Competing for position in the communal roosts of long-tailed tits ANDREW MCGOWAN * , STUART P. SHARP †, MICHELLE SIMEONI † & BEN J. HATCHWELL † *Centre for Ecology and Conservation, University of Exeter in Cornwall yEvolution and Behaviour Group, Department of Animal and Plant Sciences, University of Sheffield (Received 27 June 2005; initial acceptance 26 September 2005; final acceptance 16 February 2006; published online 22 September 2006; MS. number: 8597R) Communal roosting is widespread among social animals and has several hypothesized functions, includ- ing reduction of predation risk and thermoregulatory costs, and information sharing. The benefit derived from roosting in close proximity to conspecifics is likely to depend on an individual’s position within the roost, but there have been few quantitative studies of the process of communal roost formation and the factors influencing relative positions. We studied the communal roosting behaviour of temporarily captive flocks of long-tailed tits, Aegithalos caudatus, a cooperative breeder that forms flocks during the nonbreed- ing season. Our objectives were to determine whether (1) individuals compete for particular positions within roosts, (2) individuals achieve consistent positions within roosts, and (3) an individual’s roost po- sition is a function of its dominance status. During roost formation birds were more likely to move to inner positions regardless of whether they started in an outer or inner position. However, as the number of birds in the roost increased during roost formation, birds in outer positions were less likely to relocate to an in- ner position. The same individuals occupied outer positions in a roost on successive nights, but there was no consistency in the occupation of specific inner positions. Birds occupying outer roost positions were significantly less dominant than those occupying inner positions. Our results show that long-tailed tits compete for inner positions within communal roosts, and that an individual’s dominance status within a flock influences the outcome of this competition. Ó 2006 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. Communal roosting behaviour is widespread among animals (Eiserer 1984; Lewis 1995; Anderson 1998) and in birds it is known to have evolved independently several times (Beauchamp 1999). Communal roosts are thought to increase foraging efficiency by functioning as informa- tion centres (Ward & Zahavi 1973) or recruitment centres (Richner & Heeb 1996; Dall 2002), and they may also confer benefits by decreasing predation risk or thermo- regulatory costs (Beauchamp 1999; DuPlessis 2004). The magnitude of such benefits may depend on an individual’s position within a communal roost, which in turn may be influenced by factors such as age, sex and dominance status (Swingland 1977; Summers et al. 1987; Feare et al. 1995; Adams et al. 2000; Calf et al. 2002), although evi- dence of any benefits associated with specific roosting positions is scarce. Communal roosting appears to be relatively common in cooperatively breeding birds, although there have been few detailed studies (e.g. Chaplin 1982; Ligon et al. 1988; Williams et al. 1991; DuPlessis & Williams 1994; DuPlessis et al. 1994). Communal roosts are unlikely to function as information or recruitment centres in cooperatively breed- ing birds, as groups tend to be cohesive family-based units that spend all of their time together. Thermoregulatory benefits from group roosting are the most likely explana- tion for communal roosts in cooperative species because studies have shown that the energetic savings of commu- nal roosting can be as high as 37% relative to solitary roosting (Riehm 1970; DuPlessis 2004). However, such benefits are unlikely to be acquired equally by all roost members, so an individual’s position within a roost may have important consequences for its energy budget. Stud- ies of the jungle babbler, Turdoides striatus (Gaston 1977), varied sitella, Daphoenositta chrysoptera (Noske 1985) and Arabian babbler, Turdoides squamiceps (Zahavi 1990), have reported on the spatial organization of their commu- nal roosts, and in all three species the end positions in their linear roosting huddles were occupied by the two most dominant birds in the group, usually the top males. These structured roosts, with dominants adopting outer, Correspondence: A. McGowan, Marine Turtle Research Group, Centre for Ecology and Conservation, University of Exeter in Cornwall, Trem- ough Campus, Penryn TR0 9EZ, U.K. (email: amcgowan@seaturtle. org). S. P. Sharp, M. Simeoni & B. J. Hatchwell are at the Evolution and Behaviour Group, Department of Animal and Plant Sciences, University of Sheffield, Sheffield S10 2TN, U.K. 1035 0003e 3472/06/$30.00/0 Ó 2006 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved. ANIMAL BEHAVIOUR, 2006, 72, 1035e1043 doi:10.1016/j.anbehav.2006.02.020