know that zebras stick out like sore thumbs in the broad daylight, but we also know that they inhabit shortgrass savannas where they would be pretty visible anyway. But we also know the stripes make them indistinct at dusk and night, and that tracking an individual zebra in a herd of running zebra seems to be much harder than if they all looked different.

      And, we know that zebras have pretty darn fixed coat colors, whereas other horses show more variability. And we know, or at least there is a high probability, that stripyness is the ancestral condition for the horse group or at least certain big clades therein, and it has been lost in certain lineages. If anything, this argues that horses in lion-free regions have drifted away from the One True Stripetastic pattern that zebras still have.

      None of this creates certainty, but if one is trying to make a "best guess" about what is going on, drift is not, and should not be, close to the top of the list when trying to explain zebra stripes. Arbitrary declaration of a null hypothesis != evidence.

      But you are correct in this: if one is doing a review of all the possible explanations, as those review articles do, the nonadaptive hypotheses should be included. They are at least as good as the camouflage-from-flies idea, which has all kinds of problems, even though because flies are easier to do experiments with than lions are, there is "more" "experimental support" for the flies hypothesis.

      Delete
    4. You actually don't "know" as much as you think. Many of the things you're thinking of have not been proven. They are adaptive just-so stories.

      Delete
    5. Larry,

      I am interested to see how far you will go with this argument. A great proportion of Arctic birds and mammals are white (or white in winter), whereas a very slight proportion of non-Arctic birds and mammals are white. Do you think drift is the best explanation for these observations?

      John Werner

      Delete
    6. I think there are good adaptive explanations for a lot of things. That's not the point. The fact that you ask this question indicates that you don't understand the issue.

      Pluralism isn't about attributing everything to drift.

      Delete
    7. I may just be a dumb glorified high school teacher, but I understand the issue just fine. I am a pluralist. I accept that natural selection is just one of many mechanisms driving evolution, that most genetic change has nothing to do with selection, that most of the genome of eukaryotes is junk, and that selection may not be necessary to explain some aspects of visible morphology. Further, even alleles influencing adaptive traits are subject to the vagaries of drift. I get it.

      I agree with you most of the time when you argue with extreme adaptationist viewpoints. Like Nick Matzke, I just think you have overstepped in this case. In the one last open grassland environment with visually adept cursorial carnivores, we see great numbers of striped horses and nary an all-black steed. We know that horses can have coat patterns other than striped, that zebras can at least get mutations that disrupt the striping, and that stripes have been around in zebras for some time and are a fixed trait for at least three species. Further, to at least our mammalian eye and brain, a fleeing zebra herd is disruptive and confusing.

      To argue that zebra striping to be nonadaptive, you would have to say that a mutation that disabled striping would have just as good a chance to drift to fixation as a nucleotide substitution in a pseudogene. Do you really think that is the case? Take out the lions and hyenas (which might not be too far down the road, unfortunately), and I might agree with you. Maybe the case hasn't been proven rigorously; has anyone painted polar bear cubs black? Until an adaptive scenario proves unnecessary as a test result, I think there is enough circumstantial evidence in each of these cases to provisionally infer some adaptatiive value to coat color and pattern in each case.

      JW

      Delete
  3. If you were to ask me how the zebra got his stripes, I would look neither to adaptive nor non-adaptive answers, but to molecular mechanisms. (Non-)adaptivity is not related to how the zebra got his stripes, but to why.

    A pedantic point perhaps, but without a firm grasp on the mechanism, we're unlikely ever to be able to rigorously answer the question of adaptivity.

    ReplyDelete
    Replies
    1. Let's assume, for the sake of argument, that we know the mechanism by which genes for white hair are turned on at specific steps during development. Gould argues in "How the Zebra Gets Its Stripes" (Hen's Teeth and Horse's Toes p. 366) that the differences between the various zebra species are almost certainly due minor changes in the timing of stripe formation. That means that fixation of those alleles by random genetic drift is quite plausible.

      Gould also argues that the ability to form stripes seems to be present in all horse species although it is only obvious in zebras. It suggests that the difference between zebras and the ancestor of all horses may only require a few changes in expression.

      Knowing the exact mechanism isn't going to help us decide how the alleles were fixed in the population.

      Delete
    2. No genes for white hair ON but genes for color OFF.

      The stripes are standing waves from a Meinhardt model. Tiger stripes, leopard spots are other instances of the same. It seems a general mammalian model for fur patterning, that can drift or be tuned to adaptation.

      Meinhardt model: see: http://www.amazon.com/Algorithmic-Beauty-Sea-Shells/dp/3540440100

      Delete
    3. An excellent article here:
      http://www.scholarpedia.org/article/Gierer-Meinhardt_model

      Delete
    4. What is the difference between a Meinhardt model and a Turing model? (A quick Google search didn't turn up anything that explicitly stated what it was.)

      Delete
    5. (Note: Peter @10-02-2012 12:46 PM is not the same as Peter @10-02-2012 10:12 AM)

      While knowing the mechanism won't definitely tell us how the alleles were fixed, it can certainly help. We can look for signatures of positive selection such as rates of synonymous and nonsynonymous change in the coding sequence, or the rates of substitution in the gene promoter region relative to neighbouring intergenic sequence. In each case you have a molecular proxy for the rate of neutral variant fixation (i.e. the rate of synonymous or intergenic change) and can compare the rate of functional change to that.

      Of course, evidence of positive selection doesn't necessarily imply that the coat pattern is the relevant adaptive factor - the gene could have been under selection for a different function and the coat colour just be a side effect. However, it's a sight better than no evidence of adaptation.

      How would you test whether zebra stripes are adaptive? Short of impossibly large field trials with huge herds of endangered animals, what evidence would you accept? Aren't you in danger of putting forward an unfalsifiable null hypothesis?

      Delete
    6. @Kele
      http://www.scholarpedia.org/article/Gierer-Meinhardt_model

      Gould argues in "How the Zebra Gets Its Stripes" (Hen's Teeth and Horse's Toes p. 366) that the differences between the various zebra species are almost certainly due minor changes in the timing of stripe formation.
      The original study on zebra striping (that Gould should be or is citing) is:
      Bard JBL, 1977. Unity underlying different zebra striping patterns. Journal of Zoology 183: 527-539

      Peter 2

      Delete
  4. The stripes are basically universal in zebras, if it was drift you would expect more diversity.


    Besides Larry's answer, the stripes of zebras aren't so universal as one might expect for an adaptive feature. That is, they vary a lot -- even before you get to quaggas.

    Just to add another possibility to the mix in terms of how do you prove something like this, what if it is an adaptive feature, but no longer? How the heck would we ever show that was, or wasn't, the correct answer? If it arose via selection for very good adaptationist reasons that we could prove if we were doing so, let's say, 30,000 years ago. But due to some selection that no longer exists for them.

    There's also sexual selection as a possibility, since after all the various species vary in the nature of their stripes. But they also vary in other things, like temperament (just read the other day about the difference in tameability between mountain zebras and Burchell's zebras -- Burchell's are much easier, apparently), so even if sexual selection was a big factor in speciation, stripes may have had nothing to do with that.

    I think this is a really good illustration of how hard it is to demonstrate rather than just claim selection, even when it seems like something that "should" be due to selection. That gets back to the questions I asked in my first comment, about how on earth we can figure out these things one way or the other.

    ReplyDelete
  5. While I agree that drift should have been discussed and discarded by these authors, it does seem obvious that the stripes are not neutral in an environment filled with visually-guided predators, conspecifics, and ectoparasites. Yes, it may be hard to identify the most important drivers of the pattern, but this is only a practical difficulty, not a fundamental one. Arguing that the pattern in all three zebras is due to drift sounds purely contrarian. It reminds me of the way some people argued that the classical Bistula moth experiments did not prove the adaptive function of "industrial melanism". The critics were right, the experiments could have been improved (and the new improved experiments just published confirm the adaptive value of moth color). But anyone who has watched moths and birds interacting in the wild (or even anyone just looking at the extraordinary cryptic patterns of moths) could see that these colors were adaptive.

    ReplyDelete
  6. I find the pattern of vertical stripes on the neck and body, and the horizontal stripes on the legs and rear quarters, interesting. The patterns could accentuate some aspect of anatomy that attracts females and intimidate other males.

    ReplyDelete
    Replies
    1. Two sets of reaction-diffusion equations laying down the patterning in leg and rump, at right angles to each other, is sufficient to give the zebra backhand pattern. That is, if striping on rump and legs is independent in pattern, the joint will look like that.

      Murray, JD., 1981, A prepattern formation mechanism for animal coat markings. Journal of Theoretical Biology 88: 161-169

      Given that zebras, asses and horses all are striped to some degree, striping in equids might well be old.

      The basic question is whether the color patterning im mammals has to use a reaction diffusion process, and if so, whether striping is an unavoidable possibility of the system. Otherwise it is just parameter tuning to get particular patterns.

      Peter 2

      Delete
  7. Lou - actually, new data has just been published on Biston betularia, the peppered moth. Jerry Coyne reported on the study on his not-blog website.

    Dr. Moran - knowing the exact mechanism of the color pattern might not help directly determine which evolutionary mechanism was responsible for zebra stripes, but if the genes could be identified, population genetic models could be applied to look for differences in signatures associated with drift vs. selection.

    ReplyDelete
  8. jaxkayaker, yes, that study is the one I was referring to when I mentioned the new improved experiments. Jerry Coyne's discussion is at
    http://whyevolutionistrue.wordpress.com/2012/02/10/why-do-zebras-have-stripes/

    ReplyDelete
  9. You actually don't "know" as much as you think. Many of the things you're thinking of have not been proven. They are adaptive just-so stories.

    All hypotheses of adaptation have to be “proven”, or they are nothing but “just-so stories”. The “null hypothesis” for phenotypic evolution is always random genetic drift and who cares what people who actually know something about (say) animals think? Gould once argued X (never mind that his opinion was purely speculative, hardly “proven”), case closed.

    Dr. Moran could save a lot of time by just cutting and pasting the above quote as a stock response. Wait, he should append “sez Gould” to the end. Better.

    ReplyDelete
    Replies
    1. When discussing something like colors and patterns, there are lots of speculations about their possible adaptive significance. Most of the people who speculate have not considered non-adaptive explanations.

      When defending the idea that coloration and pattern have to be adaptive, it it isn't helpful (or logical) to quote those speculations as though they were facts.

      What should be the null hypothesis? Lots of people who claim to be experts in animal evolution act as though it should be adaptation (adaptationist program). It's true that I don't care what those so-called experts in animal evolution think about evolutionary theory because I think they're wrong and I've read lots of articles defending the idea that the null hypothesis should be random genetic drift of neutral alleles.

      If the so-called experts in animal evolution want to defend their assumption then, by all means, let them speak up and defend it. There are plenty of molecular evolutionists, paleontologists, bacteriologists, population geneticists, and philosophers who will be happy to take them on.

      Delete
  10. Is there a possible answer to the question of zebras and stripes in sexual selection? Or is the null hypothesis that the stripes, when first appearing, didn't confer a survival disadvantage and thus the stripes continue to be expressed through the generations, which is is what I think you mean?

    It seems to me that an adaptationist approach in this case, is a search for "purpose," when purpose doesn't necessarily exist.

    ReplyDelete