• condition dependence;
  • female choice;
  • mate choice;
  • sexual selection;
  • sperm competition;
  • sperm count;
  • sperm motility;
  • sperm length;
  • sperm velocity;
  • sperm viability


There are two reasons why researchers are interested in the phenotypic relationship between the expression of male secondary sexual characters (SSCs) and ‘ejaculate quality’ (defined as sperm/ejaculate traits that are widely assumed to increase female fertility and/or sperm competitiveness). First, if the relationship is positive then females could gain a direct benefit by choosing more attractive males for fertility assurance reasons (‘the phenotype-linked fertility’ hypothesis). Second, there is much interest in the direction of the correlation between traits favoured by pre-copulatory sexual selection (i.e. affecting mating success) and those favoured by post-copulatory sexual selection (i.e. increasing sperm competitiveness). If the relationship is negative this could lead to the two forms of selection counteracting each other. Theory predicts that the direction of the relationship could be either positive or negative depending on the underlying genetic variance and covariance in each trait, the extent of variation among males in condition (resources available to allocate to reproductive traits), and variation among males in the cost or rate of mating. We conducted a meta-analysis to determine the average relationship between the expression of behavioural and morphological male secondary sexual characters and four assays of ejaculate quality (sperm number, viability, swimming speed and size). Regardless of how the data were partitioned the mean relationship was consistently positive, but always statistically non-significant. The only exception was that secondary sexual character expression was weakly but significantly positively correlated with sperm viability (r = 0.07, P < 0.05). There was no significant difference in the strength or direction of the relationship between behavioural and morphological SSCs, nor among relationships using the four ejaculate quality assays. The implications of our findings are discussed.