Background/Question/Methods In many social animals, groups occur at optimal sizes that maximize foraging success; this despite the fact that theory predicts that groups should grow to equilibrium group sizes (i.e. the group size when additional members decrease fitness below that of foraging alone). Beauchamp and Fernandez (2004) outline several hypotheses for group regulation in attempt to address the group size paradox.  Prior research has shown that juvenile coho salmon (JCS) is a suitable model for studying group foraging because individuals of this species are easily observed in natural systems, occupy well defined spaces, and forage in groups.  Previous research also has shown that JCS typically occur in groups near an optimal size, suggesting active group regulation. Our study had two objectives, the first of which was to describe temporal dynamics of group membership via underwater and bank-side observations. This first objective was preliminary to our principle objective, which was to evaluate aggression and individual assessment of patch quality as mechanisms of group regulation. In this study, we conducted group size counts on 126 groups of JCS and video recorded 69 groups distributed into group size categories. Video segments were analyzed to determine foraging, aggression, and departure time for focal individuals. 
Results/Conclusions Frequency distributions of group counts satisfied our initial assumption that groups are temporally dynamic and typically maintained near a mean group size.  Results also showed that groups near the optimum are most stable.  Evaluations of the two group regulation mechanisms revealed that transient fish (departure time <100sec into individual trials) were victimized significantly more often than residents (departure time ≥200sec).  This result supported the aggression hypothesis as a mechanism of group size regulation. Results evaluating individual patch assessment showed that transient fish had a significantly shorter duration between feeding events (i.e. higher feeding rate per minute) than residents. These results were contrary to what we expected: that individuals feeding above the average group-feeding rate should stay in the group. Subsequent to this initial analysis we completed an alternative evaluation of the patch assessment hypothesis, focusing on foraging costs for transient and resident fish. From this analysis, we inferred that transient fish expended more energy in food acquisition than residents, a result suggesting that the net energy intake of transient fish might be lower than the net energy intake of resident fish. When viewed this way, our findings do lend support to the patch assessment hypothesis.