In a randomised double-masked study of 27 patients with a severe chronic idiopathic uveitis we evaluated the efficacy, safety, and tolerability of cyclosporin. All received prednisone in a low dose (0.3 mg/kg/day). In 14 patients this was combined with cyclosporin in a single daily dose of 10 mg/kg/day, while 13 patients received a placebo. The dosages were tapered off in accordance with a protocol, and we compared the number of months of successful therapy before the uveitis relapsed. The efficacy results, as expressed in a Kaplan-Meier curve, were in favour of cyclosporin. Owing to the small sample size, however, this difference did not reach statistical significance. The immunosuppressive effect of cyclosporin was not permanent, and in all but one patient the intraocular inflammation relapsed on reduction of dosage. Rather small cumulative doses of cyclosporin proved to be nephrotoxic, but subjective tolerability for cyclosporin was good.