Two-hundred and thirty patients with a diagnosis of presumptive bacterial conjunctivitis were assessed in a randomized double-blind multicentre trial. In two of the centres the patients had been treated with either trimethoprim-polymyxin B or neomycin-polymyxin B-gramicidin ophthalmic solution. In the other two centres the patients had been treated with either trimethoprim-polymyxin B or chloramphenicol ophthalmic solution. All of the preparations used were shown to be effective and very few adverse reactions were encountered. No significant difference in clinical efficacy could be demonstrated between trimethoprim-polymyxin B and neomycin-polymyxin B -gramicidin but trimethoprim-polymyxin B was found to be significantly better (P = 0.03) than chloramphenicol in reducing signs and symptoms.