How and why does this caterpillar whistle?

June 18, 2015 • 9:00 am

A piece in ScienceNews reported some new research (not yet published) that updated an old but good paper on the famous “whistling caterpillar.” I refer to the walnut sphinx caterpillar (Amorpha juglandis), which, when disturbed—as in a predator attack—emits a high-pitched whistle from its spiracles (the body openings used for respiration).

An earlier paper (from 2010) by Veronica Bura et al. in the Journal of Experimental Biology (reference and free link below) studied the whistling behavior, showed how the sound was made, and found that it apparently startles bird predators, who flee and thus allow the whistler to live another day.  Newer work, however, shows that the whistle is even more amazing than thought, for it appears to mimic the sounds that other birds make when they see a predator: an “alarm call.”

But first, let’s look at the magnificent caterpillar. Nice one, eh?


The video below (you can see similar behaviors in the supplementary material to the paper) shows how the caterpillar whistles when it’s disturbed by experimenters, a disturbance designed to mimic a bird attack. (This caterpillar may also be attacked by other vertebrates like bats, snakes, and frogs.) You may notice that the “squeak” sounds like a bird, and that’s important.

In 2010, Bura and her colleagues did a number of studies on this species to locate the source and examine the effect of this sound. The salient results:

  • They measured the characteristics of the sound with sonograms. That needn’t detain us except to note that the median length of the squeak produced by these artificially disturbed larvae was about two seconds.
  • The sound was always accompanied by a contraction of the front of the body that lasted about half a second. I suspect this contraction is what forces air through the spiracles. Here’s what it looks like (figure from the paper), along with a recording of when the sound was produced

Screen Shot 2015-06-18 at 7.19.55 AM

  • The most obvious source of the sound would be the spiracles, the holes in the body wall through which larvae (and adult insects) breathe, taking in oxygen through diffusion (or active body movement) and expelling carbon dioxide. The association of the sound with body contraction makes this hypothesis more credible. The authors then covered up the spiracles with removable make-up latex to see if the spiracles were responsible, and, if so, which ones.  They found that when all spiracles were covered, no sounds were emitted in mock attacks (pinching the beast). When the spiracles were uncovered, one by one, the authors found that only the posterior pair of spiracles (there are eight pairs) were involved, for only this pair, when covered, eliminated no sound. The posterior (“A8”) spiracles are also larger than all the others.
  • The authors also did “laser Doppler vibrometry” to see whether the air and sound emitted during the squeak came out of a specific pair of spiracles.  This involves bouncing a laser beam off of a small piece of tissue paper suspended near a spiracle. Vibrations of the tissue produced by emitted air will be detected by the laser. They found that the laser showed a big signal only when the tissue was by the A8 spiracle rather than a control (A5) spiracle, and that the vibrations disappeared when the spiracle was occluded by latex. Here’s a figure showing the results of the vibrometry. You can see a bit of vibration from A5 (which didn’t disappear when the spiracle was covered), but a a big pulse from A8 (pictures of the spiracles are also shown), which did disappear when the spiracle was covered with latex. Thus, the squeak is produced by air coming out of spiracle A8:

Screen Shot 2015-06-18 at 7.20.19 AM

  • Finally, the big question: does the squeak really deter predators? The authors tested this by exposing caterpillars to three captive (but wild-caught) yellow warblers (Dendroica petechia), who attacked the larvae regularly. In every case, when the bird attacked the caterpillar, the caterpillar produced a squeak (squeaks weren’t made when the larva wasn’t disturbed). And in every case, the bird was deterred by that squeak and fled: every bird hopped or flew away when the sound was made, and one bird even dove down into the vegetation as if it were attacked by a predator”. (That’s important; see below.) No caterpillar was harmed in the least by these attacks, so the sounds seem pretty effective in deterring predation.

Here’s a pretty lousy figure from the paper showing a bird being deterred by the sound and diving for cover:

Screen Shot 2015-06-18 at 7.21.13 AM

But crucial experment wasn’t done. If it is the caterpllar’s squeak that deters the predator, the predator wouldn’t be deterred if the sound were eliminated. In other words, they should have done the same tests using warblers with the A8 spiracles of the caterpillar occluded with latex. If the authors are right, those caterpillars should have been nommed. That would have been an easy experiment to do,  but for some reason it wasn’t. Thus the crucial information—that the sound itself deters predation—is missing.

The update by other investigators was, perhaps, motivated by one sentence in Bura et al.’s paper, which I’ve already alluded to:

In one case, the bird dived away from the caterpillar into thicker vegetation in a manner similar to how it would react to a predator.

That suggests, to the curious and atttentive mind, that the caterpillar’s squeak might imitate the “alarm” calls that birds give when predators nearby. And indeed, a new meeting abstract by Jessica Lindsay and Erik Greene (link and title below), which expands the previous results, suggesting that that the caterpillar noises aren’t just disturbing, but are disturbing because they mimic the “alarm calls” that birds give when predators like hawks are around. The birds react as if they hear the alarm call, leaving the insect alone not because they’re simply startled, but because they need to find cover fast. Here’s part of that abstract; the paper has yet to be published:

Many birds produce “seet” alarm calls in response to high-threat danger (such as flying raptors), and these cause other birds to dive for cover.  Birds also produce more broad-band “mobbing” alarm calls in response to perched or stationary predators.  We noticed that the acoustic structures of some of the caterpillar sounds are similar to bird “seet” and “mobbing” alarm calls.  Using playback experiments, we tested the hypothesis that the caterpillar whistles mimic bird alarm calls.  Many species of birds dove into cover or froze upon hearing bird “seet” calls and two types of caterpillar whistle; they also tended to fly off and orient towards the speaker in response to a third type of caterpillar whistle.  These caterpillar whistles, rather than eliciting a general startle response, may protect caterpillars by mimicking the alarm calls of their avian predators.

This will undoubtedly be published soon as a full paper, but in the meantime it’s a cool suggestion.  I’ve put below what I think is a “seet” call of a chickadee, which resembles the call of the caterpillar above. I may be wrong about the “seet”, so birder/readers who have better examples are welcome to post them below:

h/t: Barry


J. Lindsay and E. Greene. Whistling caterpillars mimic the alarm calls of birds. Presented June 12, 2015, at the 52nd annual conference of the Animal Behavior Society, Anchorage, Alaska.

V. Bura et al. Whistling in caterpillars (Amorpha juglandis, Bombycoidea): Sound-producing mechanism and function. Journal of Experimental Biology. Vol. 214, January 1, 2011, p.  30.  doi:10.1242/jeb.046805.

46 thoughts on “How and why does this caterpillar whistle?

  1. Fascinating.

    When you get in trouble and you don’t know right from wrong,
    give a little whistle!
    Give a little whistle!
    When you meet temptation and the urge is very strong,
    give a little whistle!
    Give a little whistle!
    Not just a little squeak,
    pucker up and blow.
    And if your whistle’s weak, yell “Juminy Cricket!”

    Take the straight and narrow path
    and if you start to slide,
    give a little whistle!
    Give a little whistle!
    And always let your conscience be your guide


    1. Greg,

      They may be different instars of the same species (see here). Are you sure they’re different species? Some perfunctory internet sleuthing suggests that the top photo is indeed a Walnut Sphinx and both videos confirm that the noisemaker is also of that species (as well as the paper itself). The paper also says that this is the only species known to squeak when disturbed.

      1. If you mouse over the larval pictures there, the instar number appears. This collection of shots shows instars 1 through 5; I can’t find out how many instars this moth is supposed to have, but I think that may be all of them.

        Also, Abbott’s sphinx (Sphecodina abbottii) is known to make noise as well:

        From that page: “If it is pinched or poked, the larva squeaks and bites at the attacker (Wagner 2005).”

        Note that its larvae do not at all resemble either of the spp above:

        Perhaps the first vid is of some tropical or at least non-North American species of Sphinx moth? At least it doesn’t appear to be in my Caterpillars of Eastern North America. (Which says that the Walnut Sphinx, like the Abbott’s, also thrashes when disturbed, something the larva in the first vid does not.)

      2. Those images all look like the caterpillar in the still pictures, and those include at least most if not all of the instars. So it does look like the one in the video is indeed a different species.
        An image search for walnut sphinx turns up the same sort of caterpillar in the picture except for one link that has the video with the vary different looking caterpillar. That one is written in a very different language. It is possible that the people who did the video mis-identifed it, or that they use the same common name. If so, then we have at least two kinds of squeaky caterpillars. Two is always better than one, I say.

  2. As for that second video (a side note to the main point of this post), where the text reads “Probably a way of letting the hawk know that it has been spotted,” I think that may not be probable at all, if we are to understand the work of Amotz and Avishag Zahavi. I’m referring, of course, to their book “The Handicap Principle,” which has long been one of my favorite science books.

    1. Yes, very interesting book, though I think the Zahavis try to make the handicap mechanism do a lot of extra work that can be done simpler by inclusive fitness and the geometric and physical restrictions on senses and communication modes. I haven’t seen a detailed critique that took their examples apart in this way, but it’s over a decade since I was looking for it.

  3. Shows what I know. All this time, I thought it was just because they forgot the words.

    In other words, they should have done the same tests using warblers with the A8 spiracles of the caterpillar occluded with latex.

    How common are these caterpillars? If common, seems like an experiment well within the range of a student or amateur to carry out the experiment.

    …but I’m also wondering how it would affect the caterpillar’s ability to breathe…?

    Also: how much consistency do these alarm calls have across different bird species? If it’s just the warblers that the caterpillars are mimicking, and, say, grackles have an entirely different warning call….


      1. That is an excellent idea. Using artificial prey to test predator reactions is a tried and true method to look into these things.

  4. Playing the sound alone when no caterpillars were around – that would also be something to try – would that send birds to hide?

  5. So if these caterpillars can create sounds to deter predators, does that also mean they can hear? I would think they would have to if they adapted the whistle…I guess I’m thinking they have ears, but I doubt that. Are they utilizing some other form of hearing? I don’t see antennae which could possibly detect sound. It boggles me mind.

    1. Perhaps I’m misreading it, but I don’t think there’s any suggestion that the caterpillars can modulate the sound of the squeak or learn different squeaks for use in different situations. Rather, natural selection has shaped the spiracles to emit a squeak that’s effective in deterring the most likely predator.

    2. That’s the beauty of Evolution. The caterpillar doesn’t need to hear a thing, though I wouldn’t at all be surprised if they’ve got typical insect hearing. All that needs to happen is for there to be a statistical advantage to the reflexive behavior.

      Remember, it’s not like there’s any meaningful amounts of intelligence in what passes for a brain in a caterpillar. Certainly not enough to do the sorts of signal processing to learn more effective modes of vocalization in response to predator behavior.


      1. Thanks for the clarification Gregory, Jerry and Ben, that makes sense to me and now I’m not boggled.

      2. So, there’s absolutely no chance of teaching it to whistle ‘Dixie’?

        I’m so disappointed.


  6. It reminds me of the famous hissing cockroaches from Madagascar. These big insects emit a loud hiss when disturbed. Given that they are found on the ground, it makes some sense that the hissing sound would remind a predator of a hissing snake.

  7. only this pair, when covered, eliminated no sound.

    You have an extra negation here. What you’re trying to say, I think, is that this is the only pair for which covering the spiracles suppressed the sound.

  8. If this is the only species and the only caterpiller that can do this, what is next. Determine the gene responsible and transfer to other caterpillers?

  9. Another fantastic science post. Thank-you for the time and effort you put into these. They remain my favourite flavour of WEIT posts.

  10. Aw man! I wish I would have known that when I caught that caterpillar… Nonetheless, great article! Keep up the good work. 🙂

Leave a Reply