In a double-blind trial 122 patients aged 15-40 years with insulin-dependent diabetes of recent onset were randomly assigned to cyclosporin 7.5 mg/kg per day or placebo. At the sixth month 25.4% of the cyclosporin group and 18.6% of the placebo group were in complete remission (not a significant difference). Treatment was continued in those patients with complete or partial remission (insulin requirement less than 0.25 U/kg per day) and 106 patients were followed to nine months, at which stage 24.1% of the original cyclosporin group and 5.8% of the original placebo group were in complete remission (p less than 0.01). For those patients whose whole-blood trough cyclosporin levels in the first three months averaged 300 ng/ml or more, the rates of complete remission at six and nine months were 37.5% and 37%. The rates of partial remission were also higher in the cyclosporin group and at six months the rate of complete or partial remission was 46% in the whole cyclosporin group and 65.6% in those with an average blood level exceeding 300 ng/ml in the first three months, versus 28.8% in the placebo group. The principal side-effect of cyclosporin was a modest and reversible increase in plasma creatinine. These results indicate that cyclosporin promotes the remission of type I diabetes and suggest the need for new controlled protocols aimed at evaluating the length of the effect and selecting the best drug regimen.