We used data on the abundance and distribution of fleas parasitic on small mammals in Slovakia and aimed: (i) to confirm a positive relationship between abundance and distribution fleas within and across host species; and (ii) to test if prevalence of fleas can be reliably predicted from a simple epidemiological model that takes into account flea mean abundance and its variance. Prevalence of a flea species increased with an increase in its mean abundance both within and across host species. We calculated prevalences both for each flea-host association and for each flea species across all hosts. Observed prevalences did not differ significantly from those predicted by the epidemiological model using parameters of Taylor's power relationship between mean abundance of fleas and its variance. Regressions of predicted prevalences against observed prevalences produced slope values that did not differ significantly from unity and were independent of scale (within or across host species). Our results demonstrated that up to 96% of variance in flea prevalence can be explained solely by their mean abundance. We concluded that, in general, there is no need to invoke other, more complex factors for the explanation of the variation in flea prevalence.