In spite of strong evidence for viability-based sexual selection and sex ratio adjustments, the blue tit, Parus caeruleus, is regarded as nearly sexually monomorphic and no epigamic signals have been found. The plumage coloration has not, however, been studied in relation to bird vision, which extends to the UV-A waveband (320^400 nm). Using molecular sex determination and UV/VIS spectrometry, we report here that blue tits are sexually dichromatic in UV/blue spectral purity (chroma) of the brilliant crown patch. It is displayed in courtship by horizontal posturing and erected nape feathers. A previously undescribed sexual dimorphism in crown size (controlling for body size) further supports its role as an epigamic ornament. Against`grey-brown' leaf litter and bark during pair formation in early spring, but also against green vegetation, UV contributes strongly to conspicuousness and sexual dimorphism. This should be further enhanced by the UV/bluish early morning skylight (`woodland shade') in which blue tits display. Among 18 breeding pairs, there was strong assortative mating with respect to UV chroma, but not size, of the crown ornament. We conclude that blue tits are markedly sex dimorphic in their own visual world, and that UV/violet coloration probably plays a role in blue tit mate acquisition.

The blue throat feathers of male bluethroats (Luscinia s. svecica) show a re£ectance peak in the ultraviolet (UV) waveband (320^400 nm). The throat is actively displayed during courtship, suggesting a role for sexual selection on an ultraviolet signal. Indeed, a recent aviary experiment demonstrated that females discriminated against males with arti¢cially reduced UV re£ectance (Andersson & Amundsen 1997). Here, we report the results of a similar experimental manipulation applied on free-ranging males. UVreduced (UVR) males had a lower success in attracting mates, as judged from a signi¢cantly later start of egg laying, compared with control (C) males. UVR males also spent signi¢cantly less time advertising for additional mates when their own mate was fertile, and they had a lower success in achieving extrapair fertilizations. Furthermore, UVR males tended to guard their mates more closely and lose more paternity in their own brood than C males did. We conclude that the treatment a¡ected both social and extra-pair mate choice. This is the ¢rst experimental evidence that UV signalling in£uences male mating success in free-ranging birds.