Six male volunteers received either 0 (buffer), 2.5 or 5.0 ng/kg/min PGI 2·Na for 72 hrs. Various platelet parameters were monitored for an additional 72 hrs. Each morning, for seven consecutive days, and +1 and +6 hrs after the termination of the infusion, blood was drawn and platelet rich plasma (PRP) was prepared. The PRP was immediately exposed to 100 ng/ml PGI 2·Na, and the subsequent increase in platelet cyclic AMP was measured by radioimmunoassay. Aggregation in response to 2 or 4 μM ADP was measured simultaneously. Three volunteers returned for a second 72 hr infusion of 5.0 ng/kg/min PGI 2·Na. After 72 hrs, the infusion rate was gradually “tapered off” over a 12 hr period at which time the infusion was terminated. The sensitivity of the PRP to ADP-induced aggregation was recorded before, during, and after the “tapering off” regimen. Platelet counts were not altered by any of the infusions. The responsiveness of the platelet adenylate cyclase to exogenous PGI 2·Na was inversely related to the concentration of PGI 2·Na infused. Desensitization ocurred and was more severe after 72 hrs of infusion than after either 24 or 48 hrs. For example, after 72 hrs at 5.0 ng/kg, platelets lost approximately 50% of their responsiveness to PGI 2. ADP-induced aggregation was not significantly inhibited ex vivo by the infusion of 2.5 ng/kg/min. During the infusion of 5.0 ng/kg/min PGI 2. ADP-induced aggregation was inhibited at 24 and 48 hrs, but by 72 hrs, the platelets began to respond to ADP more like control cells even though the PGI 2·Na infusion was continuing. When the infusion was abruptly terminated a hyperaggregable response (rebound) to exogenous ADP was observed. In subjects where 5.0 ng/kg/min infusion was gradually “tapered off” over a 12 hr period, there was no evidence of platelet hyperaggregability at the time points studied.