Are kea prosocial?

Abstract Prosocial behaviour is widespread in humans, but evidence for its occurrence in other species is mixed. We presented a parrot species, the kea (Nestor notabilis) with a series of experiments to test whether they exhibit prosocial tendencies. Across the first round of testing, in our first condition, two of the four kea acted prosocially, as they preferred to choose a prosocial token which rewarded both themselves and a partner, rather than a token that rewarded only themselves. Three of the four kea then showed a preference for the prosocial token in a second condition where they alternated taking turns with a partner. However, no kea showed a decrease in the third yoked control condition in which the experimenter replicated the token choice made by the partner in the previous alternating trials. This yoked condition was used to dissociate truly reciprocal behaviour, whereby the actor made choices based on their partner's choices, from a response to the amount of rewards conferred to the partner. Finally, three of the four kea continued to choose the prosocial token in the fourth asocial control condition where no partner was present. However, in round two of testing, one kea changed its token choices to a similar pattern to that expected if kea are prosocial, in that it preferred the prosocial token in the initial condition, showed a trend for the prosocial token when turns were alternated, but chose at chance in the yoked and asocial conditions. This study therefore found no evidence of spontaneous reciprocity in kea but further testing is required before we can conclude that kea are not capable of prosocial behaviour at all.1 Introduction 2 Methods 2.1 Subjects 2.2 Apparatus 2.3 Preliminary tasks and familiarisation 2.3.1 Token exchange training 2.3.2 Token preference test 2.3.3 Training on token values 2.4 General procedure 2.4.1 Description of conditions 2.4.2 Testing 2.4.3 Analyses 3 Results 4 Discussio