Abstract. In Queensland, Australia, in response to isolated cases of dengue infection, larval control of the vector Aedes aegypti (L.) (Diptera: Culicidae) is targeted at breeding sites within 200 m of a case and interior spraying with a pyrethroid adulticide is targeted at premises within 100 m. To ascertain whether these limits are appropriate, we conducted a mark–release–recapture study to measure the dispersal of female Ae. aegypti in the city of Cairns where transmission occurs. Female mosquitoes reared from wild collected eggs were differentially marked with fluorescent dust depending on whether they were to be released blood-fed or non-blood-fed, and a total of 1948 females was released. A total of 132 sticky ovitraps was set at 64 premises within a 200 m radius and collections of trapped adults were made at 5–15 days post-release. Sixty-seven females (3.4%) were recaptured, with the furthest being caught 200 m from the release point, and the mean distance travelled was 78 m. Overall, 23.1% of the recaptures outside the release site were taken beyond 100 m by day 15. Dispersal was comparable for both blood-fed and non-blood-fed releases. There was a significant tendency for dispersal to be in a north-westerly direction, probably because of the presence of numerous containers and heavy shading by trees in this direction and a busy road to the south of the release point that appeared to inhibit dispersal. The results suggest that adulticiding may have to be extended beyond 100 m if more than 8 days have elapsed since female Ae. aegypti could have fed upon a viraemic dengue case. The study also shows that dispersal is not random, and that it may be possible to maximize vector control by taking into account environmental factors that affect the direction of female mosquito flight.