Phylogénie Et Evolution Du Comportement Social Chez Les Blattes ...

Ev o l u t i o n d u c o m p o r t e m e n t s o c i a lIntroductionStudies on behaviour and phylogeny have a long common history. From the early times,comparative studies have shown that behaviour can be remarkably informative regarding therelationships of taxa, as reviewed by Hinde & Tinbergen (1958). These comparative studiesfocused on highly stereotyped and ritualized behaviours such as courting or nest-buildingbehaviours (Slikas, 1998; Johnson et al., 2000; Stuart & Currie, 2001). Being stereotyped,these behaviours were easily compared among species to assess their patterns of evolution(Lorenz, 1941; Baerends & Baerends, 1950; Hinde, 1955; Tinbergen, 1959). In this context,the homology of stereotyped behaviour has been repeatedly discussed by several seminalpapers (Baerends, 1958; Atz, 1970; Hodos, 1976; Lauder, 1986; Wenzel, 1992) with referenceto the classical criteria of homology proposed by Remane (1952), and especially the criteria ofposition and special quality. More recently, comparative studies of behaviour have taken placein the field of phylogenetic analysis and the homology of stereotyped behavioural characterswas challenged with reference to other characters, either morphological or molecular (e.g.,Coddington, 1986; McLennan, Brooks & McPhail, 1988; Wenzel, 1993; Macedonia &Stanger, 1994; Kennedy, Spencer & Gray, 1996; Stuart & Hunter, 1998; Johnson, 2000; Cap,Aulagnier & Deleporte, 2002; Noll, 2002; Price & Lanyon, 2002; Desutter-Grandcolas &Robillard, 2003; Robillard, Höbel & Gerhardt, 2006). Empirically, these studies confirmedthe old perception that behaviour is not only phylogenetically informative but also oftenimportant to study from such a perspective (De Queiroz & Wimberger, 1993; Proctor, 1996;Brooks & McLennan, 2002).Behaviour is usually described in terms of sequences which are ordered lists ofbehavioural acts, or events, expressed by one or several individuals. Stereotyped sequencesoffered good opportunities for establishing homologies as they are relatively stable withineach species. Behavioural homology can be established with reference to the criterion ofposition by considering the place of different behaviours along the behavioural sequence(Robillard et al., 2006b), and also with reference to the criterion of special quality, asreviewed by Wenzel (1992). This refined treatment of behavioural homology is akin toethologists considering the order of events in many sequences, especially when there is arelation of causality between an event and the following (see Abbott, 1995; Abbott & Tsay,2000). This situation has been analysed to study patterns of mining navigation on web sites(e.g., Hay, Wets & Vanhoof, 2004) and clearly occurs in behavioural interactions.However, stereotyped behavioural sequences represent only a small part of allpotentially available behavioural data and non-stereotyped behaviours the major part.Non-stereotyped behavioural sequences are most often observed in the context of socialrelationships. In this case, the observed behavioural sequence is a series of acts emitted bytwo individuals in alternation. These behavioural relationships vary from one individualto another, and from one context to another, preventing homologies to be hypothesizedwith reference to the criterion of position. In addition, the elementary acts which are280