Small scale spatial variation and temporal heterogeneity in mosquito densities can have important consequences for disease transmission, but the extreme variation which is observed in populations of malaria vectors makes it difficult to obtain good predictions of densities for short time periods over limited areas. We have applied Bayesian techniques derived for use in cancer epidemiology in order to map densities of Anopheles gambiae s.l. and A. funestus in a Tanzanian village where there is intense transmission of Plasmodium falciparum malaria. Estimates derived in this way should prove useful in vector population biology and in improving estimates of exposure-response relationships of the human host to malaria. The same methods can be applied in other fields of animal ecology.