Bats jamming bats: Food competition through sonar interference

See allHide authors and affiliations

Science  07 Nov 2014:
Vol. 346, Issue 6210, pp. 745-747
DOI: 10.1126/science.1259512

Competing bats jam one another's signal

Animals that live in large social colonies may benefit from many aspects of group living, but also have to contend with many of the downsides of living and foraging, with countless neighbors. Corcoran and Conner show that Mexican free-tailed bats, which live in colonies that can number in the hundreds of thousands, deal with this high level of competition for food by actively jamming competitors' echolocation. The interfering bats produce an ultrasonic signal just as the foraging bat produces its feeding call, effectively jamming the echolocation signal and causing the forager to miss its target.

Science, this issue p. 745


Communication signals are susceptible to interference (“jamming”) from conspecifics and other sources. Many active sensing animals, including bats and electric fish, alter the frequency of their emissions to avoid inadvertent jamming from conspecifics. We demonstrated that echolocating bats adaptively jam conspecifics during competitions for food. Three-dimensional flight path reconstructions and audio-video field recordings of foraging bats (Tadarida brasiliensis) revealed extended interactions in which bats emitted sinusoidal frequency-modulated ultrasonic signals that interfered with the echolocation of conspecifics attacking insect prey. Playbacks of the jamming call, but not of control sounds, caused bats to miss insect targets. This study demonstrates intraspecific food competition through active disruption of a competitor’s sensing during food acquisition.

Active sensory systems such as echolocation and electrolocation allow animals to exploit habitats where vision is ineffective (1) but are also susceptible to “jamming” (i.e., signal interference) from conspecifics. Electric fish and bats alter the frequency of electric and acoustic emissions, respectively, to avoid jamming from conspecifics [the “jamming avoidance response” (25)]. Although some insects adaptively jam their bat predators as a defense (6, 7), jamming among echolocating animals appears to occur only inadvertently (24).

Mexican free-tailed bats (Tadarida brasiliensis) form the largest known colonies of any active sensing animal, with some caves housing more than one million individuals (8). Studies of captive, roosting T. brasiliensis have revealed a complex social system mediated by a vocal repertoire of at least 15 communication, or “social,” calls (9). Echolocation signals also mediate social interactions outside the roost, as T. brasiliensis exhibit jamming avoidance responses (2, 3) and are attracted to conspecific feeding buzzes, presumably to find insect prey (10).

Here we report results of field observations and playback experiments (Fig. 1) that demonstrated the jamming function of a previously unstudied social call from T. brasiliensis (Fig. 2, A to C), which we term the sinusoidal frequency-modulated (sinFM) call after its spectral-temporal pattern. Initial field observations showed that sinFM calls are produced only when another bat is rapidly emitting echolocation calls (“feeding buzz”) in the late stages of insect pursuit—a time when bats are particularly susceptible to jamming (6, 7). We hypothesized that sinFM calls are used for jamming conspecifics while competing for insect prey. The proposed function of sinFM calls contrasts starkly with reports of bat jamming avoidance (24) and proposed functions of social calls recorded in a foraging context, namely cooperative foraging and food patch defense (1113). Alternatively, sinFM calls could startle bats without disrupting echolocation (14).

Fig. 1 Field experimental setups for (A) field observations and (B) field playbacks.

Bats and insects were attracted to foraging areas by either an ultraviolet (UV) light or a street light. For field observations (A), bat social interactions and insect capture attempts were recorded by a video camera mounted on a spotlight from a 6-m platform. 3D bat flight trajectories were determined by acoustically localizing bat calls using two four-microphone arrays. (B) During playbacks, bats attacked moths tethered to a 0.1-mm monofilament line. Attacks were recorded on two microphones and an infrared (IR) video camera. Ultrasound was broadcast from a speaker placed directly below the tethered moth. (C) Spectrograms of the three acoustic playbacks that were used—sinFM, tone, and noise—and a buzz-phase echolocation call for reference. Detailed methods are available (15). Gridlines are spaced at 1-m intervals.

Fig. 2 Acoustics and results of T. brasiliensis sinFM interactions and playback experiments.

Spectrograms show examples of (A) three-syllable and (B) one-syllable sinFM calls overlapping another bat’s feeding buzz. (C) Close-up of a sinFM syllable from (A). (D) Representative power spectra of sinFM and buzz 2 calls [for buzz 2 definition, see (15)]. Scatter plots show correlation between the duration of the competitor bat’s buzz after start of sinFM and (E) sinFM duration or (F) number of sinFM syllables. Solid lines depict linear regression. (G) Box plots of feeding buzz durations when conspecific sinFM calls were present (n = 68) or absent (n = 185; t test). (H) Capture success at two field sites with and without a competitor bat producing sinFM calls. (I) Capture success as a function of playback type for individual bats attacking tethered moths. All playbacks overlapped the feeding buzz, except sinFM early, which preceded the feeding buzz. *P < 0.05, **P < 0.01, ***P < 0.001 for Fisher’s exact test. n = 185 attacks and n = 145 for no sinFM at SWRS and Animas; n = 74 and n = 12 for sinFM at SWRS and Animas in (H). For the playback experiment (I), n = 57, no playback; n = 18, tone; n = 17, noise; n = 18, sinFM early; and n = 40, sinFM. See Fig. 1C and text for playback descriptions.

We determined the function of sinFM calls by quantifying natural social interactions between bats at two foraging sites in Arizona and New Mexico. We collected data using low-light videography, visual observations with a spotlight, and three-dimensional (3D) reconstruction of bat flight paths, using time-of-arrival-differences of sounds recorded on two microphone arrays (Fig. 1A) (15). In a second field experiment, we broadcast synthesized sinFM calls or controls [(i) no playback, (ii) 40-kHz tones played during the feeding buzz, (iii) noise bursts also played during the buzz, and (iv) sinFM calls played before the buzz; Fig. 1C] to T. brasiliensis attacking moths tethered to a monofilament line approximately 5 m above the ground (Fig. 1B). Stimulus parameters were selected to replicate the natural situation (15).

Acoustically, sinFM calls appear well adapted for jamming conspecifics. In all sinFM field recordings with good signal-to-noise ratio [n = 68 interactions from ≥ 7 individuals (15)], the calls overlapped temporally and spectrally with a conspecific’s feeding buzz (Fig. 2, A to D, and Table 1), with 187 ± 118 ms (mean ± SD) of delay between the onset of the buzz and the onset of the sinFM call. SinFM calls had one to five repeated units (syllables) separated by short silent intervals (Table 1). Total sinFM call duration and the number of sinFM syllables were both highly correlated with the duration of the competitor bat’s feeding buzz after the start of the sinFM call (Fig. 2, E and F). Mean feeding buzz duration was independent of the presence of sinFM calls (Fig. 2G). Therefore, in real time bats appear to dynamically modulate the number of sinFM syllables and the total sinFM call duration, depending on the duration of the conspecific competitor’s feeding buzz.

Table 1 Acoustic properties of T. brasiliensis sinFM and buzz calls.

Syll dur, syllable duration; syll interval, syllable interval; total dur, total duration (including silent intervals) for sinFM, buzz 1, and buzz 2; peak freq, peak frequency; max freq, maximum frequency; min freq, minimum frequency; sweep, sweep rate; sin period, sinusoid period. NA, not applicable.

View this table:

The sinFM call’s rapid frequency modulations occur at a rate that ensures that at least one full FM cycle occurs in the listening window between the successive calls of the other bat’s feeding buzz (Fig. 2C and Table 1). The maximum downward FM rate of sinFM calls is close to that found in feeding buzz calls (Table 1) and may stimulate the same FM-rate–sensitive auditory neurons that encode prey location information (16). FM signals that temporally overlap echoes are known to interfere with the estimation of target distance (17).

Bats were 85.9 and 77.3% less likely to capture insects in the presence of conspecific-produced sinFM calls at the Arizona[Southwestern Research Station (SWRS)] and New Mexico (Animas) field sites, respectively (Fig. 2H; see the figure legend for sample sizes and statistics). Playbacks of sinFM calls to individual bats attacking tethered moths caused capture success to decrease by 73.5% as compared to no playback (Fig. 2I).

Audio and video recordings show bats continuing attacks on prey after hearing sinFM calls from conspecifics (Fig. 2, A and B, and movie S1) and from the ultrasonic speaker (fig. S2 and movies S2 and S3). In natural interactions, bats continued buzzes for 213 ± 145 ms after sinFM calls began, and buzz durations were not abbreviated as compared to attacks lacking sinFM calls (Fig. 2G). In the playback experiment, bats continued attacks both when sinFM calls were broadcast during the feeding buzz (sinFM) and when sinFM calls were broadcast before the feeding buzz (sinFM early; fig. S2). SinFM playbacks only prevented capture when they overlapped the feeding buzz and not when they preceded the buzz (Fig. 2I). These results contradict the primary prediction of the food defense hypothesis: that bats abort attacks on prey after hearing the call (12, 13). In contrast, these results support the predictions of the jamming hypothesis that capture success decreases only when sinFM calls temporally overlap with the feeding buzz and that bats miss prey despite persistent attacking behavior. SinFM calls appear to interfere with the bat’s ability to determine prey position and to successfully coordinate capture.

We used quantitative criteria to examine 3D flight trajectories of bat interactions involving sinFM calls (fig. S1). Sixteen sequences lasting 3 to 10 s had sufficient quality for 3D flight trajectory reconstruction. An example sequence (Fig. 3A) and a summary of all interactions (fig. S1C) show multiple bat competitors alternately and repeatedly attempting to capture insects (indicated by buzz calls) in a central foraging area while another bat makes a sinFM call. SinFM calls do not cause bats to leave the foraging area or move away from prey, as has been shown for food defense social calls in other species (12, 13). Instead, bats frequently circle back to the foraging area after missing prey while a competitor makes sinFM calls. The bats making sinFM calls appear intent on capturing prey rather than chasing away their competitors, as shown by a detailed analysis of bat flight trajectories (Fig. 3B, C).

Fig. 3 Flight behavior of bat interactions involving sinFM calls.

(A) Overhead view of an example behavioral sequence showing two bats alternately producing feeding buzzes while the other bat produces sinFM calls. Numbers indicate time in seconds. A movie of this sequence is also available (movie S4). (B) Diagram of bat flight measurements. Bat flight, bat-bat, and bat-buzz vectors are shown for two time points: the time of the buzz (x symbols) and a time after the buzz (● symbols). The bat flight vector indicates the direction of flight. The bat-bat vector is the direction from the nonbuzzing bat to the buzzing bat. The bat-buzz vector indicates the direction from the nonbuzzing bat to the position of the buzz (and presumed location of prey). The bat-bat and bat-buzz vectors are identical at the time of the buzz and diverge after that time. (C) Flight direction of bats producing sinFM calls relative to the position of the conspecific buzz (red box plots) and the competing bat (blue box plots) at 250-ms time intervals after the beginning of the buzz. Note that bats fly toward the position of the conspecific buzz (prey) and not of the bat competitor. * P < 0.05 and **P < 0.01 for paired t tests comparing values to the 0-ms time point and for comparisons between buzz and bat conditions at each time.

Two acoustic playbacks (tones and noise bursts) further tested the predictions of the jamming hypothesis and competing hypotheses. Tone playbacks had the same spectral level and pulse durations as sinFM calls but were at a single frequency (40 kHz) not likely to jam bat echolocation. Tone playbacks had no effect on capture success (Fig. 2I), indicating that the effect of sinFM calls is not an acoustic startle reaction. Noise playbacks had the same duration and spectral distribution of acoustic energy as sinFM calls (15), and like tones, also did not affect capture success (Fig. 2I). This indicates that the specific spectral-temporal structure of sinFM calls may be important to their jamming function.

The favorable conditions of our playback experiment allowed us to acquire high-quality audio recordings to test whether bats exhibited jamming avoidance in response to playbacks. Bats did not alter the timing of echolocation emissions (fig. S2) or pulse intervals or call durations in response to playbacks (fig. S3). However, bats increased the frequency of multiple acoustic parameters in response to sinFM, sinFM early, and tone playbacks (fig. S3). This behavior is similar to how T. brasiliensis avoid jamming from conspecific echolocation (2, 3); however, it does not allow bats to avoid jamming from sinFM calls, as evidenced by decreased capture success (Fig. 2I).

Acoustic data, capture rates, and 3D flight behavior from naturally occurring interactions and playback experiments provide compelling evidence supporting the hypothesized jamming function of sinFM calls. These results provide an extraordinary example of interference competition through disruption of a competitor’s senses, a phenomenon never before documented in animals. Although echolocation allows bats to be the dominant nocturnal predator of night skies, it also makes them susceptible to jamming, a vulnerability now known to be exploited by bat competitors.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S3

References (1822)

Movies S1 to S4

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: We thank Z. Walker, K. Roman, D. Margius, N. Dowdy, and K. Irvin for field assistance; the Southwestern Research Station staff for coordination of activities; and C. Moss for thoughtful review of the manuscript. Funding was provided by the Wake Forest University Summer Fellowship Program, the American Museum of Natural History (Theodore Roosevelt Grant), an institutional training grant (UMD T32 DC-00046) from the National Institute of Deafness and Communicative Disorders of the National Institutes of Health, and the National Science Foundation (grant number IOS-1257248). Data are available in the supplementary materials.
View Abstract

Stay Connected to Science

Navigate This Article