A clever new hypothesis about insect mimicry

February 26, 2017 • 9:00 am

Over the years I’ve written here about several kinds of mimicry. The most common subjects have been Batesian mimicry, in which the evolutionary scenario involves three species: an easily identifiable and noxious or toxic model, a predator that learns (or has evolved) to avoid the model (signal receiver), and an edible mimic that evolves to resemble the model. You can easily see how an edible species would leave more offspring if it accumulated mutations that made it resemble the model, for it would be avoided by the predator more often. Here’s one example: a harmless and edible fly that mimics a bee, almost certainly to avoid bird predation:

Source: The Nature Observer’s Journal

The second form of mimicry I’ve often discussed is Müllerian mimicry, in which a group of easily identifiable species, all of them toxic or unpalatable, evolve to “converge,” or resemble each other. Such mutual resemblance gives members of the similar-looking species an advantage, for it’s easier for the predator to learn and avoid one pattern rather than several.

Here’s one example, a group of butterflies in the genus Acraea,  Such mimicry needn’t just involve one group of organisms: similar patterns have been described in mimicry “rings” that involve butterflies, beetles, “true bugs” (Hemiptera), and wasps. All of them can achieve some protection from predation by adopting similar colors and patterns.

Source: MimeticButterflies.org

For both sorts of mimicry to evolve, the signal receiver must encounter both model and mimic, so they all have to live in the same area. (There is one scenario in which model and mimic can live in different areas, but I’ll leave you to figure out how that works.)

A classic case of what was thought to be Batesian mimicry involves moths or butterflies that look like bees or wasps. The resemblance between model and putative mimic is sometimes astonishingly precise. Here’s a moth that for all the world looks like a wasp. Note all the features of the moth that have evolved to resemble the wasp: the wings are clear, are folded longitudinally like those of a wasp, the body is narrow, striped and colored like that of a wasp, and has a very thin constriction between thorax and abdomen (“petiole”) that regular moths don’t have. Believe me, if you saw this, you would mistake it for a wasp and avoid it, just like predators do.


(From Boppré et al. paper): Not a stinging wasp but a harmless day-flying moth (Lepidoptera: Erebidae: Arctiinae: Pseudosphex laticincta). These moths are “sheep in wolves’ clothing” and simulate their predators—this is not necessarily a case of classical mimicry. Photograph © courtesy of Hannes Freitag (FZE)

But this may not be a classic case of Batesian mimicry, or so claims a new paper in Ecology and Evolution by Michael Boppré et al.  (reference below, free download). There may not be a “signal receiver” that has learned to avoid the moth because it resembles a stinging wasp. Rather, as Boppé et al. suggest, the predator is said to be the “model” itself: a predatory wasp, and the scenario involves innate avoidance rather than learning.

Here’s how it works. The wasps in question, yellowjackets (wasps in the genera Vespula and Dolichovespula) are abundant social insects that make their living as predators (and scavengers) of other insects.  Here’s a yellowjacket attacking a praying mantis:

We know that on their hunting expeditions yellowjackets won’t attack members of their own species; that, after all, would be a maladaptive behavior, since members of your own species may well be members of your own nest. Further, if you attack another yellowjacket, you yourself could get stung to death.

Besides this observation, the authors noted that the mimicry between palatable moths like the one above and the yellowjackets is astonishingly precise: far too precise, they say, to have evolved to fool birds. (They argue that birds will avoid a prey simply by longer-distance recognition of general patterns like color and striping.)  Here’s are two more examples of the precision (read the caption); the first photo shows two species of wasps and a moth. Can you tell which is which? (Hint: the antennae are a giveaway).

(From paper):  Two species of eusocial wasps and a “wasp-moth” from Costa Rica—but which is which? The moth simulates not only the striped abdomen but also transparent and folded wings, petiolate abdomen, and patterned thorax of the wasps. Its true identity is revealed by its proboscis and pectinate antennae. (a) Mischocyttarus sp., (b) Polybia sp. (Hymenoptera: Vespidae), (c) Pseudosphex laticincta (Lepidoptera: Erebidae: Arctiinae)

And here’s another example of precise mimicry of body shape, color, and pattern. Again, inspection of the antennae shows that the wasp is on the left and the moth on the right:

(From paper): case of accurate resemblance between a black eusocial wasp (a, Hymenoptera: Vespidae: Parachartergus apicalis) and a neotropical moth (b, Lepidoptera: Erebidae: Arctiinae: Myrmecopsis strigosa), showing the very same simulated features (abdomen, wings, petiole, thorax) discussed for yellow jackets (Figures 2 and 3). This exemplifies that the hypothesis discussed at length for yellow jackets can also be applied to understand accurate simulation of other color patterns. (The wing folding of the moth is incomplete in this photograph.)

The authors’ hypothesis is both clever and simple–so clever and simple that, in fact, I’m surprised that nobody has suggested it before. It is this:

The moths aren’t mimicking the wasps to fool birds; they’re mimicking the wasps to fool the wasps themselves. That’s because the wasps are predators, and will avoid attacking any insect that looks very similar to their nestmates, because you don’t get a food reward by attacking another wasp. 

In other words, the predatory wasps have an innate aversion toward attacking animals whose appearance they’ve evolved to recognize (presumably because they’re eusocial and live in groups where they help each other); and the “model” takes advantage of this, evolving a precise resemblance to the wasp. The authors like this hypothesis because they think that the wasps scrutinize potential prey much more closely than do birds, a trait that “forces” the models to evolve a very precise resemblance—a resemblance that, they say, couldn’t be explained by the scrutiny of birds alone.

The difference between this and “classic” Batesian mimicry is that a.) the signal receiver and the model are the same species: the wasp; and b.) there is no learning involved: wasp recognition of its self-pattern is an evolved trait, probably evolved the same way most animals develop a recognition of members of their own species.

This hypothesis is clever, and when I read the paper I had the same reaction that Thomas Henry Huxley had after reading Darwin’s On the Origin of Species: “How extremely stupid not to have thought of that!” After all, we’ve known about mimicry since shortly after The Origin was published, and yet it took people over 150 years to come up with this simple idea.

Yes, it’s a clever idea, and may well be right, especially if the chance of predation by a wasp is much higher than by a bird. But is the hypothesis true? What’s the evidence for it?

Sadly, there isn’t much yet beyond the authors’ speculation that mimicry this precise could not be mediated by visual bird predation, but requires the acute vision of a wasp. Further, the authors describe yellowjacket wasps preying on moths in the wild but not on other yellowjackets or on moths mimicking yellowjackets. (This leads to the idea that the similar black-and-yellow striped pattern of different wasp species could be a case of “quasi-Müllerian” mimicry, but one in which they’ve evolved to resemble each other because it’s injurious to attack each other. In this case again, there is no predator learning involved, nor any signal receiver.)

Now the authors’ hypothesis doesn’t rule out that this is also a case of true Batesian mimicry: that selection occurred both by wasps avoiding attacks on prey that look like themselves, and also by birds having learned to avoid attacking anything that looks like a wasp. Both factors could be involved, and I suspect are. But how do you test whether yellowjacket predation was a driving force of selection?

The authors note that it’s hard to test that:

However, it seems likely that (2) general, visually oriented predators such as birds are additional selecting agents shaping similarity of other insects to wasps. Thus, in “wasp mimicry” two sorts of selecting agents (with different life-styles) are plausibly acting. Then, the relative abundance of predatory wasps (individuals and species) that recognize look-alikes as non-food versus various predators that learn through experience could explain the accuracy and non-accuracy of potentially profitable mimics. We would observe combinations of innate protective masquerade and learned Batesian and Müllerian mimicry, and recognize different sorts of selecting agents, namely those which respond innately and those which learn by experience. Thus, accurate mimics would be protected against both wasps and birds, whereas inaccurate mimics would be protected mainly against educated birds (which to a certain extent generalize a learned pattern) but not so well against wasps. In theory, proof could only come from studies in habitats where wasps prey on insects but learning predators do not occur—however, such places cannot be found.

But you could think of other tests. For example, put mimetic moths and non-mimetic moths in a cage with yellowjackets. If the mimetic moths are attacked less often, that would support the authors’ hypothesis. Or you could efface certain parts of the mimics’ pattern with paint and see if they get attacked more often.  I’m sure that, with some thought, other tests are possible. Here’s one I just thought of: if a palatable moth resembles something that doesn’t attack it, like a bumblebee, then it’s likely that this is a case of true Batesian mimicry rather than the new form of mimicry (not given a new name) described by Boppré et al. And in such cases you’d expect the mimicry to be less precise than that of yellowjacket mimics, because (according to the authors), birds don’t need to look as closely at potential prey as do yellowjackets.

A final point: the authors note that sometimes the moths themselves may be toxic: some species eat plants and, like monarch butterflies, sequester the unpalatable alkaloid compounds in their bodies. If this were the case (and that would be relatively easy to test), then the mimicry becomes more complicated, and harder to understand. Why wouldn’t the wasp predators evolve to recognize whatever pattern a toxic moth had in the first place—that is, why did such precise mimicry evolve? Perhaps in this case it would prevent the initial strike of the moth that kills the mimic, even if the wasp then recognizes that the mimic isn’t fit to eat. Or, the authors could be wrong about how birds recognize potential prey: birds may be more sharp-sighted than we think.

At any rate, this is a novel hypothesis and well worth considering in the case of mimics that look like their predators. I was once fooled by one of these moths that had gotten into my house in Maryland, and I had to get my fly net to catch it as I was afraid of getting stung. Only after I caught it did I realize the ruse. If a moth can fool someone who works with insects, it’s likely that it can fool a predatory wasp or bee.

h/t: Matthew Cobb


 Boppré, M., R. I. Vane-Wright, and W. Wickler. 2017. A hypothesis to explain accuracy of wasp resemblances. Ecology and Evolution 7:73-81.

42 thoughts on “A clever new hypothesis about insect mimicry

  1. Fascinating. Mimicry is one of the most amazing subjects to learn about. Now we may have expanded the sense of wonder. Now, to come up with a name for it. Something with the word “mirror” in it?

  2. Do the wasps identify fellow wasps purely by visual means? Is there a possibility that the mimicry extends to other sensory inputs? Given the detail of the mimicry there must be a strong visual component but it would be interesting to know if there were other aspects to mimicry that are beyond our vision-dominated perception.

    1. My first thought was perhaps related: how do we judge the similarity of colouration?

      To human eyes these moths look very much like wasps. But the colours (number and spectrum) seen by birds and by wasps are different. It would be very interesting to photograph them in the appropriate spectral bands — if they look very different through the bird’s eyes, that would be strong evidence for this theory.

    2. I have difficulty with the premises this hypothesis rests upon. Wasps have compound eyes of the apposition type (IIRC), their image would be less precise (lower resolution) than that of birds, known to have typically even better eyesight than we do (UV, higher resolution, accommodation over dozens of diopters, etc), let alone apposition compound eyed wasps. Which would count against the ‘new’ hypothesis.
      In other words, do these moths indeed need to visually resemble more precisely to deceive their predatory ‘model’ rather than a third party predator, ic. birds?

      Could this be a case of “a beautiful theory destroyed by an ugly little fact” in stead of a “how stupid not to have thought of that”?

      Note apposition compound eyes are very good at motion detection, maybe there could also be some ‘motion mimicry’?
      Point about pheromones: if indeed they resemble the wasp’s this would be a strong argument in favour of the ‘new’ hypothesis. Birds are notoriously bad ‘smellers’ compared to many insects and most mammals (excluding us there).

      I would go with Jerry and say: possibly both. Even if the wasp’s eyesight were not the strongest driver. Whether eaten by a wasp or a bird: you are equally dead, taken out of the genepool.
      However, it has the hallmarks of a good hypothesis: a plethora of ways to investigate further.

  3. Thanks, as always, for the time I know it takes to post these discussions of current research. Coincidentally, I just received a TOC alert that includes a paper that might complement this one. It’s in Evolutionary Biology (link below). I haven’t read the paper yet, but the abstract focuses on developmental patterns, and their role in producing similar color patterns in wasps and syrphid flies. Papers like these are important beyond their immediate contribution because they illustrate that evolutionary biology is more than a collection of “just so” stories. Readers of this blog know this, but critics and curious students want more than “the mimic benefits from its convergent similarity to the model”.


  4. Thanks for the article. Through science articles such as this one, you have opened my mind to thinking about nature that never otherwise would be the case.

  5. That is very clever! I agree that this can be tested in controlled settings, like you suggest. This sort of thing has been done before between mimics and bird predators, and even between mimics and insect predators (preying mantises avoid aposematic insects).

    Another possible example that occurs to me are insects and spiders that mimic ants. There are many cases of this, and perhaps there is a bit of this ‘new’ mimicry going on there as well. That is, an arthropod mimics an ant to be avoided by other ants.

  6. “Perhaps in this case it would prevent the initial strike of the moth that kills the mimic, even if the moth then recognizes that the mimic isn’t fit to eat.”

    I think you wanted to write wasp instead of moth in both place in this sentence.

  7. Are there any plausible examples of predator species that mimic prey species in order to get closer or have an easier time feeding? Any examples of, say, a predator mimicking the female of a species in order to lure in males of the species to eat?

    1. Yes and yes. Angler fish and snapping turtles use a lure of a small prey item to draw in prey to them.
      As for mimicking a female to lure in males, there is of course the famous example of a predatory firefly that mimics the flash of a female firefly belonging to a different species. When a male of that species comes to it, they are eaten.

  8. I wonder if the moths also use chemical cues as part of their mimicry? Short-distance visual cues plus long distance odor cues?

  9. I have read that some birds, for instance tanagers, eat yellow-jackets. I expect that if the moth mimics were avoiding bird predators rather than the wasps, the mimicry might be counter-productive around tanagers.

  10. I wish I had become an entomologist instead of an education/math major. You won’t believe this, but over the years, photographing so many insect mimics, I had, at first, assumed that they mimicked as this new theory states. Then I did some cursory reading on mimicry and learned of the two types of and found I was wrong. If I were an entomologist, I’d have written a paper and gotten famous. Now I’ll just continue to labor in poverty as I always have. Sigh. My favorite article so far. Thank you.

    1. Entomologists are not exactly rolling in $ either. But in our chosen fields, you and I may afford some time in pursuit of other learning and photography.

      1. Interesting. I thought mantises mimicked plants to avoid predation, but you are right. The flower mantis is a good example of aggressive mimicry.

        Perhaps mimicry serves both purposes in mantises.

    1. I know of a robberfly species that specializes in hunting bees here in Costa Rica that looks like a common species of euglossine bee.

  11. Very interesting post as usual PCC(E), thanks. I haven’t read the original article yet. I wonder if the authors discuss about the sensory ecology differences between wasps and birds. I know some wasps have very good vision (e.g. paper wasps), so the visual resemblance of the moths would make sense, but what about the chemical signals?

  12. Many years ago in the 80s ,i had steeled myself to get close enough to a Wolf Spider on a Nettle leaf to take a photo .
    A Scorpion Fly landed on my arm ,i knew it was harmless ,still almost wet myself .

  13. A few folks above alteady commented on other sensory means beyond visual that wasps use to recognize one another. I think that could be key. If the moths have evolved their complement of pheremones away from those of moths and toward those of wasps, that seems like strong evidence supporting the new theory. I presume avian predators of moths (and wasps even) hunt visually and do not follow pheremone scent trails and it seems unlikely that birds would possess highly sensitive pheremone detection. As I recall, a moth’s antenna is one of the most sensitive biosensors found in nature. In other words, the moths have evolved visual mimicry to fool both wasps and birds but evolving “smell” mimicry would only occur to fool wasps.

  14. Seems like “social commensalism”, similar to social parasitism. We might then predict that the moth might also have a form of chemical camouflage to go along with the morphological.

  15. I also think that those moths displaying jumping spiders on their wings (a few posts earlier on this very website) might be andidates for this ‘new’ mimicry hypothesis.

  16. Fascinating idea. I am curious to know if the moth mimics also show the behavioral mimicry seen in the beelike flies that try to “sting” when caught, moving the abdomen in a pretty good imitation of what bees/wasps do.

    Most robber flies, by the way, are capable of inflicting a nasty and painful “bite” with that proboscis. Laphria, when handled, should be handled with caution. I didn’t.

Leave a Reply