Résumé
1. Urbanization is one of the most significant contributors to the Anthropocene, and urban evolutionary ecology has become an important field of research. While it is commonly assumed that cities impose new and stronger selection, the contradictory assertion that selection may be relaxed in cities is also frequently mentioned, and overall, our understanding of the effects of urbanization on natural selection is incomplete.2. In this review, we first conduct a literature search to find evidence for patterns of natural selection on phenotypic traits including morphology, physiology, behaviour and life history, in urban and non-urban populations of animals and plants. This search reveals that coefficients of natural selection in the context of urbanization are scarce (n = 8 studies providing selection gradients/differentials that include a total of n = 200 coefficients) and a lack of standardized methods hinders quantitative comparisons across studies (e.g. with meta-analysis). These studies, however, provide interesting insight on the agents shaping natural selection in cities and improve our mechanistic understanding of selection processes at different spatial scales.3. We then perform a second literature search to review genomic studies assessing selection intensity in cities, on the genome of non-human natural populations. While this search returns 383 articles, only 34 of these truly investigate foot-prints of selection associated with urbanization, and only one study provides urban genetic selection coefficients. Here again, we find highly heterogeneous approaches, yet some studies provide strong evidence of genomic footprints of urban adaptation.4. In neither the phenotypic nor genomic literature review were we able to quantitatively assess natural selection across urban versus non-urban habitats. Thus, we propose a roadmap of how future studies should provide standardized metrics to facilitate mega- or meta-analyses and explore generalized effects of urbanization on selection.