In gynodioecious species, females and hermaphrodites coexist and the genetics of sex determination is usually nuclear cytoplasmic. Maintaining nuclear-cytoplasmic gynodioecy requires polymorphism for the feminizing genes (contained in the mitochondria) and the genes that restore male fertility (contained in the nucleus). This complex polymorphism depends, in part, on there being negative pleiotropic effects (i.e. costs) of the nuclear restorer alleles. Here, we combine information from theoretical studies and studies on the molecular action of restorer alleles in crops to interpret the probable costs of such alleles, and suggest how various aspects of the theoretical models could be tested. In doing so, we highlight how crops can be used to address evolutionary questions about the maintenance of nuclear-cytoplasmic gynodioecy.