Sperm competition may occur whenever sperm from more than one male are present in the reproductive tract of the female. Studies of field-caught Drosophila reveal that a substantial fraction (80%) of females clearly have sperm from more than one male, and the figure is probably higher because only a small number of progeny are typically surveyed, so a strong skew in paternity can make multiply-mated females appear as singly mated unless appropriate models are applied. Examination of genetic variation in aspects of sperm competition has revealed some striking patterns, particularly in the implications for the maintenance of polymorphism. The magnitude of variation in sperm competitive ability is as great as that for other fitness components, and the males with the strongest displacement also appear to be the ones with the greatest positive effect on fertility. Why then does not the most competitive allele simply go to fixation? Such synergistic pleiotropy makes the polymorphism even more unexpected. Examination of patterns of competitive success of pairs of male genotypes, and of female-male interactions, demonstrate clearly that the outcome of sperm competition is not a simple property of each male. That is, sperm competitive ability of male genotypes cannot simply be ranked from best to worst. Rather, the outcome of each competitive bout depends on the particular pair of males. These results have intriguing implications for the molecular biology of genes involved in the determination of sperm competitive success, and on the opportunity for maintenance of polymorphism in those genes.