Objectives-To determine the occupational risk of Q fever. Design-Cohort study. Setting-Community: five English local authority districts. Subjects and methods-Prevalence and incidence of immunoglobulin G (IgG) specific antibody to Coxiella burnetii phase II antigen was measured in a representative (study) cohort of farm workers in the United Kingdom, and detailed exposure data were collected. Also seroprevalence of Q fever in a (control) cohort of police and emergency service personnel was measured.Results-Prevalence was significantly (P < 0.01) higher in the study cohort (1051385 v 43/395). During the first 12 month period after enrolment no seroconversions were found (upper 95% confidence limit: 13181100 0OOOyear). During the second 12 month period after enrolment two seroconversions were found, equalling an incidence of 813/100 000, year (95% confidence interval (95% CI) 98-2937/100 0OOOyear). No association was found between seroprevalence and age. In the study cohort, extent of total contact with farm animals seemed more important than exposure to any specific animal: full time employees were more than four times more likely to be antibody positive than part time employees (P < 0.05). Exposure to cattle, but not sheep, goats, cats, raw milk, and hay (all reported sources of Q fever) was associated with being positive to Coxiella burnetii IgG by univariate analysis but this association was not independent of total farm animal contact. Conclusions-The risk of Q fever on livestock farms is related to contact with the farm environment rather than any specific animal exposure. The absence of an increasing prevalence with age suggests that exposure may occur as clusters in space and time (outbreaks). (Occup Environ Med 1995;52:644-647)