Surveillance for Ebola hemorrhagic fever was conducted in the Democratic Republic of the Congo from 1981 to 1985 to estimate the incidence of human infection. Persons who met the criteria of one of three different case definitions were clinically evaluated, and blood was obtained for antibody confirmation by IFA. Contacts of each case and 4 age- and sex-matched controls were also clinically examined and tested for immunofluorescent antibody. Twenty-one cases of Ebola infection (persons with an antibody titer of > or = 1:64, or lower if they fit the clinical case definition) were identified, with a maximum 1-year incidence of 9 and a case fatality rate of 43%. Cases occurred throughout the year, but most (48%) occurred early in the rainy season. Fifteen percent of contacts had antibody titers > or =1:64 to Ebola virus, compared with 1% of controls (P < .0001). Results suggest that Ebola virus periodically emerges from nature to infect humans, that person-to-person transmission is relatively limited, and that amplification to large epidemics is unusual.