Bats use sonar to locate cactus flowers

Many cacti bloom at night, and that’s when the bats and moths are available for pollination. But since most bats locate prey by echolocation via sonar, how do they find the flowers? This is particularly crucial when those flowers stick out of cacti, for the bats approach the flower fast and hover before it while lapping nectar; if they miss, they could get impaled. And indeed, in  a very good NYT piece by Natalie Angier, she reports that impaled bats—those who have missed their mark—can be found on cacti (see photo at very bottom).

That means there’s strong natural selection on the bats to find the flowers and home in on them accurately. How do they do this? A reference in Angier’s article led me to a piece published last year in PLoS ONE  (reference and free link below) showing that, at least at final approach, the bats use echolocation, emitting a special kind of call as they approach the flower. And apparently the flowers themselves have characteristics that help guide the bats to the nectar tube, which I suppose scientists could have realized a long time ago had they thought about it.

The authors, Tania Gonzalez-Terrazas et al., recorded the sonar calls of a group of wild-caught Lesser Long-Nosed BatsLeptonycteris yerbabuenae), a Central and North American species that subsists on nectar from night-blooming plants, particularly columnar cacti. Here’s a photo and a video of this species feeding:

Note that the bats hover when feeding:

To test how the bats locate flowers, they used two targets (tested separately): a flowering columnar cactus (Pachyceroeus pringlei) and the “control”, an acrylic hemisphere (set up next to a columnar but flowerless cactus) filled with sugar water (the flower was also filled with an equal amount of sugar water). The bats’ calls were recorded with an ultrasound microphone and their movements tracked with an infrared camera. Here’s the setup:

(From the paper): Setup used during the flight cage experiments with both targets. We worked with one target at a time, respectively, each fixed to a cactus branch. The flight and echolocation behavior was recorded with an ultrasound microphone and two synchronized video cameras supported by stroboscopic light. Flight cage dimensions (4m x 4m x 3m).

The results can be stated succinctly:

  • The bats approach and locate the flower and sphere differently. They use sonar to find both, emitting calls, but also emit a special “buzzlike” call, consisting of reduced interval between pulses, and reduced peak amplitude, only when approaching the flower. This may enable them to locate the flower opening accurately. That special call is not used when approaching the acrylic “flower”. This is in fact the only report of a “terminal-buzz-like phase” in a nectar-feeding bat, though similar calls are used by aerial insect-eating bats as they approach their prey.


  • The acrylic hemisphere, containing nectar, was never used for food; bats aborted their approach flight to it at the last second, while all the flowers were found (they may also use scent). The special call emitted when approaching flowers indicates that sonar is also important. Here’s a diagram showing how bats approach both targets (three different flight paths), with the dots indicating when the bats emit a call. The green dot shows the position of the microphone and the pink star the location of the target.

“A” is a cactus flower and “B” the acrylic hemisphere. Note that the bats never got to the hemisphere, and their calling is more variable and erratic. The top two panels are side views, the bottom two top views:

As you see, the bats have a much more uniform approach to the real flower, always coming from below and never getting above it. You can also see that all approaches to the sphere were aborted.

  • Finally, the authors suggest that the cactus flower itself has evolved to guide the bat accurately to the nectar. After all, it’s in the cactus’s reproductive interests to avoid having the bat miss the flower and get impaled! Here’s the suggestion from the paper:

“The echoes reflected from the inside of the long tube of the cactus flower might function as an acoustic guide that provides the bat with detailed information on the location of the flower opening and the orientation of the floral tube. In addition, cacti as well as many bat-pollinated flowers have particularly robust and rigid petals that may indicate the flower opening not only visually but also acoustically. Leptonycteris yerbabuenae only emitted the long terminal group while approaching the flower. We suggest that specific echo-acoustic characteristics of the flower guide the bats directly into the opening of the flower, and emission of an extended terminal group of calls at the end of the approach sequence aids this task.”

It would seem obvious that if pollination—reproduction for a cactus and a crucial factor in preserving its genes—depends on bats, then the flower itself might evolve to lure the bats as well as make sure it doesn’t get impaled on the cactus’s own spines. Well, that’s obvious in retrospect, but why wasn’t this realized until now? I can surmise only that it comes from a bias towards studying the evolution of animals compared to plants (there are a lot more zoologists than botanists). Yet even Darwin devoted several of his books to plant evolution. Let us not ignore our green relatives!

Here’s a bat that missed, impaled on a cholla cactus:

Natural selection; photo from “what I saw. . .


Gonzalez-Terrazas, T. P., Koblitz, J. C., Fleming, T. H., Medellín, R. A., Kalko, E. K. V., Schnitzler, H.-U., & Tschapka, M. (2016). How Nectar-Feeding Bats Localize their Food: Echolocation Behavior of Leptonycteris yerbabuenae Approaching Cactus Flowers. PLoS ONE, 11(9), e0163492.


  1. eric
    Posted April 26, 2017 at 9:52 am | Permalink

    It would seem obvious that if pollination—reproduction for a cactus and a crucial factor in preserving its genes—depends on bats, then the flower itself might evolve to lure the bats as well as make sure it doesn’t get impaled on the cactus’s own spines.

    It makes me wonder why the cactus hasn’t also evolved to have less spines growing near the flower sites.

    AIUI its mostly lizards, hogs, and rabbits that try and eat cactus, or at least tear into them for the moisture. Ground dwellers, IOW. So my admittedly naïve or lay person’s thought is that cacti with less needles around flower sites high off the ground might gain more from better bat pollination than it would lose from the higher risk of predation.

    (I’m aware some owls nest in cacti too, but AIUI they don’t necessarily kill the plant so there might not be as much evolutionary pressure to ‘fight’ them.)

    • rickflick
      Posted April 26, 2017 at 12:15 pm | Permalink

      Good idea. But, I can think of a reason this doesn’t happen: the evolutionary landscape might be such that reducing spines in specific areas of the plant while preserving them in others is “difficult” to achieve. Reshaping the blossom to help the bats in location might be relatively “easy” and good enough so that a sufficient supply of pollinators is maintained.

      • Richard Bond
        Posted April 26, 2017 at 1:22 pm | Permalink

        In he UK, the common holly (Ilex aquifolium) has very spiny leaves low down, but greatly reduced spines above about two metres high. Apparently it has learned that there are no giraffes in Britain.

        • rickflick
          Posted April 26, 2017 at 5:29 pm | Permalink

          A similar pattern was mentioned appearing on a tree in New Zealand reported by Jerry. The plant had strong spines until it reached a size that protected it from browsing by extinct birds. So, spines can be deployed in defense in some instances, but it’s far from universal. Some cactus plants may have some form of this as well. But apparently the bat pollinated ones did not find the idea a convenient adaptation.

          • infiniteimprobabilit
            Posted April 27, 2017 at 2:50 am | Permalink

            Almost but not quite, IIRC. The plant was the lancewood, which as a juvenile has long thin very tough leaves – unattractive to eat. But when it grows to sufficient height, it then (if I recall Jerry’s post accurately) develops far more ‘normal’ leaves. (I didn’t know this, I only know what the distinctive juvenile plant looks like).


    • Gregory Kusnick
      Posted April 30, 2017 at 11:33 am | Permalink

      The cactus clearly benefits from maximizing the number of flowers successfully pollinated. But I don’t see why it should care about minimizing the number of dead bats. Even if bats are scarce enough that bat population is a limiting factor in cactus reproduction, it’s hard to see how genes for preserving bat lives would confer an advantage on the particular cactus carrying them. So where is the selection pressure for minimizing bat mortality supposed to come from?

      • infiniteimprobabilit
        Posted April 30, 2017 at 6:40 pm | Permalink

        A flower that has an injured bat flapping near it (or a dead one) might well be avoided by other bats?


        • Gregory Kusnick
          Posted April 30, 2017 at 7:43 pm | Permalink

          Sure, we can concoct just-so stories about why cacti might want to keep bats alive — or why they might want them dead. (A bit of fertilizer, perhaps?)

          The point is that any such selection pressure is not an inevitable logical consequence of the pressure to maximize pollination, and therefore requires additional explanation in its own terms.

          • infiniteimprobabilit
            Posted April 30, 2017 at 11:34 pm | Permalink

            Not a matter of what the cactus ‘wants’. If bats avoid the flowers with bats stuck near them, the genes for having spines near the flowers will spread less than the genes from flowers that have a ‘safe zone’ around them.


            • Gregory Kusnick
              Posted May 1, 2017 at 1:59 am | Permalink

              The key word is “if”. The fact that cactus flowers are optimized for echolocation provides no evidence one way or the other about whether safe zones are advantageous to cacti. The apparent lack of such safe zones suggests that they aren’t, and that the bats care more about getting the nectar than about avoiding bat corpses.

              • infiniteimprobabilit
                Posted May 1, 2017 at 6:29 am | Permalink

                Re-reading, I see the ‘safe zones’ were just a speculation of eric’s (or rather, a query as to why cacti don’t have them). But the same end might be served by optimising the flower for echolocation.

                (At no point was I imputing altruistic motives to the cactus).


  2. Posted April 26, 2017 at 9:53 am | Permalink

    Very cool, thanks for posting. We here in Minnesota are hoping our local bat populations will recover soon from White Nose Syndrome.

    (I’m sure der Drumpfenführer and Sarah Pain would wonder out loud what this research is “good” for.)

  3. Posted April 26, 2017 at 10:00 am | Permalink

    Cool. Thanks.

  4. Christopher
    Posted April 26, 2017 at 10:02 am | Permalink

    It would be interesting to know what went wrong that led the bat to impale itself. I’ve experienced a similar error many years ago, in broad daylight, with a bat on the Lake of the Ozarks. The bats love to roost and feed over the lake on the docks, especially those with lights on at night. This one, however, was flying during the day, came flying straight at me, I ducked, and it slammed into a storage cabinet next to me, stunned for a good few minutes, then flew off. I can only imagine that it had been disturbed from its roost, confused, then panicked when I was in the way, and then crashed, but who knows.

  5. Posted April 26, 2017 at 10:11 am | Permalink

    “why wasn’t this realized until now?”

    I think it was realized long ago that flower cues guided the bats; but understandably, people concentraed on cues that we could percieve– odor and color. Night-flowering cacti are bright white, often very fragrant, and often have very long ovaries keeping the flower far from the spines. The novelty of this study is that it shows there are non-visual cues (or perhaps the same cues as used in vision, just perceived through sonar instead).

    It’s interesting that bats can eventually learn to get nectar from fake flowers that have no smell and no white color. In tropical countries they come to hummingbird feeders at night. I have seen this in Colombia and Ecuador. I don’t think it is known how they learned to do this.

    • darrelle
      Posted April 26, 2017 at 12:00 pm | Permalink


      Do the same bats feed on any of the same flowers that the local hummingbirds do? If so perhaps they’ve evolved a behavior of cuing off of hummingbirds in some way?

  6. W.Benson
    Posted April 26, 2017 at 10:18 am | Permalink

    The standard view (as I learned) it is that bat flowers are white and have a strong fragrance that aid in their location by bats, and the petals are tough to support bat’s weight when it alights to feed (although may bat species feed while in flight). Bat flowers functioning sonar-reflecting navigation devises makes good sense.

  7. ThyroidPlanet
    Posted April 26, 2017 at 10:23 am | Permalink


  8. Mark Sturtevant
    Posted April 26, 2017 at 10:39 am | Permalink

    Very interesting. I had thought that nectar feeding bats were of a kind that did not use sonar, but I am glad to be wrong.
    Of course it had to be a cholla cactus that impaled the bat. Those cacti are especially nasty.

  9. Posted April 26, 2017 at 10:42 am | Permalink

    that’s an interesting info.. thanks for sharing

  10. Randy schenck
    Posted April 26, 2017 at 11:06 am | Permalink

    Very interesting topic. Apparently bats develop different echolocation systems depending on the food source. Can’t help thinking of echocardiography in the medical world as I have had a few of these.

  11. Mark R.
    Posted April 26, 2017 at 11:48 am | Permalink


  12. ThyroidPlanet
    Posted April 26, 2017 at 2:16 pm | Permalink

    “Many cacti”

    Can’t help this – even though I’m not a Greek or Latin scholar:

    The word “cactus”, like “octopus”, is of Greek origin, so even though it’s ok to use “cacti”, I think it’s pedantic to do so, when “cactuses” is entirely adequate. However, I don’t know the precise word for the Greek plural of “cactus” – for “octopus” it is in fact “octopodes” … “Cactodes”?

    • squidmaster
      Posted April 26, 2017 at 10:31 pm | Permalink

      The greek word was kaktos, pl. kaktoi, which is Latinized to cactus, cacti, so ‘cacti’ is indeed the latin plural, although ‘cactuses’ is listed as OK by some dictionaries. BTW, the Greeks and Romans had never seen actual kaktoi. Kaktos meant artichoke or something like that.

      The plural of oktopos was, indeed, oktopodes, because the greeks called our cephalopod friend okto pous (eight foot) and the plural was okto podes (eight feet). I don’t think there was ever a Latin word, octopus, but the word was made up in the 19th century some time.

      I like this trivia because it informs our understanding of language, but don’t care terribly what people write.

      • ThyroidPlanet
        Posted April 27, 2017 at 12:33 am | Permalink

        I look at the word in Google’s tree and the furthest back word is Greek – doesn’t that mean it’s origin is Greek? Isn’t Latinization just a fancy way of saying it isn’t originally Latin?

  13. Posted April 26, 2017 at 6:48 pm | Permalink

    Thanks, great article – good to see Elizabeth Kalko’s name on the article (in memoriam). jumping cholla cactus is thoroughly obnoxious (been jumped on numerous times) – and it has hooked spines – so the impaled animal can’t escape.
    To clarify – there are 2 groups of bats that are nectar feeders – the ‘vegetarian’ micro-bats which are found only in the Americas (which have sonar), and the ‘vegetarian’ mega-bats found in Africa, Asia and the Pacific (which don’t have sonar). The mega-bats tend to be non-specialists (eat anything vegetable), whereas the micro-bats tend to be specialists.
    Hope this helps .. the top image looks suspiciously like one of Merlin Tuttle’s, yes?

  14. squidmaster
    Posted April 26, 2017 at 10:14 pm | Permalink

    I’m in AZ ATM and have been fascinated with the biology of saguaro cacti. These are the big cacti with multiple arms that are familiar from westerns. The cacti are quite long lived (up to 250 years) and they grow slowly. They don’t become sexually mature until age 70, about the time they put out their first branches.

    The flowers form on the top of the arms and main trunk and look like a crown of white blossoms. The petals are quite firm and waxy; the flowers stay open for several days. they are pollinated by birds, bats and insects. The flowers are open day and night.

    Gila woodpeckers excavate cavities in saguaro for their nests and the cactus secretes a hard shell around the excavation. You can find these secretions long after the cactus has died. If the woodpeckers abandon the nest, then sometimes elf owls take up residence.

    There are many other ecological connections of saguaro to the rest of the Sonoran desert ecosystem that are worth a read.

  15. kieran
    Posted April 27, 2017 at 6:23 am | Permalink

    As a botanist it’s easier because you don’t anthropomorphize the subject you’re looking at!

  16. Bernie
    Posted April 27, 2017 at 3:22 pm | Permalink

    I became personally aware of their use of echolocation in this regard one night while doing fieldwork with screech owls in Arizona. Evidently they mistook me for a cactus, and my headlamp (despite being on low) for a flower, at least upon initial approach. Thankfully, feeding attempts were abandoned quickly, but I was “facepalmed” by several of these bats over the course of a night. A little freaky when it first happens…

  17. Posted May 5, 2017 at 4:18 am | Permalink

    Very interesting — thanks for your concise and clear summary!

%d bloggers like this: