The spectacular diversity of personality and behaviour of animals and humans has evoked many hypotheses intended to explain its developmental and evolutionary background. Although the list of the possible contributing mechanisms seems long, we propose that an underemphasised explanation is the division of labour creating negative frequency dependent selection. We use analytical and numerical models of social division of labour to show how selection can create consistent and heritable behavioural differences in a population, where randomly sampled individuals solve a collective task together. We assume that the collective task needs collaboration of individuals performing one of the two possible subtasks. The total benefit of the group is highest when the ratio of different subtasks is closest to 1. The probability of choosing one of the two costly subtasks and the costs assigned to them are under selection. By using adaptive dynamics we show that if a trade-off between the costs of the subtasks is strong enough, then evolution leads to coexistence of specialized individuals performing one of the subtasks with high probability and low cost. Our analytical results were verified and extended by numerical simulations.

No abstract

No abstract

The emergence of animal societies offers unsolved problems for both evolutionary and ecological studies. Social spiders are specially well suited to address this problem given their multiple independent origins and distinct geographical distribution. Based on long term research on the spider genus Anelosimus, we developed a spatial model that recreates observed macroecological patterns in the distribution of social and subsocial spiders. We show that parallel gradients of increasing insect size and disturbance (rain, predation) with proximity to the lowland tropical rainforest would explain why social species are concentrated in the lowland wet tropics, but absent from higher elevations and latitudes. The model further shows that disturbance, which disproportionately affects small colonies, not only creates conditions that require group living, but also tempers the dynamics of large social groups. Similarly simple underlying processes, albeit with different players on a somewhat different stage, may explain the diversity of other social systems.

The facultatively social spider Anelosimus studiosus offers a unique opportunity for understanding how multilevel selection acts in natural populations. However, the importance of previous empirical studies are shaded by a conceptual debate about whether colony-level selection is truly present in these populations or not. Here we introduce a detailed individual based model, where practically all assumptions are supported by empirical data. The only element of the female A. studiosus life cycle missing from the literature is how maturing female spiders decide whether to disperse. This behavioural component we estimate with evolutionary simulations. This model is able to recapitulate the characteristic size and composition distributions of natural populations in different environments. The evolutionary simulations revealed that the optimal dispersal behaviour of a maturing female varies both with her ecological environment and behavioural phenotype. This finding is open for straightforward empirical testing. In agreement with empirical findings we have established parameter ranges where the population is prone to extinction without multiple-female nests. We propose that the dispersal behaviour of individuals is both the result and the prerequisite of multilevel selection in this species.