Why eyespots?

While listening to talks at the evolution meetings, I’ve mentally divided them into two groups: what I call “general” versus “anecdotal” research.  The former seeks general laws of evolution that apply across diverse species.  “Haldane’s rule” is one example:  the observation that if, in a cross between two species, only of the two sexes of hybrid offspring is sterile or inviable, it’s nearly always the heterogametic sex (males in mammals and many insects, females in birds and lepidoptera). I’ve spent a lot of years trying to explain that one. Another “law” is the repeated observation that if only one sex in a species is ornamented or brightly colored, it’s almost invariably the male sex.

“Anecdotal” research—the name is not meant disparagingly—seeks to find the evolutionary basis of a single phenomenon, often in a single species.  The “panda’s thumb”, made famous by Steve Gould, is a familar example.  In this case, a herbivorous bear has evolved a rudimentary opposable “thumb” by modifying the radial sesamoid bone of the wrist.  The thumb helps strip leaves from bamboo, the only item on the panda menu.

Both strategies are essential to answer the question, “How has evolution produced the marvels of nature?”  But young people at this meeting seem to be pursuing the “generalist” strategy, perhaps sensing that career rewards are more likely to come if you answer Big Questions rather than concentrating on a single system.

A new perspective piece by Dan Janzen and his colleages in PNAS straddles the boundary between these two areas. It takes a single element in the color pattern of caterpillars and pupae of Lepidoptera—the presence of false eyespots—and floats a theory to explain this group-restricted pattern.

Janzen, perhaps the world’s finest field naturalist, has spent much of his life studying the insects of Costa Rica, especially in Guanacaste Province in northwest Costa Rica.  (I had the privilege of being one of Dan’s students in a Tropical Ecology course in Guanacaste in 1973.) He and his colleagues noticed, as others had before them, that many of the Lepidoptera (butterflies and moths)  in the neotropics had markings on the larvae (caterpillars) and pupae that looked a lot like eyes.  Here are some caterpillars with eyespots, taken from the paper:

And here are nonmobile pupae with eyespots:

Nice, eh?  Biologists have reflected on the existence of these eyespots, suggesting that they evolved because they enhance survival.  How? By fooling predators, mostly birds, who mistake the “eyespots”  for their own enemies—snakes, lizards, predatory birds, and some mammals—and flee in fear.  The insect with the eyespots thereby avoids being eaten.

It’s clear that the eyespots have something to do with predation, because in many cases they’re displayed only when the larva is disturbed or detected by a potential predator.  Here’s one example, with the caption (taken from the paper):

The 7-mm-wide pupa of Cephise nuspesez (23) (Hesperiidae), a Costa Rican skipper butterfly as itappears to a foraging bird that (Upper) has poked into the front of the rolled leaf shelter constructed bythe caterpillar or (Lower) has opened the roll from above. When disturbed, this pupa rotates to present its face to the open end of the leaf roll.

That sure looks as if it would frighten a foraging bird that was investigating a leaf.

Here’s another example of how eyespots are displayed when a caterpillar is detected (note that this adaptation is twofold: the eyespot itself but also the evolved behavior that displays it only in a certain context).  The caption is from the paper:

The 50-mm-long last instar caterpillar of Costa Rican Ridens panche (Hesperiidae) at the moment when its leaf shelter is forced open (Upper) and a few seconds later (Lower), when it presents glowing red false eye spots directed at the invader and glowing lemon-yellow eye spots in the dark of the cavernbehind. Both kinds of false eyes are thrust at the leaf roll entrance until the invader leaves.

So far so good.  Janzen et al. are not the first to suggest that eyespots evolve to protect lepidopterans from predation.  But they go further, and suggest that the birds’ avoidance of insect eyespots is often innate (that is, a hard-wired genetic behavior that is the product of natural selection) rather than learned.

A bit of background.  Some forms of mimicry, in which an edible species of insect mimics another species that is both brightly colored and repugnant to predators, involve predator learning.  When a hand-reared and naive bird eats a ladybug for the first time, it noms it down and then, realizing how dreadful it tastes, spits it out.  You can show in the lab that after one or a few such episodes the bird learns to avoid the brightly spotted pattern of the lady bug. And, it will also avoid tasty insects that have evolved patterns that resemble the ladybug. (This resemblance has evolved in many species that birds find tasty, including cockroaches and beetles.)

This form of mimicry, in which an edible species evolves to physically resemble an inedible one, is called Batesian mimicry, after the Amazonian naturalist H. W. Bates.  Its evolution depends on the predator being able to learn that an insect is inedible, and then generalizing that experience to avoid other insect species with similar patterns.

Janzen et al. suggest, however, that the eyespot mimicry (lepidopteran patterns mimic bird-predator eyes) is based not on the predator learning to avoid the eyespots, but evolving to avoid the eyespots.  The state their reason succinctly:

[T]he bird that must learn to avoid an eye is not long for this world.

In other words, the evolution of “eye avoidance” (which generalizes to eyespot avoidance) is likely to be innate rather than learned, for it’s hard to learn to avoid an eye.  If you encounter the eye of an owl or a snake, and don’t flee right away, you’re dead. No learning can occur. On the other hand, the higher survival of individuals who flee at the sight of an eye would select for an innate avoidance of things that look like eyes.

Once that’s evolved in an insect-eating bird, it sets the stage for the evolution of eyespots in caterpillars and pupae, which gain survival benefits from the birds’ innate avoidance of anything eyelike. Janzen et al. call this phenomenon, which has apparently caused the evolution of eyespots in hundreds of diverse lepidopteran species, “diffuse seletion.”

Janzen et al. don’t mention this, but their theory about innate versus learned avoidance is eminently testable.  All you have to do is hand-raise, from eggs, some of the birds known to flee from eyespots.  If, on first encountering a pupa or caterpillar with an eyespot, they get startled and flee, then their aversion must have been innate rather than learned.

This kind of experiment was done by Susan Smith in 1975, showing that birds who avoid the black, yellow, and red striped pattern of coral snakes do so innately, not through learning (indeed, it would be hard to learn since an encounter with a coral snake is likely to be fatal).

Explaining eyespots may not yield the professional cachet of explaining something like Haldane’s rule, but the real joys of evolutionary biology are found more often in the particular than in the general.


UPDATE: In the comments, Naturalistbiologist points out the scary resemblance of the Gaudy Sphinx caterpillar to a snake.  Just to show you how far caterpillar mimicry can go, have a look at it:


Janzen, D. H., W. Hallwachs, and J. M. Burns.  2010.  A tropical horde of counterfeit predator eyes. Proc Nat. Acad. Sci. USA 107:11659-11665.

Smith, S. M. 1975. Innate recognition of coral snake pattern by a possible avian predator. Science 187: 759–760.

28 thoughts on “Why eyespots?

  1. In the second set of photos, third row, center: This is a very interesting mimic of a snake– you can see the ‘scales’ quite clearly (larger around the ‘lips’, smaller on top of the ‘head’) and the color pattern resembles a common one in snakes. There is one interesting difference. In a snake, the eye would be located in the dark eyestripe on either side. In the mimic, the ‘eye’ is more nearly where the nostrils would be, and thus much more noticeable. In snakes, the eyestripe is thought to be a camouflage feature for breaking up the outline of the head and disguising the eye. The caterpillar, however, ‘wants’ to be seen, and thus the ‘eye’ is not in the more cryptic position found in real snakes.

      1. Very nice, thanks. I’ve added a picture of the Gaudy Sphinx caterpillar to the end of my post.

  2. You would think that the bird brain would develop a distinction between eyes on something smaller than me and eyes on something bigger than me. Perhaps in the next few millennia… but if it hasn’t happened yet, perhaps the incentive for avoiding eyes out-weighs the possible gains for making the distinction.

    I wonder whether eyespots work on other mammals or lizards. Do they dissuade predators that eat other creatures with visible eyes like rodents and fish?

    1. Snakes much smaller than a bird can still kill it, and very well might if startled. Just like rattlesnakes much smaller than a person can still kill one.

    2. You would think that the bird brain would develop a distinction between eyes on something smaller than me and eyes on something bigger than me.

      Hence the term birdbrain 🙂 (sorry, couldn’t resist).

  3. Even I flinch upon seeing some of those eyes. Feels innate.

    One question I have always had about edible insects mimicking inedible ones is: If enough edibles mimic, wouldn’t the signal be lost, because the predator would find the pattern equally likely to be carried by a tasty bug as not? What stops mimicry from reaching this ultimate equilibrium?

    1. This is a recurring debate in the evolution of mimicry. With respect to a predator who learns, as in Batesian mimicry, it turns out that if the inedible mimic is sufficiently inedible, the predator can still maintain its aversion even if it encounters more mimics than inedible “models”. And with regard to these eyespots, which may be maintained by an evolved predator, all it takes to work is that predators who avoid eyespots have a marginally higher survival than those who don’t. And this can hold even when the predator encounters the fake eyespots a lot more frequently than it encounters real eyes belonging to something that could eat it.

      1. Thanks, that makes sense. So the equilibrating factor would be the ratio of tasty mimics to nasties. A real nasty might have nine mimics per real thing, say, so the predator now has a 10% chance of a nasty but 90% chance of a tasty mimic to balance the risk. The mimicry signal has lost its power, but only at the margin.

        That sounds like it has some testable implications too.

    2. If you eat at a certain restaurant frequently, but in the space of a few weeks have two bad experiences, aren’t you likely to avoid going back?

      Negative reinforcement, in my experience, tends to be much stronger than positive reinforcement.

      One insect does not a meal make, so the pressure to eat everything in a flush environment is not going to compete well with the incentive to avoid eating X, where the last X was quite repugnant.

  4. Gould was a master of using this “Anecdotal” research in his writing. The Panda’s thumb is indeed a well known example but most of his articles were of that type. From the twists of a snails shell to the Flamingo’s smile, he would take some specific biological feature and show how it revealed a more general truth.

  5. This is all very interesting but I wonder why eyes? What made birds feel compelled to avoid eyes and not something like scales?

    1. I was wondering about that too, and I wonder if part of the signal is to make the predator think it has been detected.

  6. Wait, isn’t is obvious that creatures instinctive recognize faces? Isn’t that why we think that this ^_^ is a smiling person rather than a series of nongramatical punctuation? Doesn’t the fact that an (American-style) outlet has a look of surprise mean we shouldn’t be surprised at this?

    Just my two cents.

  7. “But young people at this meeting seem to be pursuing the “generalist” strategy, perhaps sensing that career rewards are more likely to come if you answer Big Questions rather than concentrating on a single system.”

    And our mentors at big research institutions like, um, U. of Chicago have nothing to do with that…

    I met Janzen in 1992 when down there on a tropical ecology course. I can’t believe he’s STILL at it!

  8. I only have eyes for you.

    It takes a single element in the color pattern of caterpillars and pupae of Lepidoptera—the presence of false eyespots—and floats a theory to explain this group-restricted pattern.

    The astrobiology course book, using Earth as the best (planet) or only (biology) known reference case, calls this “case study” – taking an example and trying to extract general principles and theory.

    I’ve never heard of “anecdotal” research though. A naive characterization I would have made would have been “ad hoc”, since it is as yet unconnected with general theory. Sorry, but it sounds more neutral. 😀

  9. I am sure this is just one of god’s trials to test the strength of creationists’ faith.

  10. Sorry about the topic drift but I just thought of an example of how an insect could become invisible to a bird without camouflage.
    Every time I see a butterfly in the air it seems to fly upwards for a while then it stops beating its wings and falls like a leaf for a bit then resumes flapping.
    I used to think that the butterfly was just resting its muscles but then I realized that by falling like a leaf it suddenly becomes invisible to the bird. Birds can’t chase falling leaves, there are so many of them that it would waste the birds energy if they did that.

  11. But how does it happen? How does the caterpillar begin to grow one of them? they can’t choose to grow an eyespot (i’m guessing), and it can’t be just down to natural selection there are many with eyespots if Janzen is right.

    Sorry, I know it sounds simple minded but I just don’t understand how they arose?

  12. Pingback: » Monday links

Comments are closed.