- SUMMARY We investigated variation in the abundance-prevalence relationships of fleas among 17 different flea-host associations as well as among different species of hosts and fleas in the Negev desert. We explored variation in the value of exponent of Taylor’s power relationship with changes in flea community size and flea specialization (host specificity and seasonal pattern of activity). We tested if a simple epidemiological model can reproduce the pattern of the abundance-prevalence relationship. We confirmed aggregated distribution of fleas within a population of host species as well as across a whole host community and the existence of a positive relationship between local flea abundance and their prevalence. Prevalence, mean abundance and variance of abundance were significantly higher in host specific than host opportunistic fleas. When ecological specialization was considered, based on a seasonal pattern of activity, these parameters were higher in yearround-active than seasonal fleas. The degree of flea specialization and flea community richness affected the pattern of the relationship between mean abundance and its variance. Power law slopes decreased with increasing richness of flea community. A simple epidemiological model based on mean flea abundance and degree of aggregation, corrected for host sample size, can predict the observed pattern of prevalence. In some cases, observed flea prevalence was higher than that predicted from the epidemiological model. The discrepancy of the observed prevalence from that predicted by the model can be explained by either a relatively low negative effect of flea parasitism on a host (at least, in terms of pathology) or strong resistance of a host to flea parasitism or both.