Abstract:How and why individuals differ from each other is a central question in behavioral and evolutionary ecology, because selection particularly acts on this among-individual variation. It is therefore important to accurately partition phenotypic variances into their within- and between-individual components. Partitioning covariances into both components can also inform about underlying mechanistic pathways that potentially interlink trait expression. In the current study, we applied such a (co)variance partitioning approach to test key predictions of two central hypotheses in behavioral ecology, namely the immunocompetence handicap hypothesis and the challenge hypothesis. To this end, we assessed potential pleiotropic effects of testosterone on male sexual signaling, immune function, and parental care. We here repeatedly measured a set of relevant traits in 47 breeding pairs of captive canaries (Serinus canaria). We found that a within-individual increase in female testosterone level suppressed immune function. Furthermore, testosterone levels were positively related to male song repertoire size as an important component of sexual signaling at the between-male level. These were, however, the only relevant significant correlations. Overall, our data do therefore not convincingly support the hypotheses tested and suggest rather limited hormonal pleiotropic effects of testosterone on immune function, parental care, and male sexual signaling, at least in our study system.Significance statementPhenotypic variances and covariances can nowadays be partitioned easily into within- and between-individual components. These components inform about trait repeatability and the extent to which multiple traits form phenotypic suites of traits, as well as about their joint underlying mechanistic pathways. Testosterone for example, may be key to mediate the expression of suites of traits in many vertebrate species. We here used captive canaries and repeatedly measured male and female traits, relevant in the context of endocrinology, sexual signaling, immune function, and parental investment. For both sexes, we report particularly weak correlations between testosterone and all other measured traits at both within- and between-individual levels. Our case study thus questions the pleiotropic effects of testosterone, but exemplifies the applicability and relevance of (co)variance partitioning in behavioral ecological research.

Within- and between-individual (co)variance partitioning reveals limited pleiotropic effects of testosterone on immune function, sexual signaling, and parental investment