Where does a flock end from an information perspective? A comparative experiment with live and robotic birds

Behavioral Ecology doi:10.1093/beheco/arr132 Advance Access publication 30 August 2011 Original Article Where does a flock end from an information perspective? A comparative experiment with live and robotic birds Esteban Fernández-Juricica and Victor Kowalskib Department of Biological Sciences, Purdue University, 915 W. State Street, West Lafayette, IN 47907, USA and bDepartment of Biological Sciences, California State University Long Beach, 1250 Bellflower Blvd, Long Beach, CA 90840, USA a INTRODUCTION ne of the benefits of joining a group is that predator detection can be improved through the combined vigilance of group members or collective detection (e.g., Quinn and Cresswell 2005; Fairbanks and Dobson 2007; Ward and Mehner 2010). Theory assumes that when an individual detects a predator, information (e.g., alarm calls, flushing responses, chemical signals) will pass to the other members of the group immediately and accurately to ensure a timely escape (Pulliam 1973; Lazarus 1979; Pulliam et al. 1982; Hart and Lendrem 1984; Lima 1987). Recent theoretical work has relaxed the assumption of immediate information transfer to represent a decrease in information flow over distance (Proctor et al. 2003). This is because, all else being equal, distance may degrade 1) visual information through a reduction in visual contrast and the ability to resolve changes in the behavior of group mates (Fernández-Juricic et al. 2004), 2) acoustic information through an increase in attenuation and reverberation (Balsby et al. 2003), and 3) chemical information through a reduction in the concentration of chemicals cues from the source (Ferrari et al. 2010). Therefore, longer neighbor distances are expected to reduce the availability of social information and increase the perception of risk because O Address correspondence to E. Fernández-Juricic. E-mail: efernan @purdue.edu. Received 4 March 2011; revised 12 June 2011; accepted 16 June 2011.
The Author 2011. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: the protection provided by the presence and number of group mates decreases (Proctor et al. 2003). Empirical evidence suggests that greater spacing between group mates increases the perception of risk and delays the speed of response to a threat (Pöysä 1994; Lima and Zollner 1996; Hilton et al. 1999; Rolando et al. 2001; Fernández-Juricic and Kacelnik 2004; Quinn and Cresswell 2005). However, little is known about the shape of the function by which information flow decreases with neighbor distance and the extent to which the shape of this function varies between species. This is relevant from both theoretical and empirical perspectives. Theoretically, models assume either linear or nonlinear functions whose shapes can affect the speed of information transfer and as a result the area a flock occupies and the individual investment in vigilance and foraging (Proctor et al. 2003; Jackson and Ruxton 2006). Additionally, models assume that information flows without being degraded among close group mates irrespective of the number of individuals in the group (Bahr and Bekoff 1999; Jackson and Ruxton 2006; Proctor et al. 2006). However, if information quality is reduced as it transfers between close group mates, it could lead to false alarms (e.g., individuals misinterpreting a sudden flight caused by a nonthreatening stimulus as if it were a predator attack; Proctor et al. 2001). An increase in false alarms can reduce the benefits of collective detection and actually become a cost for individuals joining groups by increasing energetic expenditure in unnecessary flights and reducing the amount of time foraging (Beauchamp and Ruxton 2007). Empirically, individuals are expected to distance themselves Predator detection is improved when individuals join groups. Theory assumes that the transfer of social information about predators among individuals is immediate and accurate. However, animals in groups space themselves at different distances. Little is known about the shape of the social information transfer function over distance, which can affect group cohesion and ultimately the costs and benefits of group living. Our goal was to study the flow of social information in 3 bird species with different visual acuity (European starling, brown-headed cowbird, and house finch). We used robotic birds to manipulate the availability of social information. In a previous study, we demonstrated that birds react to robotic birds in the same way as they do to live birds. We measured the probability of 3 linearly placed live birds reacting to the flushing behavior of 2 robotic birds of the same species. Our study species were tested independently. We found a nonlinear decrease in social information flow with increasing distance between the robots and live birds; however, this decrease was more pronounced in species with lower visual acuity. Additionally, social information apparently degraded when flowing between closely spaced individuals, which could lead to false alarms. Our findings suggest that the benefits of social information flow are restricted to small neighbor distances and that larger species, with higher visual acuity, may have a greater spatial domain of collective detection. This mechanism may explain the spatial limits of flocks based on the transfer of social information. Key words: antipredator behavior, collective detection, predator detection, social information, vigilance, visual acuity. [Behav Ecol 22:1304–1311 (2011)] Fernández-Juricic and Kowalski • Where does a flock end from an information perspective? METHODS We conducted the experiment outdoors to enhance the perception of risk under semicontrolled conditions. Our experimental arena consisted of 2 robotic birds (1 male and 1 female) closely spaced but facing opposite directions. Perpendicular to the robots, we laid out an enclosure with transparent partitions such that there was 1 live bird in each of the 3 compartments (close, middle, and far in relation to the robots; Figure 1). The robots and live birds belonged to the same species. Details of the experimental arena are presented in Supplementary Appendix A. Figure 1 Probabilities of detecting robots flushing at different distances between the live birds and robots of 3 different species (European starling, brown-headed cowbird, and house finch). Probability functions estimated with generalized linear models. We controlled for several factors that can affect the flow of social information. First, we only manipulated the distance between producers and receivers but kept the receivers at the same distance among themselves. Second, we kept flock size constant and relatively small by having 2 information producers (robots) and 3 information receivers (live birds). Collective detection is expected to be more prevalent in small rather than in large flocks (e (...truncated)