Animals living in groups tend to express less variable behaviour than animals living alone. It is widely assumed that this difference reflects, at least in part, an adaptive response to contrasting selection pressures: group-living should favour the evolution of more uniform behaviour whereas lone-living should favour behaviour that is less predictable. Empirical evidence linking these contrasting selection pressures to intrinsic differences in behavioural variability is, however, largely lacking. The desert locust, Schistocerca gregaria, manifests in two very distinct eco-phenotypes, a lone-living cryptic “solitarious phase” and a swarming “gregarious phase” that aggregates into very large and dense groups. This “phase polyphenism” has evolved in response to contrasting selection pressures that change rapidly and unpredictably. Phase differences in mean behaviour are well-characterised, but no previous study has considered differences in variability. Here we used locust phase polyphenism to test the hypothesis that group living leads to the evolution of reduced intrinsic variability in behaviour. We measured two behaviours in both phenotypes: locomotor activity in the presence of conspecifics, and locomotor hesitation in approaching food when alone. We assayed each individual repeatedly and estimated variability relative to the mean in log-normal mixed-effects models that explicitly account for the means-variance dependency in the behavioural measures. Our results demonstrate that relative behavioural variability differs between the two phases in line with predictions from ecological theory: both within-individual and between-individual variability were lower in the group-living gregarious phenotype. This contrasts with previous studies on social niche construction in spiders and crickets, and highlights the importance of social ecology: in animals that form non-social collectives, such as locusts, reduced individual behavioural variability is key for coherent collective behaviour. The differences in variability persisted when gregarious locusts were tested in isolation and solitarious locusts were tested in groups, indicating that they arise not simply as flexible reactions to different social contexts, but are intrinsic to the individual animals of each phase. This “variance polyphenism” in locusts provides empirical evidence that evolutionary adaptation for group living has driven a reduction in within- and between-individual behavioural variability.
Plasticity of behavioural variability reflects conflicting selection in group-living and solitary desert locusts
