Abstract
Many epidemic models lead to an approximately Poisson distribution of parasites among hosts. This is at variance with observation, where heavy previous overdispersion is the rule. A simple model is proposed that, while treating individuals alike, nonetheless give rise to highly variable parasite loads.