Sexual–selection theory assumes that there are costs associated with ornamental plumage coloration. While pigment–based ornaments have repeatedly been shown to be condition dependent, this has been more difficult to demonstrate for structural colours. We present evidence for condition dependence of both types of plumage colour in nestling blue tits (Parus caeruleus). Using reflectance spectrometry, we show that blue tit nestlings are sexually dichromatic, with males having more chromatic (more 'saturated') and ultraviolet (UV)–shifted tail coloration and more chromatic yellow breast coloration. The sexual dimorphism in nestling tail coloration is qualitatively similar to that of chick–feeding adults from the same population. By contrast, the breast plumage of adult birds is not sexually dichromatic in terms of chroma. In nestlings, the chroma of both tail and breast feathers is positively associated with condition (body mass on day 14). The UV/blue hue of the tail feathers is influenced by paternally inherited genes, as indicated by a maternal half–sibling comparison. We conclude that the expression of both carotenoid–based and structural coloration seems to be condition dependent in blue tit nestlings, and that there are additional genetic effects on the hue of the UV/blue tail feathers. The signalling or other functions of sexual dichromatism in nestlings remain obscure. Our study shows that nestling blue tits are suitable model organisms for the study of ontogenetic costs and heritability of both carotenoid-based and structural colour in birds.

In socially monogamous species, extra–pair paternity can increase the variance in reproductive success and thereby the potential for sexual selection on male ornaments. We studied whether male secondary sexual ornaments are selected through within– and/or extra–pair reproductive success in the blue tit (Parus caeruleus). Male blue tits display a bright blue crown plumage, which reflects substantially in the ultraviolet (UV) and previously has been indicated to be an important sexual signal. We show that males with a more UV–shifted crown hue were less cuckolded, which probably resulted from female preference for more ornamented mates. By contrast, however, older males and males with a less UV–shifted hue sired more extra–pair young. This probably did not reflect direct female preference, since cuckolders were not less UV–ornamented than the males they cuckolded. Alternatively, a trade–off between UV ornamentation and other traits that enhance extra–pair success could explain this pattern. Our results might reflect two alternative male mating tactics, where more UV–ornamented males maximize within–pair success and less UV–ornamented males maximize extra–pair success. Since crown colour was selected in opposite directions by within–pair and extra–pair paternity, directional selection through extra–pair matings seemed weak, at least in this population and breeding season. Reduced intensity of sexual selection due to alternative mating tactics constitutes a potential mechanism maintaining additive genetic variance of male ornaments.