The consequences of non-homogeneous mixing (non-random contact) between vectors and hosts for the persistence, prevalence, and hence control, of any mosquito-borne disease are explored. When mosquitoes concentrate on certain hosts, the basic reproductive rate, R, of the disease (a measure of persistence) and the vectorial capacity are both greater than or equal to their values under homogeneous mixing. Field data suggest that R could be more than two and a half times as large as it would be under homogeneous mixing. Our calculations underestimate the importance of heterogeneity by ignoring mosquito patchiness and stochastic effects. Host selection limits the dependability of predictions, made from population models which assume homogeneous mixing, about the success of disease control (vaccines, chemotherapy, vector control). In particular, eradication or the maintenance of low prevalence will be more difficult than expected, unless localized control makes the distribution of infective bites on a community more nearly uniform. Age may explain most of the variability in biting in small communities (e.g., rural villages), in which case models incorporating age-specific biting will be appropriate tools in control programmes.