OBJECTIVE--To examine the efficacy and safety of conservative management of mild otitis media ("the acute red ear") in children. DESIGN--Double blind placebo controlled trial. SETTING--17 group general practices (48 general practitioners) in Southampton, Bristol, and Portsmouth. PATIENTS--232 children aged 3-10 years with acute earache and at least one abnormal eardrum (114 allocated to receive antibiotic, 118 placebo). INTERVENTIONS--Amoxycillin 125 mg three times a day for seven days or matching placebo; 100 ml paracetamol 120 mg/5 ml. MAIN OUTCOME MEASURES--Diary records of pain and crying, use of analgesic, eardrum signs, failure of treatment, tympanometry at one and three months, recurrence rate, and ear, nose, and throat referral rate over one year. RESULTS--Treatment failure was eight times more likely in the placebo than the antibiotic group (14.4% v 1.7%, odds ratio 8.21, 95% confidence interval 1.94 to 34.7). Children in the placebo group showed a significantly higher incidence of fever on the day after entry (20% v 8%, p less than 0.05), mean analgesic consumption (0.36 ml/h v 0.21 ml/h, difference 0.14, 95% confidence interval 0.07 to 0.23; p = 0.0022), mean duration of crying (1.44 days v 0.50 days, 0.94; 0.50 to 1.38; p less than 0.001), and mean absence from school (1.96 days v 0.52 days, 1.45; 0.46 to 2.42; p = 0.0132). Differences in recorded pain were not significant. The prevalence of middle ear effusion at one or three months, as defined by tympanometry, was not significantly different, nor was there any difference in recurrence rate or in ear, nose, and throat referral rate in the follow up year. No characteristics could be identified which predicted an adverse outcome. CONCLUSIONS--Use of antibiotic improves short term outcome substantially and therefore continues to be an appropriate management policy.