A new in vitro bleeding time (BT) device was applied to various surgical patients. After setting the optimal assay condition, the normal in vitro bleeding time (T2) and volume (V2) were obtained from healthy volunteers, being 114.7 secs +/- 25.8 (SD) and 272.2 microliter +/- 69.1 (SD), respectively. When the T2 was below 210 secs, the platelet count was inversely proportional to the T2 and V2 of the in vitro BT with p less than 0.01. The value of the in vitro BT was not affected by heparin. Agents, which modify platelet functions, such as PGI2, DN9693 (an inhibitor of phosphodiesterase) and aspirin, prolonged the in vitro BT (T2, V2) dose-dependently. Administration of aspirin (660 mg) to volunteers prolonged the T2 from 108 to over 300 secs and the V2 from 253 to over 600 microliter but ticlopidine (500 mg/day for 3 days) had no effect. In 8 patients with liver cirrhosis who underwent hepatectomy, one patient with a prolonged T2 (260 secs) and a normal skin BT bled postoperatively, however, 3 patients with a prolonged skin BT (greater than 15 min) and a normal T2 had no hemorrhagic complications. From these observations it was concluded that in vitro BT is a convenient and useful tool to examine primary hemostasis or platelet function.