Beyond first clutches: Second broods reshape selection on breeding timing in forest and urban great tits.

Urbanisation provides a unique framework for studying how organisms respond to human-modified environments. Although phenotypic differences between rural and urban populations are increasingly well documented, our knowledge of how natural selection operates in cities remains limited. In particular, few studies have quantified the direction and shape of natural selection acting on breeding phenology in urban environments. Using a 13-year dataset on great tits (Parus major), we investigated how first clutch laying date (FCLD) influences both fledgling output from first broods and total seasonal productivity (first and second clutches combined). We first tested whether FCLD differed between habitats and among urban plots. We then estimated natural selection on FCLD relating breeding phenology to fledgling number from (i) first broods only and (ii) total annual productivity including first and second broods. Finally, we evaluated how FCLD and habitat jointly affected the probability of multiple brooding. As in previous studies, FCLD occurred earlier in more urbanised areas. With first broods only, earlier breeding tended to be associated with higher fledging numbers in forest but not in urban habitats. In contrast, when including second clutches, a higher annual productivity was found for earlier breeders in both habitats. When controlling for clutch size, no direct selection on FCLD was detected. Earlier breeders were also more likely to initiate a second clutch, particularly in urban populations. Finally, structural equation models further showed that the effect of FCLD on annual fledgling production was largely indirect, mediated by the probability of multiple brooding and the resulting increase in annual egg number. Together, our analyses suggest that earlier laying females achieve higher annual fledgling productivity, mainly determined by a higher probability of double brooders and the resulting higher seasonal egg production in both forest and city. In general, these results highlight the importance of evaluating selection using annual reproductive output rather than single breeding attempts. We believe that the underlying proximate mechanism that may promote this advantage may be related to the dynamics of prey abundance throughout the birds' breeding season, and we encourage further studies that would compare these prey dynamics across habitats.