1. Classic evolutionary theory suggests that sexual dimorphism evolves primarily via sexual and fecundity selection. However, theory and evidence is beginning to accumulate suggesting that resource competition can drive the evolution of sexual dimorphism, via ecological character displacement between sexes. A key prediction of this hypothesis is that the extent of ecological divergence between sexes will be associated with the extent of sexual dimorphism.
2. As the stable isotope ratios of animal tissues provide a quantitative measure of various aspects of ecology, we carried out a meta-analysis examining associations between the extent of isotopic divergence between sexes and the extent of body size dimorphism. Our models demonstrate that large amounts of between-study variation in isotopic (ecological) divergence between sexes is non-random and may be associated with the traits of study subjects. We therefore completed meta-regressions to examine whether the extent of isotopic divergence between sexes is associated with the extent of sexual size dimorphism.
3. We found modest but significantly positive associations across species between size dimorphism and ecological differences between sexes, that increased in strength when the ecological opportunity for dietary divergence between sexes was greatest.
4. Our results therefore provide further evidence that ecologically mediated selection, not directly related to reproduction, can contribute to the evolution of sexual dimorphism.

We collated peer-reviewed literature available in the Web of Science Core Collection. The stable isotope literature is large, with the search term “stable isotope” returning ~76 500 studies at the time of writing. To constrain the search, we combined the following specific terms, using the default publication year range of 1900-2020, on 10/11/2020: Isotop* Nich; Isotop Nich* Male; Isotop* Nich* Female; Isotop* Nich* Male Female; Isotop* Nich* Sex Diff*; Isotop Nich* Dimorph; Isotop Dimorph*.

Our searches returned 3489 studies, which we placed into a spreadsheet to highlight duplicates for manual removal. Removing duplicates resulted in 2807 studies for title and abstract screening. At this stage, we made the decision to constrain our analysis to the nitrogen and carbon stable isotope systems, due to the relatively small number of studies using other systems that were returned by our search terms. We also rejected studies during title and abstract screening if they did not use bulk stable isotope analysis, used samples of human, museum, archeological or palaeontological origin, were review, comment, or method papers, or if the animals sampled were not wild, not adults, not vertebrates or if data were not available for both sexes. We then searched the remaining 1279 studies using the ctrl+F search function and, separately, the terms “sex”, “male” and “female”, excluding studies if they contained none of these terms, under the assumption that they did not contain stable isotope ratios for each sex and, if at least one term was present, checking for the presence of the required data. Additional reasons for exclusion were if the full text was inaccessible without purchase or contacting authors, presented incomplete data (mean, error or sample size missing), was not in English or was a paper correction. We then attempted to extract data from the remaining 210 studies.  Additional reasons for exclusion at this stage were if raw data was presented as images with >50 rows, if data were from an earlier study already included or if data extraction from figures was not possible. We extracted data from figures using a mouse pointer to individually select data points from an image of the figure, with the image calibrated to the axis values from the original figure; therefore, too much point overlap made this process inaccurate, because not all points could be selected for inclusion. The entire process provided 173 studies in which mean, standard deviation and sample sizes for each sex were presented in the manuscript, or could be calculated from raw data, or could be taken from model outputs, or extracted from figures. We collected data for any vertebrate species, from any global location and, if stable isotope ratios for each sex were presented for more than one tissue type, we entered each tissue as a separate row in our database.

Miscrosoft Excel