When the currently used larval surveillance system (visual inspection) for the dengue vector Aedes aegypti was compared with the surveillance for the presence of eggs by ovitrapping in Port of Spain, Trinidad, it was found that the latter (39.1%) was significantly more sensitive than the visual inspection system (10.1%). At the same time, the presence of the nuisance mosquito Culex quinquefasciatus was detected in 38.4% of the households. Both Ae. aegypti and Cx. quinquefasciatus showed preference for ovipositional attractants in ovitraps: hay infusion > yeast suspension > plain tap water. Although all the socioeconomic and geographic areas produced both mosquito species in 1996, upper middle class (UMC) areas (8.6-43.4%) produced more Ae. aegypti than did lower middle class (LMC) areas (7.8-38.8%), which produced more than working class (WC) areas (3.9-29.9%). For Cx. quinquefasciatus, the order of production was reversed with WC areas (50.1%) > LMC areas (30.0%) > UMC areas (26.0%). Change in vector surveillance strategies incorporating some ovitrapping and stratified sampling are recommended for Caribbean countries.