Fig. 4: Results of the third experiment, unequal reward distribution.

Results of the third experiment. The pattern of results of the fish’s preference for targets providing food only to itself (non-prosocial targets) or for targets providing a food pellet to the advantage of a neighboring fish (1/2) (unequal prosocial). a–c show the fish raw data, with each graph representing a different fish. Both Y axes represent the proportion of target selection, with the left-hand scale numbered for the yellow color and the right-hand scale for the blue color. The X-axis represents the number of the session. The middle line represents 50% chance. The left side of each graph shows the last ten sessions of the first prosocial experiment in which the prosocial target benefiting both is associated with equal distribution of rewards (1/1). The yellow bars represent the proportion of preference for the equal distribution (pro-social) target, and the blue bars represent the proportion of preference for the non-prosocial target. In the second phase (the current experiment), the yellow bars represent a social outcome to the other fish advantage (1/2) and, as before, the blue bars represent the proportion of preference for the non-prosocial target (1/0). The bars add up to 1. d represents a summary across all fish for each target type selection in both phases. Error bars represent one standard error. The data points represent the average preference of each fish in each condition. Over all, the fish stopped choosing the target that resulted in a social outcome when the passive fish received more food than the acting fish.