Post-flight initiation distance response and correlation
Squirrels ran into shelter in 30% of trials. Juveniles sheltered (as opposed to ‘stop and look’) more frequently than adults (Table S2). Squirrels tested near a burrow were more likely to run into shelter (Table S2), but no other factor significantly explained variation in likelihood of running to shelter. Notably, the decision to run into shelter was not repeatable (R = 0.038 [0, 0.119]), but time to emerge from shelter was repeatable (R = 0.39 [0.191, 0.59]). Individuals from home sites that experience greater human activity emerged from shelter more quickly than squirrels from home sites with lower human activity levels (-2.336 [-4.413, -0.431]; Table S3; Fig. 3A). Furthermore, more trappable individuals had shorter shelter emergence times than less trappable individuals (-1.359 [-2.456, -0.359]; Table S3; Fig. 3B). We also found a high positive among-individual correlation between emergence time from shelter and FID (0.57 [0.139, 0.967]; Fig. 4A), indicating that individuals that allowed walkers to approach more closely before fleeing were also faster to re-emerge from shelter. Time to emerge from shelter decreased with trial number (0.137 [0.030, 0.241]; Table S4).
The response to stop and look from a distance occurred in 70% of trials (stop look distance mean: 5.1 ± 0.3 S.E. m; range: 0.2 m to 27.7m). Stop and look distances were only weakly repeatable (R = 0.216 [0.073, 0.370]) and we found no significant predictors for stop look distance (Table S4). However, we did find a high positive among-individual correlation between stop look distance and FID, suggesting that individuals that allowed walkers to approach more closely before fleeing also fled shorter distances from the walker following their FID response (0.736 [0.428, 0.988]; Fig. 4B).