In Transkei, 143 patients with active tuberculous constrictive pericarditis without significant pericardial effusion all received the same daily 6-month antituberculosis regimen of streptomycin, isoniazid, rifampicin, and pyrazinamide for 14 weeks followed by isoniazid and rifampicin. They were randomly allocated to receive in addition either prednisolone or placebo for the first 11 weeks; the comparison was double-blind throughout treatment and follow-up. In the 114 patients assessable up to 24 months, improvement was significantly more rapid in the prednisolone group, as shown by the rate of fall in the mean pulse rate and the rate at which jugular venous pressure and level of physical activity became normal. During follow-up, 2 (4%) of the 53 prednisolone and 7 (11%) of the 61 placebo patients died from pericarditis, and 11 (21%) and 18 (30%), respectively, required pericardiectomy. By 24 months 50 (94%) prednisolone and 52 (85%) placebo patients had a favourable status. 3 patients (1 prednisolone, 2 placebo) were normally active but were classified as not having achieved a favourable status. It is recommended that, in the absence of a specific contraindication, antituberculosis chemotherapy should be initially supplemented by steroids.