A 28-day treatment regimen was undertaken by 12 volunteers, in which 5% (w/v) benzoyl peroxide (BP) in an aqueous gel was applied daily to the entire face. Clinical efficacy of the treatment was assessed after 2, 4, 9, 14 and 28 days, and the surface and follicular microbial populations were enumerated using established techniques. Viable counts were obtained for propionibacteria and Micrococcaceae. Mean numbers of propionibacteria recovered from the skin surface and follicular casts were significantly reduced after 2 days' treatment (P < 0.01), and the population was maintained at a significantly lower level throughout the study (P < 0.01), with mean values approaching the lower detection limit of the assay. Significant reductions in the surface and follicular Micrococcaceae were observed after 2 days' treatment, and at all subsequent visits (P < 0.05). After 2 days' treatment, only slight reductions in mean acne grade and mean inflamed lesion count were observed. However, at all subsequent visits the mean acne grade was significantly reduced (P < 0.05) compared with T0. The mean non-inflamed lesion count was lower than the pretreatment level at all visits, although the results were variable. The results indicate that significant reductions in surface and follicular microorganisms may be obtained after 48 h treatment with BP. Therefore, the non-specific antibacterial action of BP may be utilized in short intervening courses to reduce the carriage of antibiotic-resistant micro-organisms and thus improve the long-term efficacy of antibiotic acne treatments.(ABSTRACT TRUNCATED AT 250 WORDS)