New estimates indicate that males are not larger than females in most mammal species

Sexual size dimorphism has motivated a large body of research on mammalian mating strategies and sexual selection. Despite some contrary evidence, the narrative that larger males are the norm in mammals – upheld since Darwin’s Descent of Man – still dominates today, supported by meta-analyses that use coarse measures of dimorphism and taxonomically biased sampling. With newly available datasets and primary sources reporting sex-segregated means and variances in adult body mass, we estimated statistically determined rates of sexual size dimorphism in mammals, sampling taxa by their species richness at the family level. Our analyses of wild, non-provisioned populations representing >400 species indicate that although males tend to be larger than females when dimorphism occurs, males are not larger in most mammal species, suggesting a need to revisit other assumptions in sexual selection research.

Methods
These data on sex-segregated body mass means and variances were collected from the literature, including existing published datasets and from primary sources. We sampled mammalian families according to their species richness (including the orders and families comprising at least 10 species). We excluded data from sexually immature, captive-bred or food-provisioned, or pregnant individuals when these data were distinguished or when it was indicated that they were mixed into the data (see Methods section in manuscript for the few exceptions). Estimates from museum specimens were generally not used as we wished to make within-population comparisons between the sexes, and domesticated species were excluded apart from semi-domesticated free-ranging reindeer (Rangifer tarandus). Whenever sex-segregated means and variances for body length were presented alongside body mass data in a study included in the dataset, these data were also harvested. Prior to analyses, we filtered the full dataset (attached here) by sample size such that a minimum of 9 males and 9 females were required for inclusion, based on an analysis of the relationship between sample size and 95% confidence intervals for mean body mass (see R code), and removed rows for any duplicated species, keeping only the row with the highest combined (male + female) sample size for each unique species (subspecies were treated as different populations of the same species). Analyses were performed in RStudio (version 4.3.2) and the code used for analyses and data visualization are also appended here.