Five cases of clinical Q fever were identified amongst students and staff of a Somerset secondary school between 23 October 1987 and 21 December 1987. Five goats which were kept at the school were found to have antibodies to Coxiella burnetii phase II. A cross-sectional study was conducted at the school in July 1988. A single CF test was used to identify evidence of prior infection, and a self-administered questionnaire was used to collect data on exposure variables and illness during 1987. Four hundred and ninety-nine eligible subjects took part in the study, and serological information was obtained from 461 of these. Eighty-seven subjects (18.9%) had CF titres of 20 or greater. It was estimated that only 1 in every 30 individuals with evidence of past C. burnetii infection had been recognized as a clinical case of Q fever. Antibody positive subjects were more likely to have been off school sick and to report having suffered chest pain than negative subjects. Contact with school animals, specifically cleaning the school poultry, collecting their eggs and visiting a school goat on the day of kidding, was associated with the presence of antibodies to C. burnetii. However, a large proportion of the antibody positives (24.2%) had no known contact with the school animals. Spread of organisms, either wind-borne or in straw or manure, may have been responsible for the high prevalence of unexplained infection.