GROUP-SIZE AND THE COST OF AGONISTIC BEHAVIOR IN PUMPKINSEED SUNFISH

This laboratory study investigated the costs, in terms of energy or time, of grouping as a function of group size. I manipulated group size by randomly assembling small and medium sized pumpkinseed sunfish (Lepomis gibbosus) into groups of various sizes. Agonistic interactions within groups were highly non-random. Consequently, linear dominance hierarchies often could not be inferred due to too many dominance relationships not being expressed at all. The interaction frequency per group per hour increased, whereas the interaction frequency per individual per hour (agonistic index) decreased with group size; the agonistic activity of the most dominant individual within each group did not change with group size. A null model predicting that the agonistic index should increase in proportion to the number of potential dyadic relationships in any given group could be rejected. Instead, the data supported an alternative hypothesis derived from the theory of economic defensibility of a resource (here, a feeding space conferring food with a certain probability) in that the per capita frequency of agonistic interactions decreased with increasing intruder pressure (i.e. group size). Based only on the assessment of the probability of being engaged in a fight, potential subordinates should hence prefer larger groups (a type of intra-group dilution effect), while potential dominants should have no group size preferences other than solitary lifestyles. The most dominant individual in each group acquired more food items and gained more body mass than the average subordinate. Subordinates showed reduced growth rates relative to their food uptake, which was likely related to factors other than locomotion. No costs to subordinates in terms of injuries were apparent. No energetic costs to dominants as a function of their agonistic activity were apparent, but a cost in terms of time or <<missed opportunities>> was detectable. These findings are in contrast to studies on the costs of agonistic behavior and dominance in birds, but corroborate conclusions of studies of arthropods.