In a population of approximately 300 000 there were 150 illnesses attributable to chlamydial infection from January 1975 to June 1983 and nearly all were presumed due to Chlamydia psittaci. Diagnosis was based on the complement-fixation (CF) test, the disadvantages of which are discussed, especially anamnestic responses. There were 73 cases of pneumonia, 37 febrile respiratory infections, 35 miscellaneous conditions and five asymptomatic seropositive contacts. The annual number of cases was lowest in 1976 (3) and increased to a maximum in 1980 (28) followed by a sustained high level. No seasonal variation was observed. A presumptive bird source was present in only 17% and detailed consideration of cases in one rural practice suggests that the infection may possibly be more endemic than zoonotic.