Bleeding time measurement and investigation of platelet aggregation in platelet rich plasma (PRP) are routine procedures for the diagnosis of defects in primary hemostasis. These tests are subject to methodological difficulties and should be well standardized in each individual laboratory. - In the present study, bleeding time was measured using the Simplate II device in 40 normal subjects. Furthermore, platelet aggregation in PRP induced by ADP, collagen, arachidonate, and ristocetin was examined. 26 patients referred for investigation of a suspected mild bleeding disorder, who had a normal plasmatic coagulation profile, a normal von Willebrand factor activity, and a normal platelet count, were similarly studied. - Based on the reference values established in the 40 normal subjects, platelet aggregation was found to be pathologic in 7 patients and normal in 12. In 7 patients platelet aggregation was considered to be borderline-pathologic as defined by the range of platelet aggregability found in the 10% of our normal subjects showing the weakest aggregation responses. Bleeding time was prolonged in only 3 patients whereas it was normal in the remaining 23. There was strong evidence of a hemostatic defect as assessed by systematic patient history in 6 out of 7 patients with pathologic platelet aggregation, but in only 3 out of 19 showing normal or borderline-pathologic aggregation. - Pathologic platelet aggregation, therefore, represents not only an abnormal laboratory finding but is likely to be associated with a hemorrhagic diathesis. Platelet aggregation studies do not permit etiologic diagnosis of the thrombocytopathy except for the well-defined membrane glycoprotein deficiencies. The bleeding time appeared to be of low sensitivity for the diagnosis of mild platelet dysfunction.