Aerial alarm calling by male fowl (Gallus gallus) reveals subtle new mechanisms of risk management Artemi Kokolakis, Carolynn L. Smith * , Christopher S. Evans Centre for the Integrative Study of Animal Behaviour, Department of Brain, Behaviour and Evolution, Macquarie University article info Article history: Received 11 August 2009 Initial acceptance 12 October 2009 Final acceptance 25 February 2010 Available online 22 April 2010 MS. number: A09-00532R Keywords: alarm call antipredator behaviour communication network fowl Gallus gallus judicious risk taking kin selection referential signal signal design Alarm calling is a classic problem in evolutionary biology. Although a signaller may increase the likeli- hood of survival for group members, which typically include kin and mates, there are inherent risks associated with any behaviour that increases conspicuousness to predators. Callers can increase their indirect benefits by calling only in the presence of an appropriate audience and manage concomitant costs by judicious investment. Possible tactics for controlling costs include facultative variation in call structure and timing, as well as sensitivity to the environmental and social factors that predict personal vulnerability. We examined individual variation in the alarm-calling behaviour of male fowl in natu- ralistic social groups. Previous studies of cost management have focused on variation at the level of alarm call rate. We took advantage of recent advances in wireless sound recording and remote video moni- toring to test for more subtle variation in signal structure and timing. These were then mapped onto individual mating success and moment-to-moment changes in environmental and social context. Results replicate the previous finding that alarm calling is sensitive to both social rank and recent mating success. In addition, we detected systematic variation in call structure as a function of personal vulnerability and proximity to a rival male. The frequency bandwidth of alarms was reliably influenced by degree of vigilance prior to calling, suggesting that this acoustic dimension reflects motivational state. Taken together, these results reveal several novel tactics for risk management, complementing those previously described at the level of gross variation in alarm-calling behaviour. Ó 2010 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. Male-signalling behaviour has been shaped by costs. These costs include mate investment, competition and predation. Investment in mates and offspring is often time consuming, may be energeti- cally expensive (Clutton-Brock & Godfray 1991) and may include the production of high-risk calls (Hauser 1997; Wilson et al. 2008). The same is true of competition with other males, which can also increase predation risk (Ryan et al. 1982) and attract eavesdroppers, such as parasites (Lehmann & Heller 1998; Bernal et al. 2007). However, if signalling behaviour increases fitness, this benefit may partially offset such costs. Male signal production thus reflects the interaction among several selective forces. Alarm calls that are directed towards conspecifics are a classic problem in evolutionary biology because this costly behaviour is apparently altruistic. It is now clear that alarm calling is, in some cases, an example of judicious investment (defined as the selective expression of a behaviour when it offers the most payoff). Signals that alert conspecifics to the presence of predators can indirectly benefit the signaller by increasing the survival of mates and/or kin (Sherman 1977; Witkin & Ficken 1979). However, calling may simultaneously increase the signaller’s likelihood of detection and capture by a predator (Endler 1993). Therefore selection should favour adaptations that allow the signaller to warn conspecifics, while minimizing risk. Alarm calls often have structural character- istics that function to reduce conspicuousness (Marler 1955; Wood et al. 2000). They may be given selectively in the presence of appropriate receivers (Evans & Marler 1992); when calling does occur, there can also be flexibility in the signal structure (i.e. facul- tative variation) that further manages risk (Bayly & Evans 2003). Alarm signals have been shaped by selection from both preda- tors and conspecifics (Klump et al. 1986; Bayly & Evans 2003). A signal structure that is difficult to hear and localize by predators is favoured to maximize the survival prospects of the signaller (Bradbury & Vehrencamp 1988). This involves changes to reduce both audibility (Klump et al. 1986) and localizability (Marler 1955; Wood et al. 2000). In the presence of a predator, animals may either incur the costs of alarm calling or adopt a cryptic strategy by remaining silent. Social factors are critically important for influencing this choice. These include an appropriate ‘audience’, such as potential mates or offspring (Cheney & Seyfarth 1985; Evans & Marler 1992). By calling only in the presence of appropriate receivers, the signaller thus * Correspondence: C. L. Smith, Department of Brain, Behaviour and Evolution, Macquarie University, Sydney, NSW 2109, Australia. E-mail address: klynn.smith@gmail.com (C.L. Smith). Contents lists available at ScienceDirect Animal Behaviour journal homepage: www.elsevier.com/locate/anbehav 0003-3472/$38.00 Ó 2010 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.anbehav.2010.03.013 Animal Behaviour 79 (2010) 1373e1380