During experimental hut trials to assess the efficacy of insecticide-treated bednets against malaria mosquitoes, we observed that human subjects varied consistently in their attractiveness to mosquitoes. Attractiveness was assessed by estimating the numbers of wild Anopheles gambiae Giles mosquitoes entering a hut in which a man was sleeping, and the numbers of human-bloodfed An. gambiae sensu lato collected from each hut each morning. Five trials were carried out at Wali Kunda in rural Gambia during 2.5 yr. During each 6-wk trial a man slept under a bednet in each of the six huts. Morning collections of mosquitoes from the room, enclosed verandas, and window traps of each hut provided estimates of the number of mosquitoes that had entered during the night. Blood meals were analyzed using an ELISA technique to identify those mosquitoes feeding on humans. Specimens were collected by field workers, not the subjects; therefore, sampling was independent of the subjects' ability to catch mosquitoes. Moreover, the trials were designed to measure the relative attractiveness of individual sleepers to mosquitoes, allowing for other sources of variation (i.e., among huts, bednets, nights, and day of the week). Attractiveness of men to mosquitoes differed significantly among individuals as indicated by the consistent differences between the numbers of mosquitoes entering each man's hut and the numbers feeding on each man. However, the two measures of attractiveness were apparently independent of each other: subjects who attracted consistently high numbers of vectors into their hut did not necessarily have high numbers of mosquitoes feeding on them. These findings support the view that some individuals within a community are at greater risk from mosquito-borne pathogens than others.