829 healthy male volunteers aged 20-49 were issued with supplies of either demethylchlortetracycline or placebo in late October, 1969, with instructions to take a course of treatment in the event of getting a cold or influenza-like illness between November, 1969, and April, 1970. Clinical and other information was self-recorded on reply-paid postcards, one card for each month. 91% of cards were returned, and 836 courses of treatment were taken by 543 of the patients. Morbidity was assessed in terms of the number of days in each illness on which purulent spit was noted and the number of days off work. Closely similar patterns of morbidity were found for patients taking antibiotic and placebo; whether treatment was started early in the illness, before or after onset of cough, purulent nasal discharge or purulent spit had no effect. Being a non-smoker rather than a 20-plus cigarettes a day smoker was of greater benefit than taking the antibiotic rather than the placebo. Work loss averaged 1·1 days per illness in patients on antibiotic and 1·5 days per illness in those on placebo. Medical advice was sought in 109 illnesses in which the trial tablets had been used. The need for insurance certificates seemed to be a more prominent reason for attendance than did morbidity referable to absence of antibiotic treatment. Side-effects were recorded in 6·7% of illnesses on antibiotic treatment and 1·8% of illnesses on placebo treatment. No justification has thus been found to support the early use of a tetracycline for minor respiratory illness in otherwise healthy patients.