Individuals that are consistent in risk-taking benefit during collective foraging

Abstract

It is well established that living in groups helps animals avoid predation and locate resources, but maintaining a group requires collective coordination, which can be difficult when individuals differ from one another. Personality variation (consistent behavioural differences within a population) is already known to be important in group interactions. Growing evidence suggests that individuals also differ in their consistency, i.e. differing in how variable they are over time, and theoretical models predict that this consistency can be beneficial in social contexts. We used three-spined sticklebacks (Gasterosteus aculeatus) to test whether the consistency in, as well as average levels of, risk taking behaviour (i.e. boldness) when individuals were tested alone affects social interactions when fish were retested in groups of 2 and 4. Behavioural consistency, independently of average levels of risk-taking, can be advantageous: more consistent individuals showed higher rates of initiating group movements as leaders, more behavioural coordination by joining others as followers, and greater food consumption. Our results have implications for both group decision making, as groups composed of consistent individuals are more cohesive, and personality traits, as social interactions can have functional consequences for consistency in behaviour and hence the evolution of personality variation.

Figure 1. The experimental tank (to scale).

Fish were habituated in a mesh-covered refuge (shaded grey) before the door (thick dotted black line) was raised remotely. The fish had to swim past a visual barrier before being able to see a food stimulus in one of two arms at the other end of the arena; the dotted lines represent lines of sight when the food stimulus was placed in the right side arm (bottom right of the figure, indicated by the arrow). The dashed lines connected to the food stimulus represent lines of sight when the fish first leave the refuge, and the other line is the point at which the fish were deemed to have made a decision as at this point the stimulus in the other arm (if present) would not be visible. Once this decision had been made, 2 bloodworms per fish being tested were released.

Figure 2. The effect of relative personality trait scores on whether a fish initiated exploration of the arena (i.e. was the first to leave).

(a) Shows the effect of the difference between the two fish’s scores for each trait in two-fish trials. (b) Shows the effect of the difference between a fish’s score in a trait and the median of the group for that trait for groups of 4 fish. Initiators are represented by filled circles and non-initiators by open circles.

Figure 3. The effect of the initiator’s boldness score on their latency to leave the refuge in trials of two fish.

Figure 4. The effect of personality trait scores on following (the time delay between initiators and non-initiators) in two fish trials.

Filled circles are observed data. The colours represent the non-initiator’s consistency score (a) and non-initiator’s boldness score (b), binned every 0.5 units of each score. Coloured lines are fits for each binned interval, calculated from the coefficients of the GLMM which includes the two significant interaction terms (Table S1; note that the models use continuous, not binned, data). The main effect of trial order is fixed at its mean value in the data, as is the value of non-initiator boldness in (a), and non-initiator consistency in (b).

Figure 5. The effect of relative personality trait scores on the proportion of food eaten by each fish in two-fish trials.

The proportion of food is represented on a colour scale. Relatively bolder and more consistent fish ate more food (i.e. there are darker points in the top right quarter compared to the bottom left quarter).