A follow-up study to 52 weeks is reported on 95 corticosteroid-dependent asthmatics who took part in a double-blind, controlled trial of beclomethasone dipropionate for 28 weeks. During the 28 weeks two dosage schedules (800 microgram and 400 microgram daily) were compared with a placebo. If the corticosteroid tablet dosage had been at least halved and the clinical condition was satisfactory, the allocated regimen was continued 'blind' for the second period (29-52 weeks); 51 patients continued for this period, 46 on beclomethasone dipropionate. If progress was unsatisfactory, known beclomethasone dipropionate 800 microgram daily was substituted, the supervision remaining unchanged, the clinical supervisors remaining unaware which schedule had been received during the first 28 weeks. Nearly all the patients who at least halved their corticosteroid tablet dosage by 28 weeks continued to do well up to 52 weeks on the allocated regimen. When the placebo patients with unsatisfactory progress at 28 weeks were changed to 800 microgram beclomethasone dipropionate they showed a substantial improvement. Patients changed from 400 to 800 microgram showed little improvement on average. Approximately 30% of the whole group of patients failed to have their dosage of corticosteroid tablets, despite receiving 800 microgram beclomethasone dipropionate daily for six or more months. The cumulative incidence of 'oral candidiasis' (as defined for this report) in placebo patients changed at 28 weeks to 800 microgram beclomethasone dipropionate daily was 30% at 36 weeks and 43% at 52 weeks. Of eight patients increased from 400 microgram to 800 microgram daily, one had 'candidiasis' by 36 weeks and three by 52 weeks. Oral candidiasis was often asymptomatic and never led to the discontinuation of the corticosteroid aerosol, although the dosage was reduced in nine of the 30 who received 800 microgram beclomethasone dipropionate during the first 28 weeks. The incidence of bacterial infections in patients on beclomethasone dipropionate was similar to that of the placebo group and systematic laboratory investigations revealed no increase in the rate of isolation of potential pathogens.