Behavioural Processes 76 (2007) 149–151 Commentary Cooperation and communication networks Rita Covas a,* , Peter K. McGregor b , Claire Doutrelant a a CEFE-CNRS, 1919 Rte de Mende, 34293 Montpellier Cedex 5, France b Duchy College, Stoke Climsland, Callington PL17 8PB, Cornwall, UK Received 15 December 2006; accepted 20 December 2006 Indirect fitness benefits have long been accepted as one of the main explanations for the evolution of cooperative breeding. However, recent studies have shown that helpers can be unre- lated to the breeders and may have access to reproduction or to other direct benefits. Understanding the basis for cooperation in the absence of relatedness is therefore a renewed challenge for research on cooperative breeding. In this context, the review by Bergm¨ uller and colleagues represents an important contri- bution to the field by bringing together for the first time the theories of classic cooperation and cooperative breeding and thereby providing new avenues for research. This paper is likely to be particularly relevant to theoretical studies of cooperation in cooperatively breeding systems given the emphasis on clarifying concepts and the attempt to define which types of ‘traditional’ cooperation fall into the different theories that attempt to account for direct benefits for cooperative breeders. Although the parti- tioning of different potential direct benefits (pay to stay, prestige and group augmentation) into the proposed categories of coop- eration may be debatable to some extent, it does constitute an important first step. A central issue when dealing with cooperation in cooper- atively breeding systems, as acknowledged by the authors, is how to deal with the fact that cooperative breeding typically involves multiple players. As pointed out, situations where sev- eral individuals (as opposed to only two) cooperate to perform a certain task are problematic to address from a theoretical point of view because individuals invest into common goods, which should cause cooperation to collapse. Public goods allow societies composed largely of co-operators to outperform soci- eties composed mainly of non-cooperators. However, public goods also provide an incentive for individuals to be selfish by benefiting from the public good. The authors refer to cur- rent solutions to this problem. In particular, they mention a study on humans, where being altruist could increase one’s * Corresponding author. E-mail address: rita.covas@cefe.cnrs.fr (R. Covas). ‘reputation’ (Milinski et al., 2002) and another study on pop- ulation dynamics in structured populations where reproduction in groups, combined with dispersal between groups, results in variations in group size and high investors having greater fitness than low investors (Killingback et al., 2006). We believe there is another possible solution to this ‘tragedy of the commons’ that deserves attention: considering interactions in the framework of animal communication networks (McGregor, 2005). Two key features of communication networks are that several individuals are involved and information can be gathered from interactions by individuals (“eavesdroppers”) that do not take part in the interactions (McGregor and Dabelsteen, 1996; Naguib and Todt, 1997; Oliveira et al., 1998; Otter et al., 1999; Doutrelant and McGregor, 2000; Doutrelant et al., 2001; Peake et al., 2001). Although not part of the classic cooperation theory, Bshary and d’Souza (2005) pointed out that recent developments in coop- eration theory can be linked to the concept of communication networks and the network context has been pivotal in their under- standing of a classic example of cooperation: cleaner fish–client interactions. In communication networks, it has been argued that some behaviour not primarily intended as signals can evolve a sig- nalling function if they contain reliable information used by an audience of eavesdroppers (see also Lotem et al., 1999; Borgia and Colerman, 2000; Johnstone and Bshary, 2004; Bshary and Grutter, 2006). Individuals are expected to eavesdrop when such behaviour increases their fitness; similarly, if individuals gain from being observed they should perform the behaviour relevant to the eavesdroppers more frequently in presence of an audience of eavesdropper(s) (Doutrelant et al., 2001; Bshary and Grutter, 2006; Doutrelant and Covas, 2007). It has been suggested that helping behaviour in cooperatively breeding species might have evolved a secondary signalling function (Zahavi, 1995; Putland, 2001; Doutrelant and Covas, 2007). Helping can be associated with direct fitness benefits for the helpers such as increased likelihood of reproducing (Reyer, 1986; Whittingham et al., 1997; Richardson et al., 2002; Webster et al., 2004) or being accepted in a coalition (Heinsohn et al., 0376-6357/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.beproc.2006.12.019