Changes in movement characteristics in response to private and social information acquisition of socially foraging fish

Abstract

To overcome the cost of competition resulting from foraging socially, individuals may balance their use of private (i.e. acquired from personal sampling) and social (i.e. acquired by watching other individuals) information to adjust their foraging strategy accordingly. Reliability of private information about environmental characteristics, such as the spatial distribution of prey, is thus likely to affect individual movement and social interactions. We aimed to investigate how movement characteristics of foraging individuals changed as they acquired reliable information about the spatial occurrence of prey in a foraging context. We allowed guppies, Poecilia reticulata, to develop the reliability of their private knowledge about prey spatial occurrence by repeatedly testing shoals in a foraging task under three experimental distributions of prey: (1) aggregated prey forming three patches located in fixed locations, (2) scattered distribution of prey with random locations, or (3) no prey (used as a control). Using individual time series of spatial coordinates, we computed a suite of movement variables reflecting search effort, social proximity and locomotion characteristics during foraging, to examine changes occurring over repeated trials. Over time, individuals foraging on either scattered or aggregated prey travelled greater distances, showed an increasing distance to their closest neighbour and became slightly more stochastic in their acceleration profile, compared to control individuals. We found that behaviour changed as private information increased over time, with a clear behavioural shift and an increase of collective foraging efficiency occurring on the third testing day. Social proximity was the major predictor of foraging success in the absence of prior foraging information, while search effort became the most important predictor of foraging success as information increased. In conclusion, we showed that individuals' movement patterns changed as they acquired private information. Contrary to our predictions, the spatial distribution of prey did not affect any of the movement variables of interest.