Aim Recently, Choudhury & Dick (2000) argued that helminth parasite communities in tropical freshwater fish are not richer than those in temperate fish. Here their data is re-examined in an analysis that takes into account important confounding variables, and their conclusions are revisited.Location The data set covers fish species from many tropical areas and from the north temperate area of the nearctic zone.Methods Multivariate analyses are used to control the influence of sampling effort and host body size, two factors that may confound the comparison between tropical and temperate helminth communities. In addition, phylogenetically independent contrasts are derived from the fish taxa included in the data set and used to remove the potential effect of shared phylogenetic history on the richness of helminth communities.Results Both host body size and the number of hosts examined correlated positively with the number of species found in a helminth community, and fish body size also correlated with measures of richness corresponding to the mean number of helminth species per fish individual. After controlling for the effects of these two confounding variables, temperate helminth communities still had higher richness scores than those from the tropics. The analysis using independent contrasts removed the remaining potential confounding factor, i.e. host phylogeny, and it again showed that helminth communities of temperate fish taxa are richer than those of related fish taxa from the tropics. This last result, however, applies only to the number of helminth species found in a helminth community, and not to the mean number of helminth species per individual fish host.Main conclusions The general conclusion of the present study is similar to that of Choudhury & Dick (2000). However, the new analysis has ruled out some confounding factors as potential explanations for the patterns, and has served as an indirect test of a version of the area hypothesis (i.e. helminth diversity is explained by host size). The reason why this latitudinal gradient in richness runs counter to those usually observed in other plant and animal assemblages remains to be found.