Final Report Summary - SIGNAL RELIABILITY (Signal reliability and decision making in social birds)
In many social animal species, distinct vocalisations are often used for communication between group members. These distinct vocalisations have been shown to encode detailed acoustic information about the signaller, such as identity, age, sex, and dominance status. In making a decision about whether to respond to a call, receivers therefore have the opportunity to take such information into account. Recently, increasing research interest is focusing on the possibility that receivers also gain vocal information about a signaller's reliability. If receivers identify signallers and adjust their responses according to reliability, they could potentially minimise the costs inherent in responding to false information, such as lost foraging opportunites. Despite a large literature on avian vocal signals, no study has so far considered the influence of signaller reliability on the decisions made by other group members.
The overall main objective with this project was to investigate signal reliability in pied babblers (Turdoides bicolor), a highly vocal social bird species living in the South African desert. By using behavioural observations, sound recordings, acoustic analyses, model predator presentations and playback experiments, the aim was to focus on two major areas:
(1) the reliability of alarm calling (alarm calls are given to warn other group members of predators) and the resultant decisions made by receivers in response to others' alarm calls; and
(2) the decisions made by foraging group members in response to sentinels (one group member scanning for predators from a raised position while the rest of the group forages) that vary in their reliability.
Although acoustic analyses suggest that different group members exhibit individually distinct alarm calls, and there is hence a potential for receiver responses that vary depending on caller reliability (aim 1), preliminary analyses of natural observations made it apparent that accurate estimates of reliability are difficult to obtain from such data. Due to unforeseen logistical reasons and maternity leave, our plan to test this experimentally became unfeasible. As a consequence, we decided that the remaining time on the project should be focussed on sentinel behaviour and reliability (aim 2).
Pied babblers provide an ideal opportunity to address receiver decision making in response to sentinels. Babbler groups spend approximately 95 % of their foraging time on the ground, in a variety of habitats (ranging from open areas with little vegetation to sites covered with thick long grass), probing into the sand and sifting through vegetation for invertebrate prey. These methods of foraging often result in restricted vigilance, making individuals particularly vulnerable to predation from a variety of raptors, mammals and reptiles. Foraging groups frequently have a sentinel in place who often gives alarm calls, and continuously produces a low-amplitude watchman's song while on guard. Sentinels select a wide range of positions from which to guard, potentially affecting the magnitude of the benefits that arise to themselves and/or other group members; there is considerable variation in sentinel behaviour.
We first investigated a range of behavioural responses shown by group members to sentinel presence and, in particular, in response to the sentinel's use of the watchman's song. We used observational and experimental data to provide the first evidence that group members gain an increase in foraging success by responding to these vocal cues given by sentinels. Foraging pied babblers spread out more, used more exposed patches, looked up less often, and spent less time vigilant in response to sentinel calling: they feel safer with a sentinel present. Crucially, we demonstrated that these behavioural alterations lead to an increase in biomass intake by foragers, which is likely to enhance survival. We argue that this benefit may be the reason for sentinel calling, making it a truly cooperative behaviour. This work has been published in the internationally renowned journal Current Biology.
We then investigated whether foraging group members adjust their behaviour according to a sentinel's height; if a sentinel is perched higher, it may be more likely to detect predators sooner and hence be regarded as more reliable by other group members. As a consequence, we may expect foragers to feel safer in the sentinel's presence and adjust their behaviour accordingly. We found that after a sentinel moved to a higher position, foraging group members spent a smaller proportion of time vigilant, looked up less often, spread out more widely and were more likely to forage in the open compared to under vegetation cover. As a consequence of foragers feeling safer, they spent a greater proportion of time foraging and had a higher foraging efficiency following an increase in sentinel height. The increases in foraging time and efficiency after a sentinel moved higher resulted in an increase in the biomass intake rate of foragers. Foraging individuals also altered their behaviour in the opposite direction when a sentinel moved to a lower position and therefore, suffered a reduction in biomass intake rate. With playback experiments we showed that foragers appeared capable of assessing sentinel height from vocal cues alone. This work has been published in Proceedings of the Royal Society of London, Series B9.
We now have good evidence that the quiet vocalisations produced throughout a sentinel bout acts to advertise sentinel presence and height, amplifying the benefits of having a sentinel. However, it is unclear whether sentinel bouts are actively coordinated or the likelihood of sentinel replacement is adjusted to the level of risk. We investigated whether the fast-rate chuck call given at the end of a sentinel bout in the pied babblers functions to aid in coordinating sentinels by conspicuously advertising the absence of a sentinel and the need for replacement or whether they act as a signal of risk, advertising the urgency with which replacement is needed. Contrary to expectation, we found little evidence that these calls are involved in regulating the pied babbler sentinel system: observations revealed that their utterance is influenced only marginally by wind conditions and not at all by habitat, factors which can clearly affect current risk. For example, in dense habitats and/or high wind when predators are harder to spot10 there could be a crucial need for quick sentinel replacement and we would predict these calls to be given more often. Experimental playback also showed that the giving of these calls had no effect on the length of intervals between consecutive sentinel bouts. This work has been published in PLoS ONE.
It is well known that ecological conditions can affect behavioural decision making, especially choices relating to anti-predator behaviour. Predator detection, for example, can be compromised by the conditions in which an animal is foraging, leading to changes in vigilance behaviour. Given that sentinel behaviour involves anti-predator vigilance, we might expect sentinels to adjust their behaviour in response to ecological conditions, but this has not previously been considered. We examined the impact of habitat type and wind on latency between sentinel bouts, bout duration, and the initial position chosen by sentinels (height and distance to foragers) and found that sentinels started guarding sooner and guarded for longer in long grass compared to more open habitats, and also initiated sentinel bouts sooner in high wind, probably because of the increased predation risk in such circumstances. Sentinels also selected positions that were both lower and closer to the foraging group when it was windy, potentially improving transmission of vocal signals that are valuable to foragers. This work has been published in Animal Behaviour.
Despite the prevalence of sentinels in a wide variety of social bird and mammal species, surprisingly little is known about their effect on the foraging group. This project has provided ample evidence that sentinel decisions can have a major influence on the behaviour of foragers, decisions that can ultimately increase their chance of survival. By considering the influence of signaller reliability on receiver decision making, this project aimed to address key questions in biology concerning social communication, animal signalling and behavioural plasticity. Although we investigated reliability in terms of sentinel height, further work is required to investigate whether individual sentinels differ in their reliability (in terms of alarm calling) and what effect that has on receiver decisions. Such data would make a major contribution to our knowledge of how information sharing mediates group coordination and cooperation in socially living species, a prerequisite for stable, productive and well-functioning groups.
The overall main objective with this project was to investigate signal reliability in pied babblers (Turdoides bicolor), a highly vocal social bird species living in the South African desert. By using behavioural observations, sound recordings, acoustic analyses, model predator presentations and playback experiments, the aim was to focus on two major areas:
(1) the reliability of alarm calling (alarm calls are given to warn other group members of predators) and the resultant decisions made by receivers in response to others' alarm calls; and
(2) the decisions made by foraging group members in response to sentinels (one group member scanning for predators from a raised position while the rest of the group forages) that vary in their reliability.
Although acoustic analyses suggest that different group members exhibit individually distinct alarm calls, and there is hence a potential for receiver responses that vary depending on caller reliability (aim 1), preliminary analyses of natural observations made it apparent that accurate estimates of reliability are difficult to obtain from such data. Due to unforeseen logistical reasons and maternity leave, our plan to test this experimentally became unfeasible. As a consequence, we decided that the remaining time on the project should be focussed on sentinel behaviour and reliability (aim 2).
Pied babblers provide an ideal opportunity to address receiver decision making in response to sentinels. Babbler groups spend approximately 95 % of their foraging time on the ground, in a variety of habitats (ranging from open areas with little vegetation to sites covered with thick long grass), probing into the sand and sifting through vegetation for invertebrate prey. These methods of foraging often result in restricted vigilance, making individuals particularly vulnerable to predation from a variety of raptors, mammals and reptiles. Foraging groups frequently have a sentinel in place who often gives alarm calls, and continuously produces a low-amplitude watchman's song while on guard. Sentinels select a wide range of positions from which to guard, potentially affecting the magnitude of the benefits that arise to themselves and/or other group members; there is considerable variation in sentinel behaviour.
We first investigated a range of behavioural responses shown by group members to sentinel presence and, in particular, in response to the sentinel's use of the watchman's song. We used observational and experimental data to provide the first evidence that group members gain an increase in foraging success by responding to these vocal cues given by sentinels. Foraging pied babblers spread out more, used more exposed patches, looked up less often, and spent less time vigilant in response to sentinel calling: they feel safer with a sentinel present. Crucially, we demonstrated that these behavioural alterations lead to an increase in biomass intake by foragers, which is likely to enhance survival. We argue that this benefit may be the reason for sentinel calling, making it a truly cooperative behaviour. This work has been published in the internationally renowned journal Current Biology.
We then investigated whether foraging group members adjust their behaviour according to a sentinel's height; if a sentinel is perched higher, it may be more likely to detect predators sooner and hence be regarded as more reliable by other group members. As a consequence, we may expect foragers to feel safer in the sentinel's presence and adjust their behaviour accordingly. We found that after a sentinel moved to a higher position, foraging group members spent a smaller proportion of time vigilant, looked up less often, spread out more widely and were more likely to forage in the open compared to under vegetation cover. As a consequence of foragers feeling safer, they spent a greater proportion of time foraging and had a higher foraging efficiency following an increase in sentinel height. The increases in foraging time and efficiency after a sentinel moved higher resulted in an increase in the biomass intake rate of foragers. Foraging individuals also altered their behaviour in the opposite direction when a sentinel moved to a lower position and therefore, suffered a reduction in biomass intake rate. With playback experiments we showed that foragers appeared capable of assessing sentinel height from vocal cues alone. This work has been published in Proceedings of the Royal Society of London, Series B9.
We now have good evidence that the quiet vocalisations produced throughout a sentinel bout acts to advertise sentinel presence and height, amplifying the benefits of having a sentinel. However, it is unclear whether sentinel bouts are actively coordinated or the likelihood of sentinel replacement is adjusted to the level of risk. We investigated whether the fast-rate chuck call given at the end of a sentinel bout in the pied babblers functions to aid in coordinating sentinels by conspicuously advertising the absence of a sentinel and the need for replacement or whether they act as a signal of risk, advertising the urgency with which replacement is needed. Contrary to expectation, we found little evidence that these calls are involved in regulating the pied babbler sentinel system: observations revealed that their utterance is influenced only marginally by wind conditions and not at all by habitat, factors which can clearly affect current risk. For example, in dense habitats and/or high wind when predators are harder to spot10 there could be a crucial need for quick sentinel replacement and we would predict these calls to be given more often. Experimental playback also showed that the giving of these calls had no effect on the length of intervals between consecutive sentinel bouts. This work has been published in PLoS ONE.
It is well known that ecological conditions can affect behavioural decision making, especially choices relating to anti-predator behaviour. Predator detection, for example, can be compromised by the conditions in which an animal is foraging, leading to changes in vigilance behaviour. Given that sentinel behaviour involves anti-predator vigilance, we might expect sentinels to adjust their behaviour in response to ecological conditions, but this has not previously been considered. We examined the impact of habitat type and wind on latency between sentinel bouts, bout duration, and the initial position chosen by sentinels (height and distance to foragers) and found that sentinels started guarding sooner and guarded for longer in long grass compared to more open habitats, and also initiated sentinel bouts sooner in high wind, probably because of the increased predation risk in such circumstances. Sentinels also selected positions that were both lower and closer to the foraging group when it was windy, potentially improving transmission of vocal signals that are valuable to foragers. This work has been published in Animal Behaviour.
Despite the prevalence of sentinels in a wide variety of social bird and mammal species, surprisingly little is known about their effect on the foraging group. This project has provided ample evidence that sentinel decisions can have a major influence on the behaviour of foragers, decisions that can ultimately increase their chance of survival. By considering the influence of signaller reliability on receiver decision making, this project aimed to address key questions in biology concerning social communication, animal signalling and behavioural plasticity. Although we investigated reliability in terms of sentinel height, further work is required to investigate whether individual sentinels differ in their reliability (in terms of alarm calling) and what effect that has on receiver decisions. Such data would make a major contribution to our knowledge of how information sharing mediates group coordination and cooperation in socially living species, a prerequisite for stable, productive and well-functioning groups.