Global meta-analysis reveals urban-associated behavioural differences among wild populations.

Urbanization causes fundamental changes to natural environments, causing rapid and substantial adaptive phenotypic change in wild populations. While many studies have investigated how urbanization may shape interspecific behavioural variation, for example, via urban environmental filtering, no study has yet quantitatively assessed broad, global patterns of urban-associated intraspecific behavioural variation. Here, we conducted a phylogenetic meta-analysis to assess urban-associated behavioural differences in wild populations of birds, mammals, amphibians, reptiles and insects. We focused on four commonly measured behaviours (boldness, aggressiveness, activity, exploration) and extracted paired urban-nonurban effect size estimates for behavioural means and variances (k = 279), repeatability (k = 13) and correlations (k = 14) from 81 unique studies. We found evidence that urban populations exhibited heightened average boldness, aggressiveness, exploration and activity compared to nonurban conspecifics, a result that was robust across geographic regions and ecological niches. However, avian species were strongly overrepresented in our meta-analysis (N = 49 studies) compared to other taxa (Mammalia, N = 20; Reptilia, N = 7; Insecta, N = 3; Amphibia, N = 2). Consequently, most behaviours in non-avian taxa were under-sampled, and effect sizes were generally not statistically significant: among non-avian species, only boldness differed significantly between urban and nonurban populations. This strong taxonomic bias potentially affects our other inferences as well. We did not find strong evidence linking urbanization to changes in behavioural variation, repeatability or correlations. Our results summarize data from the rapidly growing field of urban evolutionary ecology and demonstrate geographically widespread differences in behaviour between urban and nonurban populations. These patterns suggest that urban populations experience parallel directional selection or that urban invaders experience environmental filtering by common urban conditions that favour certain behavioural types. Taxonomic biases and methodological heterogeneity continue to limit inference in the field, constraining our ability to test preregistered hypotheses in this study. Broadening research to include understudied taxa such as invertebrates, herpetofauna and nocturnal species; incorporating common-garden approaches; and emphasizing clearer definitions for behaviour and urbanization will yield a more comprehensive understanding of how urbanization shapes animal behaviour and other ecologically meaningful phenotypes.