In many gynodioecious species the nuclear inheritance of male fertility is complex and involves multiple (restorer) genes. In addition to restoring plants from the female (male sterile) to the hermaphrodite (male fertile) state, these genes are also thought to play a role in the determination of the quantity of pollen produced by hermaphrodites. The more restorer alleles a hermaphroditic plant possesses, the higher the pollen production. To test this hypothesis I combined the results of crossing studies of the genetics of male sterility with phenotypic data on investment in stamens and ovules among the progeny of plants involved in these studies. The sex ratio (i.e. the frequency of hermaphrodites among the progeny), being a measure of the number of restorer alleles of the maternal plant, was positively related to the investment in pollen (male function), but negatively related to the investment in ovules (female function), in both field and greenhouse experiments. Consequently, a negative correlation between male and female function was observed (trade–off) and it is suggested that antagonistic pleiotropic effects of restorer genes might be the cause. Phenotypic gender, a measure combining investment in both pollen and ovules, was highly repeatable between field and greenhouse, indicating genetic determination of a more male– or female–biased allocation pattern among the studied plants.