Organisms that are abundant locally in a habitat patch are commonly observed to be frequent regionally, or among patches. In parasites, species present in high numbers in host individuals are also present in many individuals in the host population. On a larger scale, however, when host species are considered as patches, we may expect the opposite pattern because of the cost of producing mechanisms to evade the immune responses of several host species. Thus parasite species exploiting many host species may achieve lower average abundance in their hosts than parasite species exploiting fewer host species. This prediction was tested with data from 188 species of metazoan parasites of freshwater fish, using a comparative approach that controlled for study effort and phylogenetic influences. A negative correlation was found between the number of host species used by parasites and their average abundance in hosts, measured as either prevalence or intensity of infection. There was no evidence that parasite species fall into distinct categories based on abundance patterns, but rather that they fall along a continuum ranging from a generally low abundance in many host species, to a generally high abundance in few host species. These results applied to both ecto- and endoparasites. The pattern observed suggests the existence of a trade-off between how many host species a parasite can exploit and how well it does on average in those hosts.