Accurately determining the amount of blood lost during bleeding time tests has been little exploited. Equipment is now available for measuring both the time and volume automatically--a technique called hemorrhagometry. This quantitative procedure has generated a number of interesting and unsuspected observations. Bleeding times lengthen and blood loss is exaggerated as the skin is cooled. This is particularly true of patients who have hemostatic defects, such as hemophilia A; Not only is the hemophiliac different from normal in this regard but, in Sutor's preliminary observations, the asymptomatic carrier of hemophilia A also seems to be distinguishable from the noncarrier on this basis. The test seems potentially useful in evaluating the effect of drugs on platelet function and in assessing mild bleeding diatheses. Variations induced by either warming or cooling the skin may well have clinical applicability.