There has been an increased focus on the role of natural and sexual selection in shaping cognitive abilities, but the importance of the interaction between both forces remain largely unknown. Intersexual selection through female mate choice might be an important driver of the evolution of cognitive traits, especially in monogamous species, where females may obtain direct fitness benefits by choosing mates with better cognitive abilities. However, the importance given by females to male cognitive traits might vary among species and/or populations according to their life-history traits and ecology. To disentangle the effects of natural and sexual selection, here we use an agent-based simulation model and compare the model’s predictions when females mate with the first randomly encountered male (i.e. under natural selection) versus when they choose among males based on their cognitive trait values (i.e. under natural and intersexual selection). Males and females are characterized, respectively, by their problem-solving ability and assessment strategy. At each generation, agents go through (1) a choosing phase during which females assess the cognitive abilities of potential mates until eventually finding an acceptable one, and (2) a reproductive phase during which all males compete for limited resources that are exploited at a rate which depends on their cognitive abilities. Because males provide paternal care, the foraging success of mated males determines the breeding success of the pair through its effect on nestling provisioning efficiency. The model predicts that intersexual selection plays a major role in most ecological conditions, by either reinforcing or acting against the effect of natural selection. The latter case occurs under harsh environmental conditions, where intersexual selection contributes to maintaining cognitive diversity. Our findings thus demonstrate the importance of considering the interaction between both selective forces and highlight the need to build a conceptual framework to target relevant cognitive traits.

The dataset was extracted between 28/09/2021 and 05/10/2021 from an agent-based model composed by a simulation (i.e. producing the raw data), and a script that convert the raw data into the actual shared dataset. The dataset is composed by two files (i.e. NaturalSelection and NaturalSexualSelection), each including 36 files of datasets (i.e. TypeOfSelection_ForagingPeriod_ChoosingPeriod_PatchQuality_CognitiveCost). Every file is composed by three document.txt (i.e. AnalyseLinkSRFemaleCriterionEvolutionFile, AnalyseLinkSRFemalePrecisionEvolution, AnalyseLinkSRMaleTraitEvolution) representing the datasets for each relevant trait in our simulation (i.e. female selectivity trait, female accuracy trait and male cognitive trait respectively). Each dataset contains the mean and sd reproductive success for each the trait value, the generation (i.e. from 0 to 999) and the range (i.e. from 5 to 185 food patch number, by 5). These datasets allowed us to create all figures of the mean frequency of each male and female trait values in 360 environmental conditions, under natural selection alone and under natural and sexual selection. 

The simulation model was developed using the Python programming language, while the simulations were run on Canada compute servers (Calcul Quebec). All figures were generated in R studio (R Development Core Team, 2021; version 4.0.2).