A core hypothesis in coevolutionary theory proposes that parasites adapt to specifically infect common host genotypes. Under this hypothesis, parasites function as agents of negative frequency-dependent selection, favouring rare host genotypes. This parasite-mediated advantage of rarity is key to the idea that parasites maintain genetic variation and select for outcrossing in host populations. Here, we report the results of an experimental test of parasite adaptation to common versus rare host genotypes. We selected the bacterial parasite
Serratia marcescens
to kill
Caenorhabdiis elegans
hosts in uneven mixtures of host genotypes. To examine the effect of commonness itself, independent of host identity, each of four host genotypes was represented as common or rare in experimental host mixtures. After experimental selection, we evaluated a parasite line's change in virulence—the selected fitness trait—on its rare and common host genotypes. Our results were consistent with a slight advantage for rare host genotypes: on average, parasites lost virulence against rare genotypes but not against common genotypes. The response varied substantially, however, with distinct patterns across host genotype mixtures. These findings support the potential for parasites to impose negative frequency-dependent selection, while emphasizing that the cost of being common may vary with host genotype.