Abstract

Conspicuous, sexually dimorphic plumage in birds is most likely a consequence of sexual selection favouring more ornamented males at obtaining a territory and/or a mate. Recent comparative analyses suggest that, among socially monogamous species, extra-pair paternity has also contributed to the elaboration of male ornaments. If females prefer more ornamented males as social or extra-pair mates this could translate into strong directional selection for ornament elaboration, since these males might sire more offspring in their own brood (within-pair success) or in broods of other males (extra-pair success). In this thesis I study the expression of the UV/blue crown coloration in the blue tit (Parus caeruleus) and investigate whether this trait could be selected through increased male within- or extra-pair success. Blue tits are socially monogamous passerines with relatively high levels of extra-pair paternity, and males in this species display more ultraviolet(UV)/blue reflectant crown feathers than females. Based on three years of data I found that crown coloration could be a cue used by females to assess male age since blue tits became more UV-ornamented as they aged. Crown coloration, however, did not correlate with survival to the next breeding season, suggesting that more UV-ornamented males are not necessarily of higher quality. While crown UV-ornamentation increased between years, it declined in the course of a year due to feather wear and dirt accumulation and this could affect female perception of male attractiveness. However, although the decline in UV ornamentation between winter and spring was large, it had no effect on male reproductive success, and winter and spring colour were still positively correlated. Using genetic paternity analysis I could show that more UV-ornamented males do not benefit through increased within-pair or extra-pair success. On the contrary, less UV-ornamented, adult males sired most of the extra-pair offspring. Hence the most successful males of the population were adult males that resembled juveniles in their crown colour. Accordingly, females seem to recognise less UV-ornamented males as highly successful, since they biased brood sex ratio towards male offspring if paired to these males. The causality of these patterns was tested in a colour manipulation experiment, where I treated males to become more (UV+) or less (UV-) ornamented within the natural range of variation. Against expectations UV(+) males sired significantly more extra-pair offspring than UV(-) males while the proportion of within-pair offspring was unaffected by the manipulation. Brood sex ratios did not differ between treatments but depended on male colour before manipulation. While these results do suggest that crown colour plays a role in paternity and brood sex allocation, they do not provide experimental support for the observed correlational patterns. I discuss the discrepancy between observational and experimental data, emphasising potential problems with the experimental manipulation of structural plumage colour. Finally, given that more UV-ornamented males did not sire more offspring, I explore the possibility that they would benefit by pairing with high quality females. High quality females in other species are often more ornamented, and birds of high quality pair assortatively based on ornament expression. This was not the case in this blue tit population, since female colour did not appear to indicate relevant female qualities (fecundity, seasonal reproductive success) and blue tits did not mate assortatively by crown colour in any of the three study years. To conclude, selection seems to favour older, less UV-ornamented males in this population. Whether this is due to female preference is unclear. Alternatively I hypothesise that being less ornamented may enable males searching for extra pair copulations to intrude into other territories without eliciting aggression by territory owners, perhaps by mimicking juveniles. Detailed behavioural observations in the wild coupled with choice chamber experiments in captivity are necessary to test this idea.