Sexual conflict, where male and female reproductive interests differ, is probably widespread and often mediated by male or sperm proteins and female or egg proteins that bind to each other during mating or fertilization. One potential consequence is maintenance of polymorphism in these proteins, which might result in reproductive isolation between sympatric subpopulations. I investigate the conditions for polymorphism maintenance in a series of mathematical models of sexual conflict over mating or fertilization frequency. The models represent a male or sperm ligand and a female or egg receptor, and they differ in whether expression of either protein is haploid or diploid. For diploid expression, the conditions imply that patterns of dominance, which involve neither overdominance nor underdominance, can determine whether polymorphism is maintained. For example, suppose ligand expression is diploid, and consider ligand alleles L₁ and L₂ in interactions with a given receptor genotype; if L₁/L₁ males are fitter than L₂/L₂ males in these interactions, then polymorphism is more likely to be maintained when L₁/L₂ males more closely resemble L₁/L₁ males in these interactions. Such fitter-allele dominance might be typical of a ligand or its receptor due to their biochemistry, in which case polymorphism might be typical of the pair.