Celebrations and bursts of communal joy can occur spontaneously in human communities based on mechanisms of emotional contagion. Some examples of similar collective excitement have been reported in animals when they reunite or anticipate rewards, but little is known about the processes and meaning of these multiple interactions. We experimentally studied such collective arousals in two captive groups of Tonkean macaques (Macaca tonkeana) within the context of reunions following the temporary separation of two subgroups. We compared the behaviors of individuals after separation periods of 2 and 48 h with a control period with no separation. This study showed that it is possible to reproducibly induce bursts of friendly interactions in which groupmates run around over a period of several minutes, embracing and grasping one another while displaying numerous affiliative vocalizations and facial expressions. The longer the period of separation, the higher and longer-lasting the rates of affiliative interactions were. Individuals affiliated more frequently with groupmates from a previously separated subgroup than with those having stayed in their own subgroup. Collective arousal was followed by a quieter period characterized by high rates of contact-sitting and social grooming. These results point at the role of collective arousals in social cohesion; they could resolve social tension and renew social relationships. We propose that the emotional state experienced by Tonkean macaques during such events represents a disposition similar to that giving rise to what we humans call ''shared joy. Collective celebration is common practice in human communities. Amongst various other social functions, it can be argued that it either has a role in social cohesiveness by renewing the bonds that hold society together, or that it provides individuals with occasional relief from hierarchies and social constraints (Ehrenreich, 2007). Whereas certain events such as dances, rituals, and ceremonies follow formalized rules and/or use techniques of ectasy, others occur as spontaneous bursts of excitement where groups appear united by communal joy and exaltation, all individuals experiencing a similar emotional state (Ehrenreich, 2007). In the second case, social interactions mainly rely on nonverbal communication; individuals synchronize and mimic each other through vocalizations, facial expressions, postures and movements conveying emotional contagion (Hatfield et al., 1994).There is no need to resort to cognitive empathy to account for emotional contagion. The ability to react to the feelings of others appears early in life (Hoffman, 1975;Hatfield et al., 1994), and its neurobiological substrates have been documented (Decety and Jackson, 2004; Rizzolatti et al., 2007). Moreover, it is widely known that nonhuman primates react to communication signals expressing emotions (Seyfarth and Cheney, 2003;de Waal, 2008), and synchronize by mirroring the behaviors of conspecifics (Anderson and Bayart, 1985;Ferrari et al., 2006;Meunier et ...