Fig. 2
The relationship between the day of testing in both noise (lighter colour) and control (darker colour) treatments and the: a mean speed of individuals; b mean group cohesion (standard deviation of the number of fish in the same arm at any time); c total number of movements into arms during a trial; d proportion of moves into arms that were leadership events; e number of leadership attempts; and f number of following events. Fitted lines are calculated from fixed-effect estimates from the model with all main effects and the treatment*day interaction term. Shaded areas are 95% confidence intervals. Points are individual data points (a, b and c: 18 groups, 72 trials; d, e and f: 18 groups, 68 trials) and are jittered (0.1 units) to reduce overplotting. The plot was generated using the sjplot package in r54.
