The effects of increasing in vitro cyclosporine concentrations (0, 50 100 or 200 ng./ml.) on lymphocyte blastogenesis, measured by incorporation of tritiated thymidine and induced by varying levels of concanavalin A (0, 0.25, 1.0 or 5.0 ng./ml.), were studied in regard to mean serum level of cyclosporine in 26 renal allograft recipients. Results were compared to similar data obtained in healthy controls. Patients were divided into group 1 (13 patients, mean serum cyclosporine trough level less than 150 ng./ml.) and group 2 (13 patients, cyclosporine level greater than 150 ng./ml.). With no cyclosporine added to the assay proliferation of lymphocytes obtained from all patients inversely correlated to the mean serum trough cyclosporine level at all stimulatory levels of concanavalin A (0.25 ng./ml., p less than 0.01; 1.0 ng./ml., p less than 0.001 and 5.0 ng./ml., p less than 0.001) and was significantly lower than in controls (p less than 0.0002). Whereas increasing in vitro cyclosporine concentrations has produced the expected increase in suppression of blastogenesis in controls and group 1, a paradoxical effect became evident in group 2. Under stronger stimulatory conditions (concanavalin A 1.0 or 5.0 ng./ml.) increasing in vitro cyclosporine concentrations were associated with significantly decreased suppression of blastogenesis (p less than 0.01) compared to group 1. These results confirm previous reports and suggest that the duality of effect of cyclosporine in this in vitro model may be related to its functional relationship to the calcium ion (Ca++)/calmodulin complex and to its cellular concentration/solubility curve. These considerations may be of importance in adjusting cyclosporine dosage based on serum trough levels of cyclosporine.