Marked Aedes aegypti (L.) (5-6 d old) were released inside 2 groups of 5 houses (100 females per house) in a residential community in Florida, PR, to compare behavior of gravid females at sites where oviposition containers were absent to sites where containers were abundant (i.e., 2 tires and 10 ovipots were added to each yard). Two sequential releases were made so that both groups of houses were evaluated with oviposition containers removed and added. Mosquitoes resting inside the 10 release houses plus 20 additional neighboring houses were collected with backpack aspirators for 4 consecutive days, beginning 2 d after release. Because 172 of the 185 recaptured females (93%) were collected in the same houses in which they had been released, dispersal patterns were not directly comparable. However, the recapture rate in houses with containers added (13%) was significantly higher (P < 0.0001) than in houses with containers removed (6%). No difference was observed in the mean number of potential oviposition containers among the nonrelease houses at the 2 sites (3.9 versus 3.8 aquatic containers per house in the prerelease survey). Our results are consistent with the hypothesis that availability of oviposition sites is inversely correlated with the potential for female Ae. aegypti to disperse. These results have important implications because campaigns to reduce Ae. aegypti larval sites during dengue epidemics could have the undesirable effect of inducing the dispersal of infected adult female mosquitoes.