Why do zebras have stripes?

Well, I can’t give a definitive answer to that question, but I can start by telling you that zebras are not white with black stripes, but black with white stripes. The ground color in embryos is black, and the white stripes appear later in development in areas where the deposition of melanin pigment is inhibited.

But of course you’re wondering why they have stripes at all.  Various hypotheses have been suggested, the most famous being camouflage, confusion of predators, or even thermoregulation.  None of these have been experimentally supported, for it’s hard to test them. (The camouflage story, though, doesn’t hold up since it depends on zebras hiding in tall grass, but they live on the savanna where tall grass is almost nonexistent.)

There is, however, a new hypothesis that has garnered some experimental support: it’s presented and tested in a new paper in the Journal of Experimental Biology by Ádám Egri et al. This is a rather long and complicated paper, and perhaps I’d best refer you to the BBC report of the findings, which contains a quote from our own Matthew Cobb.

In short, experiments on Hungarian horse farms using model horses and other targets for flies suggest that the striped pattern reduces the kind of polarized light that attracts blood-sucking tabanid flies (“horseflies”). Those flies are attracted to horizontally polarized light, presumably because that’s a sign of water, and water is where female tabanids lay their eggs.  Model horses and pans of fly-attracting oil painted with a dark color are very attractive to flies; white models less so.  Surprisingly, striped models attracted the fewest flies, and the stripe width that was the most deterrent was that actually present on the heads and legs of all three existing species of zebras.

The authors theorize that the adaptive significance of the stripes is to deter tabanids in Africa, because fly bites can reduce grazing, and hence survival and fertility.

It’s a reasonable theory, but has a few problems.  First, there’s some special pleading about the width of stripes: as I said, the most deterrent widths are those found on the zebra’s heads and legs, not on their backs, where the stripes are wider (see below). To explain this, the authors argue that the heads are important for survival because they contain vital sensory organs, and the legs are critical because they “are indispensable to escape from predators.” (The authors also claim the blood vessels are closest to the skin in these places.) But that doesn’t explain why the stripes aren’t the same fly-deterring width all over the body.

Here’s a Grant’s Zebra  (Equus quagga boehmii, a subspecies of the Plains zebra).  Note that the stripes are narrowest on the head and legs:

Matthew Cobb, quoted in the BBC piece, raises another problem:

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.

Indeed!  To add to that, there’s another fly that could be a serious pest on zebras in Africa: the tsetse fly, although the authors claim that tsetses don’t bite zebras as often as they do other mammals. But zebras are susceptible to the trypanosome parasite that causes human sleeping sickness (and makes horses chronically ill), and the authors didn’t test whether stripes deterred tsetse flies. (The experiments were, after all, done in Hungary where the tsetse doesn’t live.)

I find the authors’ hypothesis intriguing and the data somewhat convincing, but there could be other reasons for stripes that we just don’t understand.  And, as Matthew noted above, we also need to explain why most species of equids lost their stripes, because it’s pretty clear that the ancestral equid was striped. (Darwin noted this in The Origin when he pointed to the existence of occasional stripes in domesticated horses as evidence for the occasional re-expression of an ancestral trait). Do unstriped equids not encounter horseflies? Or was a stripey pattern too conspicuous to predators?

And someone needs to investigate this problem in felids.


Egri, A. M. B., G/ Kriska, R. Farkas, M. Gyurkovszky, S. Åkesson and, and G. Horváth. 2012. Polarotactic tabanids find striped patterns with brightness and/or polarization modulation least attractive: an advantage of zebra stripes. J. Experimental Biol. 215:736-745.


  1. daveau
    Posted February 10, 2012 at 9:47 am | Permalink

    Oh, look, a tabby…

    Really interesting post. I cannot tell you how much I learn from this site. Thank you.

    • daveau
      Posted February 10, 2012 at 10:24 am | Permalink

      I had always unscientifically assumed it was camouflage, but now I’m curious as to how this might apply to Okapi, or even if there is a reationship to the spots on a Girrafe. Could they all be expressions of similar selection pressures?

      • truthspeaker
        Posted February 10, 2012 at 11:35 am | Permalink

        I, too, had always assumed it was for camouflage.

        • GBJames
          Posted February 10, 2012 at 11:43 am | Permalink

          Camouflage from flies!

          • neil
            Posted February 10, 2012 at 12:16 pm | Permalink

            If it works, I will wear a black and white striped shirt the next time I hike in the mountains.

          • Filippo
            Posted February 11, 2012 at 4:37 pm | Permalink

            A flea and a fly
            Were caught in a flue,
            So what on Earth
            Were they going to do?

            Said the flea,” Let us fly!”
            Said the fly, “Let us flee!”

            So they flew thru a flaw in the flue.

        • Asura
          Posted February 10, 2012 at 4:41 pm | Permalink

          It’s amazing just how many assumptions people make on these sorts of things; namely adaptive selection.

          Jerry at Sandwalk talks about the same study with that as a focus.

  2. Posted February 10, 2012 at 9:48 am | Permalink

    I have a hypothesis about why the stripes are different sizes on different parts of the body. The head of an equine fetus, like the head of most fetuses, is larger in proportion to the body than in an adult. The proportions actually reverse with development. It makes sense, it seems to me, that if the striped portions of the body started out equidistant from each other, as the animal developed, the stripes on the body would get wider, and the stripes on the head would stay narrower.

    What do you think, Jerry? Could I be onto something?

    • qbsmd
      Posted February 11, 2012 at 11:19 pm | Permalink

      Implying the effects of the stripes are most important for young zebras? I wonder if anyone’s done an experiment on health complications from fly bites on young vs. mature zebras.

      • Diane G.
        Posted February 12, 2012 at 11:34 pm | Permalink

        Now this seems like an excellent hypothesis!

    • Posted February 17, 2012 at 12:08 pm | Permalink

      I think I read this somewhere in Gould.

      • Posted February 17, 2012 at 1:00 pm | Permalink

        Cool, I’ve never read Gould. This is not even my field. I’m a graphic designer.

  3. Posted February 10, 2012 at 9:49 am | Permalink

    This would explain why other striped animals show the same kind of small striped head and large striped body as Zebras e.g. tigers, domesticated cats, etc.

  4. Dominic
    Posted February 10, 2012 at 9:49 am | Permalink

    Is the trypanosome parasite one that has spread from one original host, or is it happy with any host? I am think of the way diseases hop to species in close contact – I suppose some parasites are less picky than others? It is fascinating the way that flies drive caribou so potty that they migrate to escape them. Do flies play any part in zebra migration, or are they just after new pasture?

    • Dominic
      Posted February 10, 2012 at 9:51 am | Permalink

      PS I think sexual selection may be more important in the stripe development, but that is just a guess!

      • Chris Granger
        Posted February 10, 2012 at 9:54 am | Permalink

        I agree with that possibility as well.

  5. Posted February 10, 2012 at 9:52 am | Permalink

    most species of equids lost their stripes

    So… quaggas were losing their stripes? I’d always thought of them as “almost a zebra” rather than “almost a horse” … 


  6. Chris Granger
    Posted February 10, 2012 at 9:53 am | Permalink

    Not all camouflage needs to make the camouflaged thing blend in perfectly with its surroundings. See military ship camouflage for example. It’s possible that the larger belly stripes do deceive the eyes of predators somewhat. Just a thought.

    • Posted February 10, 2012 at 10:05 am | Permalink

      And a good one. But not original! See Motion dazzle and camouflage as distinct anti-predator defenses
      by Martin Stevens, W Tom L Searle1, Jenny E Seymour, Kate LA Marshall1 and Graeme D Ruxton (via Wikipedia).


      • Posted February 10, 2012 at 10:08 am | Permalink

        * nix those trailing “1”s :-/

      • Chris Granger
        Posted February 10, 2012 at 10:09 am | Permalink

        Excellent, and more relevant to zebras. Thank you!

    • DrBrydon
      Posted February 10, 2012 at 11:12 am | Permalink

      Yes, came here to say this. Unlike land vehicle or aircraft camoflage, most ship camoflage in WWI and WWII was designed to make the shape/position of the ship harder to determine (which probably worked even better from several thousand yards away). This would make it harder for the enemy to aim. I would think this could apply to animal strips as well, since their must be better and poorer places to latch onto a zebra you are trying to bring down.

  7. GBJames
    Posted February 10, 2012 at 9:57 am | Permalink

    I can see it coming… whole new lines of zebra-striped anti-fly-bite camping and hiking gear.

    • GBJames
      Posted February 10, 2012 at 9:57 am | Permalink

      (and this time… the check box)

      • Posted February 10, 2012 at 9:59 am | Permalink

        Chequered zebras? :-O


        • Chris Booth
          Posted February 10, 2012 at 10:42 pm | Permalink

          That’s worse. Checkered zebras attract clowns. [shudders]

  8. s
    Posted February 10, 2012 at 9:59 am | Permalink

    I always thought the camouflage idea was that they blended, not with their surroundings, but with each other. Making it harder to detect individuals in groups?

    • Boko999
      Posted February 11, 2012 at 5:22 am | Permalink


    • Posted February 17, 2012 at 12:09 pm | Permalink

      I think I read this somewhere in Gould as well.

      • Diane G.
        Posted February 18, 2012 at 2:19 am | Permalink

        What can’t be found in Gould?


        (Besides gnuism, of course…)

  9. Posted February 10, 2012 at 10:11 am | Permalink

    I don’t remember if it was here or over in the squidly domain, but I seem to recall sometime in the past several months a post on a certain type of developmental lines along which all sorts of significant features get arranged. And zebra stripes seem to me to be similar to the pattern I remember of those lines.

    If so, that would explain a lot. Perhaps the selection pressure that caused this particular expression is related to parasites, but it could well be that the mechanism of the expression only permits the observed patterns.

    Any real biologists care to interpret what I think I’m trying to get at?



    • Posted February 10, 2012 at 10:34 am | Permalink

      I remember reading about striped and spotted cats and why you can have a spotted cat with a striped (ringed) tail but not vice versa. (I thought it was in Steve Jones’s The Language of Genes, but riffling through it quickly, I can’t find the discussion… annoying!)

      Anyway, that might be a related phenomenon… 


      • Posted February 10, 2012 at 10:34 am | Permalink

        * italics fail!

      • troy
        Posted February 10, 2012 at 10:44 am | Permalink

        I think you are referring to the work of JD Murray, who showed this with reaction diffusion models of development, based on Turing’s original pattern formation work.

        Here is an interesting link to that stuff.

        • Posted February 10, 2012 at 10:59 am | Permalink

          Those figures look very familiar! Thanks a lot.

          Oh — and that talks about zebras too.

          (I still wish I could find/remember where I read about, though!)


        • jwarton
          Posted February 12, 2012 at 2:38 pm | Permalink

          First off, thanks for this comment and link. One of the truly meaningful responses to this post in my opinion. I’d also like to preface my comments with the fact that I’m an architect and not in any way trained as a biologist, and I personally think there is far too much dependence on natural selection as a mechanism for understanding biological development.

          It seems to me that an understanding of physics must come before selection criteria , and since selection is a never ending cycle. We can’t be absolutely certain what instance in the process toward optimization we are looking at. We can however research the mathematical and self organizing principles you’ve provided with greater degrees of certainty.

          As a designer, I tend to be more concerned with the performative aspects of biology and their potential for reproducible optimization and less with the anecdotal or speculative justifications for why a species arrived at it’s current configuration. I hope it doesn’t come off as though I’ve dismissed the importance of natural selection and this area of research. It’s only that there seem to be a great deal of speculation and anecdotal explanation.

        • Posted February 20, 2012 at 1:35 am | Permalink

          This just in: Alan Turing’s 1950s tiger stripe theory proved!


      • Posted February 10, 2012 at 11:45 am | Permalink

        Not what I was thinking of…a bit of Googling tells me that it’s Blaschko’s lines that I was thinking of, and that they probably have nothing to do with zebra stripes after all….


  10. Posted February 10, 2012 at 10:57 am | Permalink

    Thanks for the great explanations. The color or pattern of all animals and plants is in the eyes of the beholder. Insects and birds may see different colors and patterns in flowers. Predators who do not see in color may see prey differently than humans. We must keep in mind that we are not only arbitrators of truth. Truth “as we see it” is relative.

  11. Deborah
    Posted February 10, 2012 at 10:57 am | Permalink

    Another possible advantage of stripes would be to deter Bot Flies (multiple species) which lay their eggs on the front legs and heads (near the mouth) of horses (and presumably zebras). The eggs find a way into the mouth of the host and proceed to develop inside and are eventually excreted and hatch into flies, competing the cycle. Equines frequently rub their noses and mouths on their front legs and the bot fly eggs will detach upon contact with warm liquid (which is one way to remove them).

  12. Tim
    Posted February 10, 2012 at 11:26 am | Permalink

    What is the evidence that shows that zebra stripes confer a selective advantage? In other words, are zebra stripes actually an adaptation at all?

    • qbsmd
      Posted February 11, 2012 at 11:39 pm | Permalink

      Be honest: is anyone here tempted to write an NSF proposal for research involving dozens of gallons of hair dye and a round trip ticket to Africa?

  13. Posted February 10, 2012 at 11:34 am | Permalink

    i’ve always wondered why they don’t get ulcers. 🙂

  14. BilBy
    Posted February 10, 2012 at 12:36 pm | Permalink

    tsetse fly traps were (past tense as I’m not sure how much fly trapping is done in the Zambesi valley now) a solid darkish blue, and the Zimbabweans who used to check them would blithely tell me that the flies liked solid blocks of bluish colour, which is why wildebeest and some cattle got hammered by fly bites but zebras did not

    • Posted February 10, 2012 at 7:52 pm | Permalink

      Blue pants attract large tabanids to hikers in Neotropic rainforests and cloud forests. Light pants attract fewer. Dark pants of other colors also attracted fewer than blue pants.
      I never thought to wear striped pants….will have to test!

  15. Posted February 10, 2012 at 1:37 pm | Permalink

    I’m still not clear on why a striped animal would attract fewer horseflies than a solid white one. Wouldn’t an all-white coat reflect the least (or no) polarized light?

    Or is the explanation that the stripes actively *confuse* the horsefly in some way? I read the National Geographic article on this, but it didn’t really explain this part. It just says:

    “But why would striped skin be more effective than white, which has the lowest refectivity of polarized light?

    The black-and-white pattern, Åkesson said, turns out to be “ideal in its function of disrupting this signal of reflected polarized light.”

    Because the coat reflects light in alternately polarized and nonpolarized patterns, the zebra “is more difficult to single out relative to the surroundings.” It is, in effect, camouflaged to flies as well as to big cats.”

    That still doesn’t sound like an explanation. Simply SAYING that stripes are ideal is not explaining WHY they are. What about the alternate pattern is better than purely non-polarized light?

  16. Stonyground
    Posted February 10, 2012 at 2:12 pm | Permalink

    An interesting question with regard to mammals and camouflage is why are mammals never green? Reptiles, birds and insects are often green. Mammals can be yellow, grey, black, white, reddish brown or various combinations of colours, but never green. Green is a really obvious colour to be if you don’t want to be seen but mammals seem to be incapable of that particular adaptation.

    • GBJames
      Posted February 10, 2012 at 2:15 pm | Permalink

      The sloth doth object.

      • Posted February 10, 2012 at 2:40 pm | Permalink

        Isn’t that because of algae in the fur, rather than an inherent colour?


        • GBJames
          Posted February 10, 2012 at 2:45 pm | Permalink

          Not sure if “inherency” matters. My digestion works because of intestinal bacteria that aren’t inherent.

    • Posted February 10, 2012 at 2:45 pm | Permalink

      From Yahoo!

      The short answer is that mammals are hairy. Mammalian hair has only two kinds of pigment: one that produces black or brown hair and one that produces yellow or reddish-orange hair. Mixing those two pigments is never going to yield a bright, contestable green.

      Mammalogist Maria Rutzmoser of Harvard’s Museum of Comparative Zoology suggests a more complex explanation: that small mammals – the ones needing protective coloration the most – typically live on the ground, scurrying in leaf litter. “Dead leaves aren’t green,” she points out. “They’re brown.”

      Finally, most predators of mammals are other mammals, and mammals usually have poor color vision; ergo, green wouldn’t help.

      I’m not wholly convinced…


      • Posted February 10, 2012 at 3:02 pm | Permalink

        Ah… there’s a critique of that answer here. Too involved to easily précis… 😉


        • Posted February 10, 2012 at 3:03 pm | Permalink

          * link fail: http://shkrobius.livejournal.com/176833.html

          • GBJames
            Posted February 10, 2012 at 3:09 pm | Permalink

            I do like the photo at the bottom:

            • Posted February 10, 2012 at 3:15 pm | Permalink

              Now you’re confusing sloth with lust … :-O


              • GBJames
                Posted February 10, 2012 at 3:31 pm | Permalink

                LOL. Nice!

              • Chris Booth
                Posted February 10, 2012 at 10:49 pm | Permalink

                I now see green hair in a completely different light.

  17. ohioobserver
    Posted February 10, 2012 at 2:12 pm | Permalink

    I remember reading that military camouflage is designed to be “pattern-disruptive”, that is, to make it hard to discern the outline of a subject person or animal against any background — it doesn’t appear to have the shape of a soldier/zebra because the camouflage breaks up the outline. Is this relevant to the selection of stripes and spots?

    Of course, there is another answer, since the evilutionists don’t have one — goddidit, and that’s all. Now we don’t have think about this fascinating, and potentially useful problem, and we don’t have to spend any money on it! Now we can think about Jeebus. Solved.

  18. BilBy
    Posted February 10, 2012 at 2:19 pm | Permalink

    Somewhere I have seen it suggested that having black next to white would cause enough of a difference in heating rates between the two types of fur that it would cause a flow of air across the whole body and therefore might have a cooling function. That would be eminently testable.

  19. Marella
    Posted February 10, 2012 at 3:11 pm | Permalink

    It doesn’t explain why the different species of zebra have different stripe widths either. If this were the reason for stripes you’d expect the finer striping pattern to be found on all zebras.

  20. Nicholas
    Posted February 10, 2012 at 3:56 pm | Permalink

    I am astounded at the utter lack of scientific thinking on this post. EVERYBODY knows that Zebras have stripes because polka-dots make them look fat. Is everybody dumb?

  21. Posted February 10, 2012 at 5:25 pm | Permalink

    I’m not so sure about the stripes being white. Non-zebra equids can have a few stripes, and the stripes are dark or black. I was first struck by this when I came across a herd of wild feral equids on the island of Anegada in the British Virgin Islands. They were so big, and had such prominent stripes, that I couldn’t believe they were donkeys, but I think I was too accustomed to the small domestic varieties. Darwin (in the Origin and/or Variation under Domestication) commented on equid striping. I can’t put images in the comments, but I can put a video in: notice the dark shoulder stripe in this donkey:

    • Dominic
      Posted February 13, 2012 at 3:21 am | Permalink

      Perhaps you could write on this next time Jerry is away?

  22. evodevo
    Posted February 10, 2012 at 5:26 pm | Permalink

    Having raised horses for 50 years, and dealt extensively with large, albeit US, horseflies, which are tabanids, too, I can say that their favorite place to land and feed is the middle of the back, toward the posterior, especially right in front of the root of the tail. A horse can accurately swat a fly with the tail almost everywhere else, and shake the head and neck or stomp the legs to dislodge a tabanid, but right in the middle of the back is almost impossible to reach, except by rubbing on a tree.
    By these lights, the stripe width hypotheses don’t hold up. I’m still inclined to go with the optical illusion created when a predator attacks a herd of flashing stripes and gets momentarily visually confused by the op art effect. If you are a predator, you have to be able to fix your attention on only one prey to the exclusion of all the others and follow its elusive maneuvers closely. Maybe the op art effect disrupts that just long enough? Other than that, I have no clue.

    • Dominic
      Posted February 13, 2012 at 3:22 am | Permalink

      Is there any evidence then that zebras are less favoured prey rather than say Wildebeest by lions etc?

  23. Occam
    Posted February 10, 2012 at 9:49 pm | Permalink

    For the photographically inclined among you: could it be that zebra stripes were designed so that Nikon could sell twice as many D800 cameras to photo safari junkies? The D800E, without anti-alias filter, for sharp vistas; the vanilla D800, with AA filter, for the zebras.

    (It’s damn cold outside, I’m getting my coat…)

    • Kudu
      Posted February 11, 2012 at 6:03 am | Permalink

      Referees in NFL football other sports wear vertical stripes. Why dont we just ask them why they do it, and it will probably shed some light on the zebras.

  24. Kudu
    Posted February 11, 2012 at 6:00 am | Permalink

    I dont understand why we would assume that the stripe width should be the most effective at the point on the zebra where protection from flies is most important. After all, the stripes have no intelligent designer, but are the result of random changes filtered through natural selection. If no random change ever happened to put the most effective stripe width on the most critical point on the zebra, then natural selection has no way to propagate those genes through the population, right?

  25. Rob Knell
    Posted February 12, 2012 at 5:38 am | Permalink

    Just a note to say this isn’t actually a new idea: people have been talking about zebra stripes being in some way an anti-tsetse fly adaptation for at least 15 years. I’ve always thought it’s a great subject for some good experimental science and it’s nice to see some appear.

    One other thing to point out is that stripes (like everything else) might have multiple adaptive effects, so they could both interfere with insect vision and act as dazzle-pattern camouflage. I tend not to like to camouflage explanation though because one thing that zebras tend to do is herd with wildebeest, and a single stripy zebra in a herd of black wildebeest stands out like a sore thumb. If the main benefit of the stripes were to dazzle and confuse predators then I think you’d predict that zebras should always herd with other zebras and not with noon-stripy animals.

  26. Rob Knell
    Posted February 12, 2012 at 5:38 am | Permalink

    ps sorry about the spelling rubbish keyboard

  27. W.Benson
    Posted February 12, 2012 at 11:16 am | Permalink

    Too much speculation, no. Too little data, yes!
    There are many hypotheses to account for the zebras of zebras, and it would be good if someone with access to these animals in Africa would get to work. I have seen zebras in films and cannot buy that they are cryptic. Zebras also do not seem to run in tight packs that might confuse predators. In fact, they seem to have few predators, perhaps only lions are important, and in the open savannas where zebras seem to live (am I mistaken?) ambush seems difficult. In any case, zebras are so big that it would be hard for one to hide on an open savanna, no matter its color. This makes the idea of an anti-fly function more likely. In woodlands, and in places with winter snow and wolves, equids might not be so free to become conspicuous.
    If crypticity is unnecessary, color pattern may evolve to other functions, at times more than one at the same time. Thus, even if it could be shown that zebra stripes significantly reduced the transmission of deadly parasites, this would not refute alternative functions such as individual recognition, social signaling and predator disorientation. Stripes could even evolve to signal to experienced predators that the zebra is almost uncatchable and potentially dangerous target, and thereby reduce harassing and perhaps fatal lion attacks on distinctly striped individuals. Alas, many hypotheses, few data!

  28. Rumtopf
    Posted February 16, 2012 at 8:51 pm | Permalink

    You do get Donkeys with stripes on their legs, check out the Somali Wild Ass.

    • Rumtopf
      Posted February 16, 2012 at 8:54 pm | Permalink

      Or I should say African Wild Ass in general, the Somali WAs are an endangered subspecies :c, but they all have the stripes.

One Trackback/Pingback

  1. […] Why does the Zebra have stripes? Here is a possible answer. And some problems that Jerry Coyne raises. […]

%d bloggers like this: