Report
Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls Graphical Abstract
Authors Robert D. Magrath, Tonya M. Haff, Jessica R. McLachlan, Branislav Igic
Correspondence [email protected]
In Brief Many vertebrates get information about danger by eavesdropping on other species’ alarm calls. Magrath et al. show experimentally that a wild bird species can learn to recognize previously unfamiliar sounds as alarm calls, supporting indirect evidence that learning is critical in establishing anti-predator ‘‘information webs’’ in natural communities.
Highlights d
Vertebrates benefit by eavesdropping on other species’ alarm calls
d
We used model predators and playback to test if wild birds learn about novel alarms
d
Individual birds learned to recognize novel sounds as alarm calls
d
Such rapid learning explains eavesdropping on acoustically diverse alarm calls
Magrath et al., 2015, Current Biology 25, 1–4 August 3, 2015 ª2015 Elsevier Ltd All rights reserved http://dx.doi.org/10.1016/j.cub.2015.06.028
Please cite this article in press as: Magrath et al., Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls, Current Biology (2015), http:// dx.doi.org/10.1016/j.cub.2015.06.028
Current Biology
Report Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls Robert D. Magrath,1,* Tonya M. Haff,1 Jessica R. McLachlan,1 and Branislav Igic1,2 1Division
of Evolution, Ecology and Genetics, Research School of Biology, Australian National University, Canberra 2601, Australia address: Department of Biology, University of Akron, Akron, OH 44325-3908, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.cub.2015.06.028 2Present
SUMMARY
Many vertebrates gain critical information about danger by eavesdropping on other species’ alarm calls [1], providing an excellent context in which to study information flow among species in animal communities [2–4]. A fundamental but unresolved question is how individuals recognize other species’ alarm calls. Although individuals respond to heterospecific calls that are acoustically similar to their own, alarms vary greatly among species, and eavesdropping probably also requires learning [1]. Surprisingly, however, we lack studies demonstrating such learning. Here, we show experimentally that individual wild superb fairy-wrens, Malurus cyaneus, can learn to recognize previously unfamiliar alarm calls. We trained individuals by broadcasting unfamiliar sounds while simultaneously presenting gliding predatory birds. Fairy-wrens in the experiment originally ignored these sounds, but most fled in response to the sounds after two days’ training. The learned response was not due to increased responsiveness in general or to sensitization following repeated exposure and was independent of sound structure. Learning can therefore help explain the taxonomic diversity of eavesdropping and the refining of behavior to suit the local community. In combination with previous work on unfamiliar predator recognition (e.g., [5]), our results imply rapid spread of anti-predator behavior within wild populations and suggest methods for training captive-bred animals before release into the wild [6]. A remaining challenge is to assess the importance and consequences of direct association of unfamiliar sounds with predators, compared with social learning—such as associating unfamiliar sounds with conspecific alarms. RESULTS Playback experiments confirm that eavesdropping on heterospecifics occurs in terrestrial vertebrates worldwide, among both closely and distantly related species, and may be almost universal in birds and mammals [1]. Eavesdropping entails gain-
ing information from calls intended for others, can provide detailed information on danger, and helps structure animal communities by promoting mixed-species groups [1–4]. In some cases, individuals respond to unfamiliar heterospecific alarm calls similarly to conspecific calls because they share similar acoustic features (e.g., [7–12]). However, alarm calls are extremely variable among species, so community-wide eavesdropping probably requires learning (e.g., [13–17]). Despite its potential importance in explaining eavesdropping among species, the evidence for learned recognition of alarm calls is insufficient. First, playback experiments have merely probed natural variation in response to alarm calls (e.g., [13– 16, 18]), whereas demonstrating learning requires experimentally changing the behavior of individuals through manipulation of their experience [19]. Natural variation in response to alarm calls can be confounded by uncontrolled factors, such as physiological development instead of learning causing changes with age [20], or environmental or genetic differences rather than learning causing different responses in different locations [21– 23]. Second, the only experimental study to investigate learned recognition of alarm calls directly did not rule out alternative explanations [24]. Five golden-mantled ground squirrel (Spermophilus lateralis) individuals appeared to learn to recognize novel alarm calls, but the results were also consistent with increased wariness in general, sensitization following repeated exposure, and/or behavioral responses to the acoustic differences between training and control sounds [19, 24]. We tested experimentally whether the superb fairy-wren Malurus cyaneus, a small passerine bird, can learn to recognize unfamiliar sounds as alarm calls. Previous work has shown that fairy-wrens flee to cover in response to unfamiliar sounds that are acoustically very similar to conspecific aerial alarm calls— those given in response to predatory birds in flight—but usually do not flee in response to calls that are acoustically different [7, 8]. By contrast, fairy-wrens flee in response to alarm calls of locally common species regardless of acoustic differences [13, 14], which suggests that individuals learn to recognize the alarm calls of local species but does not rule out developmental, environmental, or genetic differences. We therefore used paired playback and predator model presentations in the wild to test directly whether individuals can learn to recognize previously unfamiliar sounds as alarm calls. We also tested alternative explanations, including response due to increased wariness, sensitization, or acoustic structure. We trained individual fairy-wrens to respond to one of two different unfamiliar sounds (Table 1; Figure 1A; Supplemental Experimental Procedures). First, we tested the response of
Current Biology 25, 1–4, August 3, 2015 ª2015 Elsevier Ltd All rights reserved 1
Please cite this article in press as: Magrath et al., Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls, Current Biology (2015), http:// dx.doi.org/10.1016/j.cub.2015.06.028
Table 1. Experimental Design Stage of Experiment Treatment
Pre-training
Training (eight playbacks over two days)
Post-training 1 (same day)
Post-training 2 (next day)
Predator treatment
sound 1 and sound 2 playbacks
looming playback of sound 1 or sound 2 with predator model
sound 1 and sound 2 playbacks
sound 1 and sound 2 playbacks
Looming control
sound 1 and sound 2 playbacks
looming playback of sound 1 or sound 2 without predator model
sound 1 and sound 2 playbacks
Non-looming control
sound 1 and sound 2 playbacks
non-looming playback of sound 1 or sound 2 without predator model
sound 1 and sound 2 playbacks
Birds were first tested for response to two novel sounds (allopatric thornbill alarm call or synthetic sound) in pre-training playbacks, to ensure that birds included in the experiment did not flee before training, then exposed to one of the two sounds over a two-day period, and finally tested in post-training playbacks (n = 10 for each treatment). Five birds in each treatment were exposed to each of the two sounds, with the other acting as a within-treatment, unexposed control. The order of playback within pre- and post-training tests was alternated. The looming playback consisted of a sequence of notes with increasing tempo and amplitude, mimicking a natural pattern of alarm calls to an approaching threat. The non-looming playback consisted of the same notes, but in a slower tempo and random order. The predator treatment consisted of a looming playback sequence with simultaneous presentation of a gliding predator model. The controls had looming or non-looming playback sequences without a predator presentation. Figure 1 shows the novel sounds and predator models, Figure S1 shows the temporal sequences, and further details are in the text and Supplemental Experimental Procedures.
individuals to each unfamiliar sound (pre-training), to ensure that the individuals included in the experiment did not flee in response to either sound before training. Next, we exposed each individual to eight playbacks of either one or the other of those two sounds over two days (training) and then finally tested the response to each sound after training (post-training). The training period included three treatments, each carried out on ten different birds in separate social groups: the predator treatment, in which unfamiliar sounds were accompanied by gliding predator models (Figure 1B), and two control treatments in which the sounds were presented without predators. The two controls used the same unfamiliar sounds as the predator treatment, but in one control they were presented in a different temporal pattern. The predator treatment and one of the controls used a ‘‘looming’’ playback consisting of a sequence of notes with increasing tempo and amplitude, mimicking a natural pattern of alarm calls to an approaching threat. The non-looming control consisted of the same notes, but in a slower tempo and random order, which is more natural for non-alarm vocalizations (Table 1; Figure S1; Supplemental Experimental Procedures). The two controls were included in case a looming pattern facilitated learning or was itself a cue used in learning to recognize unfamiliar alarm calls; in some species, looming stimuli can prompt attention or be aversive [25–27]. Half of the birds in each treatment were trained to one sound and half to the other, to ensure that sound type did not confound interpretation. The critical prediction for learned recognition of specific sounds as alarm calls in the predator treatment was flight to cover in response to the exposed sound but not in response to the unexposed sound when tested post-training. Response to Unfamiliar Sounds Most individuals learned to treat unfamiliar sounds as alarm calls after they were presented with a gliding predator. After training, 9 of 10 fairy-wrens fled to cover after playback of the exposed (training) sound, and of these, only one also fled in response to the unexposed sound (Figure 1C; same-day playbacks, McNemar’s test, exact binomial one-tailed p = 0.004). There
was a similar pattern when tested again at least one day later, with 7 of 10 fleeing in response to the exposed sound and only two also fleeing in response to the unexposed sound (McNemar’s test, exact binomial one-tailed p = 0.03). Combining data over the same- and next-day playbacks showed that 8 of 10 birds fled more often in response to the exposed than in response to the unexposed sound, with none showing the opposite pattern (McNemar’s test, exact binomial one-tailed p = 0.004; proportion fleeing did not differ significantly between days: binomial test on paired data, without a directional hypothesis, two-tailed p = 0.25). An equal number of birds learned to respond to each of the two different sounds (4 of 5 learned each). The response of birds to the exposed sounds in the predator treatment was not due to sensitization. None of the 10 birds in either control treatment fled in response to the exposed or unexposed sound when tested post-training (predator treatment compared to each control in same-day post-training playbacks: McNemar’s test, exact binomial one-tailed p = 0.004). DISCUSSION Learning to Recognize Alarm Calls Our results demonstrate learned recognition, supporting indirect evidence that learning provides a mechanism to explain widespread eavesdropping in vertebrates. After only two days’ training, fairy-wrens responded to previously unfamiliar sounds as if they were heterospecific aerial alarm calls warning of flying predators. The fairy-wrens learned to flee in response to either sound, depending on which was associated with the appearance of the predator model during training, showing that acoustic structure did not explain the response to trained versus untrained sounds. They did not respond after training to playback of the unexposed sound, showing that they discriminated between the sounds and did not flee in response to the exposed (training) sound simply because they had become wary of playbacks in general. The learned recognition of the predator-trained sound was not due to sensitization. No bird responded to sounds after
2 Current Biology 25, 1–4, August 3, 2015 ª2015 Elsevier Ltd All rights reserved
Please cite this article in press as: Magrath et al., Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls, Current Biology (2015), http:// dx.doi.org/10.1016/j.cub.2015.06.028
Figure 1. Experimental Sounds, Predator Models, and Results (A) Spectrograms of single notes that were used to compose unfamiliar sounds and training sequences. TB, allopatric, chestnut-rumped thornbill aerial alarm; SYN, note synthesized on computer; FW, superb fairy-wren aerial alarm (for comparison). (B) Ventral views of predator models, based on pied currawongs (CW) and collared sparrowhawks (SH). Superb fairy-wrens (FW) are shown at the same scale; photographs by Simon Bennett. (C) Post-training response of fairy-wrens to unfamiliar sounds that had been paired with the appearance of predator models (exposed), compared with their response to unexposed sounds of equal amplitude. Dark bars, same-day playbacks; hatched bars, next-day playbacks.
repeated playback alone in the control treatments, even if the sounds were played in a looming pattern of increasing amplitude and rate (Figure S1). The looming sequence was the pattern we associated with the predator model to simulate a natural pattern of calling when a predator approaches. Learning therefore entailed associating the unfamiliar sound with danger, potentially through direct association of the sound with the predator (Pavlovian associative learning [19]). It is also possible that individuals associated the unfamiliar sounds with anti-predator behavior of other individuals, such as fleeing or alarm calling, and so there could also have been social learning [19, 28]. Distinguishing between direct and social learning would be valuable because the mechanism could affect the rate of learning; currently we do not know whether it is essential to detect the predator directly, or whether individuals can learn to associate conspecific alarm calls or behavior with unfamiliar sounds. Regardless of the specific mechanism, individuals learned to treat appropriate sounds as alarm calls.
Ecological and Evolutionary Consequences Our study addresses the ecological question of learning to eavesdrop within natural communities. This requires testing for learned recognition of sounds that have properties of alarm calls but are unfamiliar, as if a new alarm-calling species had joined the community. A related issue is whether alarm signal design affects learning. Abrupt, broad-band, and non-linear sounds—such as some alarm calls, but not the sounds in this study—can capture attention and in doing so may facilitate learning about danger [29]. Furthermore, looming sounds can prompt attention or be aversive [25, 27], as can similarities to conspecific or familiar heterospecific alarm calls [8], and so either might facilitate learning. Although our playbacks were presented in a looming sequence and had general characteristics of aerial alarm calls, these features did not determine which sound became recognized as an alarm. Overall, the evolution and use of alarm calls could depend in part on what acoustic features affect both learned and unlearned responses, as well as the benefits and costs of being eavesdropped on [1, 16, 30, 31]. In addition to potentially affecting the evolution of signal design, the ability to learn to recognize heterospecific signals is important for understanding the establishment of community ‘‘information webs’’ [32] and could have conservation significance. Previous experiments show that alarm calls can promote learning about predators themselves [33, 34], whereas we show here that predators can prompt learning about heterospecific alarm calls. Together these could lead to the rapid spread of anti-predator behavior, with known predators facilitating learning about unfamiliar alarm calls, and learned recognition of calls aiding subsequent recognition of unfamiliar predators. Rapid spread of anti-predator behavior is likely to help populations cope with changing community composition, as is occurring with climate change and invasive species [35, 36]. Our methods also suggest that in conservation programs it would be possible to train captive-bred individuals to recognize heterospecific signals of danger, and not just predators themselves [6, 37, 38]. Such work should also consider the importance of social learning and examine the retention of learned anti-predator behavior. SUPPLEMENTAL INFORMATION Supplemental Information includes one figure and Supplemental Experimental Procedures and can be found with this article online at http://dx.doi.org/10. 1016/j.cub.2015.06.028. ACKNOWLEDGMENTS We thank B. Pitcher and A. Smith for models; T.H. Bennett for work on a pilot experiment; and K. Cain, A. Cockburn, N. Langmore, P. Re˛k, and three reviewers for detailed and insightful comments. This work was funded by the Australian National University and was conducted under permits from Environment ACT, the Australian Bird and Bat Banding Scheme, Australian National Botanic Gardens, and the ANU Ethics Committee. Received: March 12, 2015 Revised: May 7, 2015 Accepted: June 11, 2015 Published: July 16, 2015
Current Biology 25, 1–4, August 3, 2015 ª2015 Elsevier Ltd All rights reserved 3
Please cite this article in press as: Magrath et al., Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls, Current Biology (2015), http:// dx.doi.org/10.1016/j.cub.2015.06.028
REFERENCES 1. Magrath, R.D., Haff, T.M., Fallow, P.M., and Radford, A.N. (2015). Eavesdropping on heterospecific alarm calls: from mechanisms to consequences. Biol. Rev. Camb. Philos. Soc. 90, 560–586. 2. Goodale, E., Beauchamp, G., Magrath, R.D., Nieh, J.C., and Ruxton, G.D. (2010). Interspecific information transfer influences animal community structure. Trends Ecol. Evol. 25, 354–361. 3. Schmidt, K.A., Dall, S.R.X., and van Gils, J.A. (2010). The ecology of information: an overview on the ecological significance of making informed decisions. Oikos 119, 304–316. 4. Seppa¨nen, J.T., Forsman, J.T., Mo¨nkko¨nen, M., and Thomson, R.L. (2007). Social information use is a process across time, space, and ecology, reaching heterospecifics. Ecology 88, 1622–1633.
19. Shettleworth, S.J. (2010). Cognition, Evolution, and Behavior, Second Edition. (Oxford University Press). 20. Holle´n, L.I., and Radford, A.N. (2009). The development of alarm call behaviour in mammals and birds. Anim. Behav. 78, 791–800. 21. Bell, M.B.V., Radford, A.N., Rose, R., Wade, H.M., and Ridley, A.R. (2009). The value of constant surveillance in a risky environment. Proc. Biol. Sci. 276, 2997–3005. 22. Davies, N.B., Madden, J.R., Butchart, S.H.M., and Rutila, J. (2006). A hostrace of the cuckoo Cuculus canorus with nestlings attuned to the parental alarm calls of the host species. Proc. Biol. Sci. 273, 693–699. 23. Rabin, L.A., Coss, R.G., and Owings, D.H. (2006). The effects of wind turbines on antipredator behavior in California ground squirrels (Spermophilus beecheyi). Biol. Conserv. 131, 410–420.
5. Curio, E., Ernst, U., and Vieth, W. (1978). Cultural transmission of enemy recognition: one function of mobbing. Science 202, 899–901.
24. Shriner, W.M. (1999). Antipredator responses to a previously neutral sound by free-living adult golden-mantled ground squirrels, Spermophilus lateralis (Sciuridae). Ethology 105, 747–757.
6. Griffin, A.S., Blumstein, D.T., and Evans, C.S. (2000). Training captivebred or translocated animals to avoid predators. Conserv. Biol. 14, 1317–1326.
25. Bach, D.R., Neuhoff, J.G., Perrig, W., and Seifritz, E. (2009). Looming sounds as warning signals: the function of motion cues. Int. J. Psychophysiol. 74, 28–33.
7. Fallow, P.M., Gardner, J.L., and Magrath, R.D. (2011). Sound familiar? Acoustic similarity provokes responses to unfamiliar heterospecific alarm calls. Behav. Ecol. 22, 401–410.
26. Carlile, P.A., Peters, R.A., and Evans, C.S. (2006). Detection of a looming stimulus by the Jacky dragon: selective sensitivity to characteristics of an aerial predator. Anim. Behav. 72, 553–562.
8. Fallow, P.M., Pitcher, B.J., and Magrath, R.D. (2013). Alarming features: birds use specific acoustic properties to identify heterospecific alarm calls. Proc. Biol. Sci. 280, 20122539.
27. Hall, D.A., and Moore, D.R. (2003). Auditory neuroscience: the salience of looming sounds. Curr. Biol. 13, R91–R93.
9. Johnson, F.R., McNaughton, E.J., Shelly, C.D., and Blumstein, D.T. (2003). Mechanisms of heterospecific recognition in avian mobbing calls. Aust. J. Zool. 51, 577–585.
28. Hoppitt, W., and Laland, K.N. (2013). Social Learning: An Introduction to Mechanisms, Methods, and Models. (Princeton University Press). 29. Rendall, D., Owren, M.J., and Ryan, M.J. (2009). What do animal signals mean? Anim. Behav. 78, 233–240.
10. Randler, C. (2012). A possible phylogenetically conserved urgency response of great tits (Parus major) towards allopatric mobbing calls. Behav. Ecol. Sociobiol. 66, 675–681.
30. Dabelsteen, T. (2005). Public, private or anonymous? Facilitating and countering eavesdropping. In Animal Communication Networks, P.K. McGregor, ed. (Cambridge University Press), pp. 38–62.
11. Russ, J.M., Jones, G., Mackie, I.J., and Racey, P.A. (2004). Interspecific responses to distress calls in bats (Chiroptera: Vespertilionidae): a function for convergence in call design? Anim. Behav. 67, 1005–1014.
31. Westrip, J.R.S., and Bell, M.B.V. (2015). Breaking down species boundaries: selective pressures behind interspecific communication in vertebrates. Ethology 121, 1–8.
12. Teichroeb, L.J., Riede, T., Kotrba, R., and Lingle, S. (2013). Fundamental frequency is key to response of female deer to juvenile distress calls. Behav. Processes 92, 15–23.
32. Holt, R.D. (2007). The unravelling of nature’s information webs: the next depressing frontier in conservation? Isr. J. Ecol. Evol. 53, 229–236.
13. Magrath, R.D., and Bennett, T.H. (2012). A micro-geography of fear: learning to eavesdrop on alarm calls of neighbouring heterospecifics. Proc. Biol. Sci. 279, 902–909. 14. Magrath, R.D., Pitcher, B.J., and Gardner, J.L. (2009). Recognition of other species’ aerial alarm calls: speaking the same language or learning another? Proc. Biol. Sci. 276, 769–774. 15. Ramakrishnan, U., and Coss, R.G. (2000). Recognition of heterospecific alarm vocalizations by bonnet macaques (Macaca radiata). J. Comp. Psychol. 114, 3–12. 16. Wheatcroft, D., and Price, T.D. (2013). Learning and signal copying facilitate communication among bird species. Proc. Biol. Sci. 280, 20123070. 17. Zuberbu¨hler, K. (2000). Causal knowledge of predators’ behaviour in wild Diana monkeys. Anim. Behav. 59, 209–220. 18. Haff, T.M., and Magrath, R.D. (2013). Eavesdropping on the neighbours: fledgling response to heterospecific alarm calls. Anim. Behav. 85, 411–418.
33. Crane, A.L., and Ferrari, M.C.O. (2013). Social learning of predation risk: a review and prospectus. In Social Learning Theory: Phylogenetic Considerations across Animal, Plant, and Microbial Taxa, K.B. Clark, ed. (Nova), pp. 53–82. 34. Griffin, A.S. (2004). Social learning about predators: a review and prospectus. Learn. Behav. 32, 131–140. 35. Garcia, R.A., Cabeza, M., Rahbek, C., and Arau´jo, M.B. (2014). Multiple dimensions of climate change and their implications for biodiversity. Science 344, 1247579. 36. Guisan, A., Petitpierre, B., Broennimann, O., Daehler, C., and Kueffer, C. (2014). Unifying niche shift studies: insights from biological invasions. Trends Ecol. Evol. 29, 260–269. 37. Frair, J.L., Merrill, E.H., Allen, J.R., and Boyce, M.S. (2007). Know thy enemy: experience affects elk translocation success in risky landscapes. J. Wildl. Manage. 71, 541–554. 38. Seddon, P.J., Armstrong, D.P., and Maloney, R.F. (2007). Developing the science of reintroduction biology. Conserv. Biol. 21, 303–312.
4 Current Biology 25, 1–4, August 3, 2015 ª2015 Elsevier Ltd All rights reserved
Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls Graphical Abstract
Authors Robert D. Magrath, Tonya M. Haff, Jessica R. McLachlan, Branislav Igic
Correspondence [email protected]
In Brief Many vertebrates get information about danger by eavesdropping on other species’ alarm calls. Magrath et al. show experimentally that a wild bird species can learn to recognize previously unfamiliar sounds as alarm calls, supporting indirect evidence that learning is critical in establishing anti-predator ‘‘information webs’’ in natural communities.
Highlights d
Vertebrates benefit by eavesdropping on other species’ alarm calls
d
We used model predators and playback to test if wild birds learn about novel alarms
d
Individual birds learned to recognize novel sounds as alarm calls
d
Such rapid learning explains eavesdropping on acoustically diverse alarm calls
Magrath et al., 2015, Current Biology 25, 1–4 August 3, 2015 ª2015 Elsevier Ltd All rights reserved http://dx.doi.org/10.1016/j.cub.2015.06.028
Please cite this article in press as: Magrath et al., Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls, Current Biology (2015), http:// dx.doi.org/10.1016/j.cub.2015.06.028
Current Biology
Report Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls Robert D. Magrath,1,* Tonya M. Haff,1 Jessica R. McLachlan,1 and Branislav Igic1,2 1Division
of Evolution, Ecology and Genetics, Research School of Biology, Australian National University, Canberra 2601, Australia address: Department of Biology, University of Akron, Akron, OH 44325-3908, USA *Correspondence: [email protected] http://dx.doi.org/10.1016/j.cub.2015.06.028 2Present
SUMMARY
Many vertebrates gain critical information about danger by eavesdropping on other species’ alarm calls [1], providing an excellent context in which to study information flow among species in animal communities [2–4]. A fundamental but unresolved question is how individuals recognize other species’ alarm calls. Although individuals respond to heterospecific calls that are acoustically similar to their own, alarms vary greatly among species, and eavesdropping probably also requires learning [1]. Surprisingly, however, we lack studies demonstrating such learning. Here, we show experimentally that individual wild superb fairy-wrens, Malurus cyaneus, can learn to recognize previously unfamiliar alarm calls. We trained individuals by broadcasting unfamiliar sounds while simultaneously presenting gliding predatory birds. Fairy-wrens in the experiment originally ignored these sounds, but most fled in response to the sounds after two days’ training. The learned response was not due to increased responsiveness in general or to sensitization following repeated exposure and was independent of sound structure. Learning can therefore help explain the taxonomic diversity of eavesdropping and the refining of behavior to suit the local community. In combination with previous work on unfamiliar predator recognition (e.g., [5]), our results imply rapid spread of anti-predator behavior within wild populations and suggest methods for training captive-bred animals before release into the wild [6]. A remaining challenge is to assess the importance and consequences of direct association of unfamiliar sounds with predators, compared with social learning—such as associating unfamiliar sounds with conspecific alarms. RESULTS Playback experiments confirm that eavesdropping on heterospecifics occurs in terrestrial vertebrates worldwide, among both closely and distantly related species, and may be almost universal in birds and mammals [1]. Eavesdropping entails gain-
ing information from calls intended for others, can provide detailed information on danger, and helps structure animal communities by promoting mixed-species groups [1–4]. In some cases, individuals respond to unfamiliar heterospecific alarm calls similarly to conspecific calls because they share similar acoustic features (e.g., [7–12]). However, alarm calls are extremely variable among species, so community-wide eavesdropping probably requires learning (e.g., [13–17]). Despite its potential importance in explaining eavesdropping among species, the evidence for learned recognition of alarm calls is insufficient. First, playback experiments have merely probed natural variation in response to alarm calls (e.g., [13– 16, 18]), whereas demonstrating learning requires experimentally changing the behavior of individuals through manipulation of their experience [19]. Natural variation in response to alarm calls can be confounded by uncontrolled factors, such as physiological development instead of learning causing changes with age [20], or environmental or genetic differences rather than learning causing different responses in different locations [21– 23]. Second, the only experimental study to investigate learned recognition of alarm calls directly did not rule out alternative explanations [24]. Five golden-mantled ground squirrel (Spermophilus lateralis) individuals appeared to learn to recognize novel alarm calls, but the results were also consistent with increased wariness in general, sensitization following repeated exposure, and/or behavioral responses to the acoustic differences between training and control sounds [19, 24]. We tested experimentally whether the superb fairy-wren Malurus cyaneus, a small passerine bird, can learn to recognize unfamiliar sounds as alarm calls. Previous work has shown that fairy-wrens flee to cover in response to unfamiliar sounds that are acoustically very similar to conspecific aerial alarm calls— those given in response to predatory birds in flight—but usually do not flee in response to calls that are acoustically different [7, 8]. By contrast, fairy-wrens flee in response to alarm calls of locally common species regardless of acoustic differences [13, 14], which suggests that individuals learn to recognize the alarm calls of local species but does not rule out developmental, environmental, or genetic differences. We therefore used paired playback and predator model presentations in the wild to test directly whether individuals can learn to recognize previously unfamiliar sounds as alarm calls. We also tested alternative explanations, including response due to increased wariness, sensitization, or acoustic structure. We trained individual fairy-wrens to respond to one of two different unfamiliar sounds (Table 1; Figure 1A; Supplemental Experimental Procedures). First, we tested the response of
Current Biology 25, 1–4, August 3, 2015 ª2015 Elsevier Ltd All rights reserved 1
Please cite this article in press as: Magrath et al., Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls, Current Biology (2015), http:// dx.doi.org/10.1016/j.cub.2015.06.028
Table 1. Experimental Design Stage of Experiment Treatment
Pre-training
Training (eight playbacks over two days)
Post-training 1 (same day)
Post-training 2 (next day)
Predator treatment
sound 1 and sound 2 playbacks
looming playback of sound 1 or sound 2 with predator model
sound 1 and sound 2 playbacks
sound 1 and sound 2 playbacks
Looming control
sound 1 and sound 2 playbacks
looming playback of sound 1 or sound 2 without predator model
sound 1 and sound 2 playbacks
Non-looming control
sound 1 and sound 2 playbacks
non-looming playback of sound 1 or sound 2 without predator model
sound 1 and sound 2 playbacks
Birds were first tested for response to two novel sounds (allopatric thornbill alarm call or synthetic sound) in pre-training playbacks, to ensure that birds included in the experiment did not flee before training, then exposed to one of the two sounds over a two-day period, and finally tested in post-training playbacks (n = 10 for each treatment). Five birds in each treatment were exposed to each of the two sounds, with the other acting as a within-treatment, unexposed control. The order of playback within pre- and post-training tests was alternated. The looming playback consisted of a sequence of notes with increasing tempo and amplitude, mimicking a natural pattern of alarm calls to an approaching threat. The non-looming playback consisted of the same notes, but in a slower tempo and random order. The predator treatment consisted of a looming playback sequence with simultaneous presentation of a gliding predator model. The controls had looming or non-looming playback sequences without a predator presentation. Figure 1 shows the novel sounds and predator models, Figure S1 shows the temporal sequences, and further details are in the text and Supplemental Experimental Procedures.
individuals to each unfamiliar sound (pre-training), to ensure that the individuals included in the experiment did not flee in response to either sound before training. Next, we exposed each individual to eight playbacks of either one or the other of those two sounds over two days (training) and then finally tested the response to each sound after training (post-training). The training period included three treatments, each carried out on ten different birds in separate social groups: the predator treatment, in which unfamiliar sounds were accompanied by gliding predator models (Figure 1B), and two control treatments in which the sounds were presented without predators. The two controls used the same unfamiliar sounds as the predator treatment, but in one control they were presented in a different temporal pattern. The predator treatment and one of the controls used a ‘‘looming’’ playback consisting of a sequence of notes with increasing tempo and amplitude, mimicking a natural pattern of alarm calls to an approaching threat. The non-looming control consisted of the same notes, but in a slower tempo and random order, which is more natural for non-alarm vocalizations (Table 1; Figure S1; Supplemental Experimental Procedures). The two controls were included in case a looming pattern facilitated learning or was itself a cue used in learning to recognize unfamiliar alarm calls; in some species, looming stimuli can prompt attention or be aversive [25–27]. Half of the birds in each treatment were trained to one sound and half to the other, to ensure that sound type did not confound interpretation. The critical prediction for learned recognition of specific sounds as alarm calls in the predator treatment was flight to cover in response to the exposed sound but not in response to the unexposed sound when tested post-training. Response to Unfamiliar Sounds Most individuals learned to treat unfamiliar sounds as alarm calls after they were presented with a gliding predator. After training, 9 of 10 fairy-wrens fled to cover after playback of the exposed (training) sound, and of these, only one also fled in response to the unexposed sound (Figure 1C; same-day playbacks, McNemar’s test, exact binomial one-tailed p = 0.004). There
was a similar pattern when tested again at least one day later, with 7 of 10 fleeing in response to the exposed sound and only two also fleeing in response to the unexposed sound (McNemar’s test, exact binomial one-tailed p = 0.03). Combining data over the same- and next-day playbacks showed that 8 of 10 birds fled more often in response to the exposed than in response to the unexposed sound, with none showing the opposite pattern (McNemar’s test, exact binomial one-tailed p = 0.004; proportion fleeing did not differ significantly between days: binomial test on paired data, without a directional hypothesis, two-tailed p = 0.25). An equal number of birds learned to respond to each of the two different sounds (4 of 5 learned each). The response of birds to the exposed sounds in the predator treatment was not due to sensitization. None of the 10 birds in either control treatment fled in response to the exposed or unexposed sound when tested post-training (predator treatment compared to each control in same-day post-training playbacks: McNemar’s test, exact binomial one-tailed p = 0.004). DISCUSSION Learning to Recognize Alarm Calls Our results demonstrate learned recognition, supporting indirect evidence that learning provides a mechanism to explain widespread eavesdropping in vertebrates. After only two days’ training, fairy-wrens responded to previously unfamiliar sounds as if they were heterospecific aerial alarm calls warning of flying predators. The fairy-wrens learned to flee in response to either sound, depending on which was associated with the appearance of the predator model during training, showing that acoustic structure did not explain the response to trained versus untrained sounds. They did not respond after training to playback of the unexposed sound, showing that they discriminated between the sounds and did not flee in response to the exposed (training) sound simply because they had become wary of playbacks in general. The learned recognition of the predator-trained sound was not due to sensitization. No bird responded to sounds after
2 Current Biology 25, 1–4, August 3, 2015 ª2015 Elsevier Ltd All rights reserved
Please cite this article in press as: Magrath et al., Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls, Current Biology (2015), http:// dx.doi.org/10.1016/j.cub.2015.06.028
Figure 1. Experimental Sounds, Predator Models, and Results (A) Spectrograms of single notes that were used to compose unfamiliar sounds and training sequences. TB, allopatric, chestnut-rumped thornbill aerial alarm; SYN, note synthesized on computer; FW, superb fairy-wren aerial alarm (for comparison). (B) Ventral views of predator models, based on pied currawongs (CW) and collared sparrowhawks (SH). Superb fairy-wrens (FW) are shown at the same scale; photographs by Simon Bennett. (C) Post-training response of fairy-wrens to unfamiliar sounds that had been paired with the appearance of predator models (exposed), compared with their response to unexposed sounds of equal amplitude. Dark bars, same-day playbacks; hatched bars, next-day playbacks.
repeated playback alone in the control treatments, even if the sounds were played in a looming pattern of increasing amplitude and rate (Figure S1). The looming sequence was the pattern we associated with the predator model to simulate a natural pattern of calling when a predator approaches. Learning therefore entailed associating the unfamiliar sound with danger, potentially through direct association of the sound with the predator (Pavlovian associative learning [19]). It is also possible that individuals associated the unfamiliar sounds with anti-predator behavior of other individuals, such as fleeing or alarm calling, and so there could also have been social learning [19, 28]. Distinguishing between direct and social learning would be valuable because the mechanism could affect the rate of learning; currently we do not know whether it is essential to detect the predator directly, or whether individuals can learn to associate conspecific alarm calls or behavior with unfamiliar sounds. Regardless of the specific mechanism, individuals learned to treat appropriate sounds as alarm calls.
Ecological and Evolutionary Consequences Our study addresses the ecological question of learning to eavesdrop within natural communities. This requires testing for learned recognition of sounds that have properties of alarm calls but are unfamiliar, as if a new alarm-calling species had joined the community. A related issue is whether alarm signal design affects learning. Abrupt, broad-band, and non-linear sounds—such as some alarm calls, but not the sounds in this study—can capture attention and in doing so may facilitate learning about danger [29]. Furthermore, looming sounds can prompt attention or be aversive [25, 27], as can similarities to conspecific or familiar heterospecific alarm calls [8], and so either might facilitate learning. Although our playbacks were presented in a looming sequence and had general characteristics of aerial alarm calls, these features did not determine which sound became recognized as an alarm. Overall, the evolution and use of alarm calls could depend in part on what acoustic features affect both learned and unlearned responses, as well as the benefits and costs of being eavesdropped on [1, 16, 30, 31]. In addition to potentially affecting the evolution of signal design, the ability to learn to recognize heterospecific signals is important for understanding the establishment of community ‘‘information webs’’ [32] and could have conservation significance. Previous experiments show that alarm calls can promote learning about predators themselves [33, 34], whereas we show here that predators can prompt learning about heterospecific alarm calls. Together these could lead to the rapid spread of anti-predator behavior, with known predators facilitating learning about unfamiliar alarm calls, and learned recognition of calls aiding subsequent recognition of unfamiliar predators. Rapid spread of anti-predator behavior is likely to help populations cope with changing community composition, as is occurring with climate change and invasive species [35, 36]. Our methods also suggest that in conservation programs it would be possible to train captive-bred individuals to recognize heterospecific signals of danger, and not just predators themselves [6, 37, 38]. Such work should also consider the importance of social learning and examine the retention of learned anti-predator behavior. SUPPLEMENTAL INFORMATION Supplemental Information includes one figure and Supplemental Experimental Procedures and can be found with this article online at http://dx.doi.org/10. 1016/j.cub.2015.06.028. ACKNOWLEDGMENTS We thank B. Pitcher and A. Smith for models; T.H. Bennett for work on a pilot experiment; and K. Cain, A. Cockburn, N. Langmore, P. Re˛k, and three reviewers for detailed and insightful comments. This work was funded by the Australian National University and was conducted under permits from Environment ACT, the Australian Bird and Bat Banding Scheme, Australian National Botanic Gardens, and the ANU Ethics Committee. Received: March 12, 2015 Revised: May 7, 2015 Accepted: June 11, 2015 Published: July 16, 2015
Current Biology 25, 1–4, August 3, 2015 ª2015 Elsevier Ltd All rights reserved 3
Please cite this article in press as: Magrath et al., Wild Birds Learn to Eavesdrop on Heterospecific Alarm Calls, Current Biology (2015), http:// dx.doi.org/10.1016/j.cub.2015.06.028
REFERENCES 1. Magrath, R.D., Haff, T.M., Fallow, P.M., and Radford, A.N. (2015). Eavesdropping on heterospecific alarm calls: from mechanisms to consequences. Biol. Rev. Camb. Philos. Soc. 90, 560–586. 2. Goodale, E., Beauchamp, G., Magrath, R.D., Nieh, J.C., and Ruxton, G.D. (2010). Interspecific information transfer influences animal community structure. Trends Ecol. Evol. 25, 354–361. 3. Schmidt, K.A., Dall, S.R.X., and van Gils, J.A. (2010). The ecology of information: an overview on the ecological significance of making informed decisions. Oikos 119, 304–316. 4. Seppa¨nen, J.T., Forsman, J.T., Mo¨nkko¨nen, M., and Thomson, R.L. (2007). Social information use is a process across time, space, and ecology, reaching heterospecifics. Ecology 88, 1622–1633.
19. Shettleworth, S.J. (2010). Cognition, Evolution, and Behavior, Second Edition. (Oxford University Press). 20. Holle´n, L.I., and Radford, A.N. (2009). The development of alarm call behaviour in mammals and birds. Anim. Behav. 78, 791–800. 21. Bell, M.B.V., Radford, A.N., Rose, R., Wade, H.M., and Ridley, A.R. (2009). The value of constant surveillance in a risky environment. Proc. Biol. Sci. 276, 2997–3005. 22. Davies, N.B., Madden, J.R., Butchart, S.H.M., and Rutila, J. (2006). A hostrace of the cuckoo Cuculus canorus with nestlings attuned to the parental alarm calls of the host species. Proc. Biol. Sci. 273, 693–699. 23. Rabin, L.A., Coss, R.G., and Owings, D.H. (2006). The effects of wind turbines on antipredator behavior in California ground squirrels (Spermophilus beecheyi). Biol. Conserv. 131, 410–420.
5. Curio, E., Ernst, U., and Vieth, W. (1978). Cultural transmission of enemy recognition: one function of mobbing. Science 202, 899–901.
24. Shriner, W.M. (1999). Antipredator responses to a previously neutral sound by free-living adult golden-mantled ground squirrels, Spermophilus lateralis (Sciuridae). Ethology 105, 747–757.
6. Griffin, A.S., Blumstein, D.T., and Evans, C.S. (2000). Training captivebred or translocated animals to avoid predators. Conserv. Biol. 14, 1317–1326.
25. Bach, D.R., Neuhoff, J.G., Perrig, W., and Seifritz, E. (2009). Looming sounds as warning signals: the function of motion cues. Int. J. Psychophysiol. 74, 28–33.
7. Fallow, P.M., Gardner, J.L., and Magrath, R.D. (2011). Sound familiar? Acoustic similarity provokes responses to unfamiliar heterospecific alarm calls. Behav. Ecol. 22, 401–410.
26. Carlile, P.A., Peters, R.A., and Evans, C.S. (2006). Detection of a looming stimulus by the Jacky dragon: selective sensitivity to characteristics of an aerial predator. Anim. Behav. 72, 553–562.
8. Fallow, P.M., Pitcher, B.J., and Magrath, R.D. (2013). Alarming features: birds use specific acoustic properties to identify heterospecific alarm calls. Proc. Biol. Sci. 280, 20122539.
27. Hall, D.A., and Moore, D.R. (2003). Auditory neuroscience: the salience of looming sounds. Curr. Biol. 13, R91–R93.
9. Johnson, F.R., McNaughton, E.J., Shelly, C.D., and Blumstein, D.T. (2003). Mechanisms of heterospecific recognition in avian mobbing calls. Aust. J. Zool. 51, 577–585.
28. Hoppitt, W., and Laland, K.N. (2013). Social Learning: An Introduction to Mechanisms, Methods, and Models. (Princeton University Press). 29. Rendall, D., Owren, M.J., and Ryan, M.J. (2009). What do animal signals mean? Anim. Behav. 78, 233–240.
10. Randler, C. (2012). A possible phylogenetically conserved urgency response of great tits (Parus major) towards allopatric mobbing calls. Behav. Ecol. Sociobiol. 66, 675–681.
30. Dabelsteen, T. (2005). Public, private or anonymous? Facilitating and countering eavesdropping. In Animal Communication Networks, P.K. McGregor, ed. (Cambridge University Press), pp. 38–62.
11. Russ, J.M., Jones, G., Mackie, I.J., and Racey, P.A. (2004). Interspecific responses to distress calls in bats (Chiroptera: Vespertilionidae): a function for convergence in call design? Anim. Behav. 67, 1005–1014.
31. Westrip, J.R.S., and Bell, M.B.V. (2015). Breaking down species boundaries: selective pressures behind interspecific communication in vertebrates. Ethology 121, 1–8.
12. Teichroeb, L.J., Riede, T., Kotrba, R., and Lingle, S. (2013). Fundamental frequency is key to response of female deer to juvenile distress calls. Behav. Processes 92, 15–23.
32. Holt, R.D. (2007). The unravelling of nature’s information webs: the next depressing frontier in conservation? Isr. J. Ecol. Evol. 53, 229–236.
13. Magrath, R.D., and Bennett, T.H. (2012). A micro-geography of fear: learning to eavesdrop on alarm calls of neighbouring heterospecifics. Proc. Biol. Sci. 279, 902–909. 14. Magrath, R.D., Pitcher, B.J., and Gardner, J.L. (2009). Recognition of other species’ aerial alarm calls: speaking the same language or learning another? Proc. Biol. Sci. 276, 769–774. 15. Ramakrishnan, U., and Coss, R.G. (2000). Recognition of heterospecific alarm vocalizations by bonnet macaques (Macaca radiata). J. Comp. Psychol. 114, 3–12. 16. Wheatcroft, D., and Price, T.D. (2013). Learning and signal copying facilitate communication among bird species. Proc. Biol. Sci. 280, 20123070. 17. Zuberbu¨hler, K. (2000). Causal knowledge of predators’ behaviour in wild Diana monkeys. Anim. Behav. 59, 209–220. 18. Haff, T.M., and Magrath, R.D. (2013). Eavesdropping on the neighbours: fledgling response to heterospecific alarm calls. Anim. Behav. 85, 411–418.
33. Crane, A.L., and Ferrari, M.C.O. (2013). Social learning of predation risk: a review and prospectus. In Social Learning Theory: Phylogenetic Considerations across Animal, Plant, and Microbial Taxa, K.B. Clark, ed. (Nova), pp. 53–82. 34. Griffin, A.S. (2004). Social learning about predators: a review and prospectus. Learn. Behav. 32, 131–140. 35. Garcia, R.A., Cabeza, M., Rahbek, C., and Arau´jo, M.B. (2014). Multiple dimensions of climate change and their implications for biodiversity. Science 344, 1247579. 36. Guisan, A., Petitpierre, B., Broennimann, O., Daehler, C., and Kueffer, C. (2014). Unifying niche shift studies: insights from biological invasions. Trends Ecol. Evol. 29, 260–269. 37. Frair, J.L., Merrill, E.H., Allen, J.R., and Boyce, M.S. (2007). Know thy enemy: experience affects elk translocation success in risky landscapes. J. Wildl. Manage. 71, 541–554. 38. Seddon, P.J., Armstrong, D.P., and Maloney, R.F. (2007). Developing the science of reintroduction biology. Conserv. Biol. 21, 303–312.
4 Current Biology 25, 1–4, August 3, 2015 ª2015 Elsevier Ltd All rights reserved
