Sexual dichromatism may not be a good index of sexual or natural selection in the blue cardinalids (Aves: Passeriformes)

Abstract

Abstract More than a century ago, Darwin and Wallace started a still ongoing debate over which are the predominant forces driving sexual dichromatism (i.e. differences in body coloration between males and females): is it sexual selection on males to become more attractive, or is it natural selection on females to become more cryptic? While these are not mutually exclusive, the degree of sexual dichromatism has been extensively used as a proxy of the intensity of one type of selection (sexual) on one of the sexes (males). Here, we evaluated the relationship between sexual dichromatism and two body-color features that can be under natural and/or sexual selection in each sex: conspicuousness against the background and colorfulness (which we defined as the variety of colors and mechanisms to produce them within an individual’s plumage). We focused on the “blue clade” of the Cardinalidae bird family and considered the properties of their own visual system and those of potential raptor predators. We found that all blue cardinalids are sexually dichromatic, but levels of dichromatism vary within the clade. Males are on average more colorful than females, but neither male nor female colorfulness correlates with sexual dichromatism. Males are not more conspicuous than females against a vegetated background but are significantly more conspicuous against a nesting background than females. Yet, we found no correlation between conspicuousness and degree of sexual dichromatism. Our results suggest that, while both natural and sexual selection can drive color differences between the sexes, levels of sexual dichromatism do not necessarily reflect the intensity of selection forces in this clade. Our results highlight the importance of testing assumptions regarding the relationship between sexual dichromatism and color evolution in each sex, considering the properties of different visual systems, relevant to the ecology of the study model.