Several papers have suggested that host species with larger geographic ranges tend to harbour a greater number of parasite species than those with smaller ranges. However, few of these investigations have dealt quantitatively with problems related to host sampling and/or phylogeny. This paper describes a positive relationship between the number of helminth parasite species per host and host species geographic range in Holarctic waterfowl. The analysis controls for the effects of differential sampling of host species and for the effects of taxonomic association. Neither host body size, population size, population density, nor social tendency correlate significantly with the number of parasite species per host corrected for number of hosts examined, and therefore association with these variables is unlikely to confound the relationship between parasites and host geographic range. The analysis illustrates the strong correlation between number of parasite species recorded and sampling effort regardless of the methodology used. The number of host individuals examined is clearly the most useful measure of sampling effort in such data sets. The discussion outlines the implications the present findings may have in relation to studies of other taxa. Assessment of parasite richness may depend critically upon the sample size for each host species. Unless sampling effort is controlled for, apparent relationships between parasites and biological variables (such as geographic range) may be spurious.