Fast and accurate decisions through collective vigilance in fish shoals

Abstract

Although it has been suggested that large animal groups should make better decisions than smaller groups, there are few empirical demonstrations of this phenomenon and still fewer explanations of the how these improvements may be made. Here we show that both speed and accuracy of decision making increase with group size in fish shoals under predation threat. We examined two plausible mechanisms for this improvement: first, that groups are guided by a small proportion of high-quality decision makers and, second, that group members use self-organized division of vigilance. Repeated testing of individuals showed no evidence of different decision-making abilities between individual fish. Instead, we suggest that shoals achieve greater decision-making efficiencies through division of labor combined with social information transfer. Our results should prompt reconsideration of how we view cooperation in animal groups with fluid membership.

Fig. 1.

Experimental apparatus.

Fig. 2.

(A) proportion of focal fish (×) and of all fish (○) avoiding the predator as a function of group size. (B) Mean swimming speed (±SEM) of focal fish in each group size. (C) Mean time (±SEM) spent in each zone by focal fish. (D) Mean path tortuosity (±SEM) of focal fish in each group size. In B–D, open bars refer to the approach zone, and shaded bars refer to the decision zone.

Fig. 3.

Mean proportion (±SEM) of focal fish choosing the branch that avoids the predator as a function of the number of companions that avoid the predator minus the number that pass the predator. The fit of the logistic model is made assuming the probability of a single fish is equal to that over all our single-fish observations, i.e., 55.7%. The solid line shows the best fit of the model given this assumption.

Fig. 4.

Correlations between the leading pair of fish. (A) Example of fish positions through time from a 2-fish trial. Solid circles are the position of the leading fish, and open circles are the following fish. Labels “1s” and “2s” indicate the position of that fish at the time in seconds since the first fish entered the approach zone. (B) Spatial correlation of fish direction over the first 2 fish in all 2-, 4-, 8-, and 16-fish trials. The change in spatial position during each time step of 0.2 s was calculated for 2 leading fish at any point in time (note that the leading fish can change during a trial). The correlation coefficient was calculated between the direction of the leading and the second fish, with binning according to distance between the fish. Error bars show 95% confidence intervals. The number of observations varied between bins: 0–2 cm, n = 14; 2–4 cm, n = 142; 4–6 cm, n = 49; 6–8 cm, n = 26; ≥8 cm, n = 48. (C) Temporal correlation. For all leading pairs within 6 cm of each other, the correlation coefficient was calculated for time lags of 0 up to 2 s, in steps of 0.2 s. The solid line is the temporal correlation between the leader fish direction and the second fish direction at different time lags. Error bars show 95% confidence intervals. The dotted line gives the autocorrelation of the direction of the leader fish, i.e., correlation in the direction time-lagged for various times.