Third party interventions on affiliative interactions protect social bonds

Third-party interventions are defined as the interruption of dyadic interactions by third animals through direct physical contact, interposing or threats. Previous studies focused on analysing interventions against agonistic encounters. However, evaluations of interventions against affiliative behaviours in relation to the intervening animal’s social relationships and its social and spatial position are missing. Therefore, horses are an interesting model species, as interventions occur against affiliative interactions most frequently. We analysed 67 interventions in affiliative interactions, displayed between 64 feral horses (Equus ferus caballus) in three groups. Particular, high ranking females intervened in the affiliative interactions of group mates in the stable horse harems. Intervening animals took an active part in affiliative and agonistic encounters within the group, but did not occupy particularly social roles or spatial positions in the group. They generally supported preferred group mates when intervening and thus protect their own social bonds. Some species may prevent competition for preferred partners from escalating into more costly agonistic encounters by protecting their social bonds through affiliative interventions. Intervention behaviour primarily has been evaluated in monkeys and apes (deWaal 1992, Silk 1992, Prud’homme & Chapais 1996, Flack et al. 2006), in goats (Keil & Sambraus 1998), fallow deer (Jennings et al. 2009), in goats (Keil & Sambraus 1998), fallow deer (Jennings et al. 2009), hyenas (Zabel et al. 1992), dogs (de Villiers et al. 2003, Ward et al. 2009), dolphins (Connor et al. 1992), and cichlid fish (Walter & Trillmich 1994). Previous evaluations have dealt with third-party interventions on agonistic interactions. It has been assumed that intervener animals occupy certain social positions as only a few particular individuals in a group display intervention behaviour in many species, with others rarely or never intervening (spotted hyaena: Zabel et al. 1992, wild dogs: Villiers et al. 2003, pigtailed macaques: Flack et al. 2006, zebras: Schilder 1990, horses: Sigurjόnsdόttir et al. 2003, van Dierendonck et al. 2009). For agonistic encounters in primates Flack et al. (2006) found interveners to be strongly connected in three of four affiliative social networks, but not occupying any unique position. When they removed the intervening animals, aggression levels rose within the group. They concluded that interveners may engage in policing to prevent the escalation of aggressions within the group. However, deWaal (1992), and Petit & Thierry (1994) argued that interveners support preferred animals by protecting them in agonistic encounters. Several studies have evaluated the position of interveners in dominance hierarchies. Interveners in agonistic encounters were often high-ranking (Engh et al. 2000, Flack et al. 2006, Jennings et al. 2009), and usually they ranked higher than the supported and challenged animals (e.g. bovines: Sambraus 1969, goats: Keil & Sambraus 1998, fallow deer: Jennings et al. 2009). However, rank relationships were not found to influence interventions on affiliative interactions in zebras (Schilder 1990), and in horses interveners were claimed to be lower in rank than both the supported and the targeted animal (Heitor et al. 2006, van Dierendonck et al. 2009). In addition to social bonds and social ranks, the spatial position of animals in their groups could be an important factor in the decision to intervene. In prior