A phylogenetically informed comparative analysis of sexual testosterone dimorphism across mammals in relation to paternal care and sexual size dimorphism
Laubi, B. N.; Burkart, J. M.; Willems, E. P.; van Schaik, C. P.
Show abstract
Within species, male testosterone is often linked to mating competition and paternal care, suggesting that sex differences in endogenous testosterone values across mammals may covary with broader reproductive strategies. Using a structured literature search, we compiled 63 studies, spanning 31 non-human species and 9 human populations, reporting endogenous, non-experimentally manipulated testosterone values for both adult males and females within the same population and context. From these studies, we calculated male-to-female testosterone ratios, and analysed these data using Bayesian phylogenetic multilevel models. We tested whether testosterone dimorphism was associated with paternal care and sexual size dimorphism while accounting for sampling matrix, assay method, breeding context, and wild versus captive setting. Across non-human mammals, neither paternal care nor sexual size dimorphism (indexing competition) showed a clear association with testosterone ratios, and the same pattern emerged in the primate-only subset. By contrast, sampling matrix was consistently associated with testosterone dimorphism across all analyses, with lower male-to-female ratios in non-blood than in blood-based measures. In primates, testosterone ratios were also lower in captive than in wild populations, although this pattern was not clearly supported in the broader non-human dataset. In the human-only analysis, testosterone ratios did not clearly differ between industrialized and small-scale societies, whereas the matrix effect remained evident. Overall, our results suggest that sampling matrix is a major source of variation even for ratio-based measures, highlighting the need for caution when inferring between-species endocrine differences from studies using different substrates. More broadly, directly comparable, non-experimentally manipulated testosterone data for both sexes remain rare across mammals, limiting comparative inference.
Matching journals
The top 8 journals account for 50% of the predicted probability mass.