Forty-three diabetic patients on maintenance chlorpropamide (100-750 mg daily) drank 0.2 ml/kg 90% ethanol after equilibration in a room controlled at 20 degrees C. Twenty-five patients had already noted marked alcohol flushing since starting chlorpropamide therapy (group A), while 13 had not observed this (group B). The remainder were teetotal or unsure of their reaction. Cheek temperature rise correlated with plasma chlorpropamide concentration (r = 0.6, p less than 0.001) in all patients and was inversely related to basal cheek temperature (r = -0.35, p less than 0.02). Plasma chlorpropamide correlated with daily chlorpropamide dose (r = 0.8, p less than 0.001) but not with basal cheek temperature. The correlation between chlorpropamide level and cheek temperature rise was strengthened on analysis of group A alone (r = 0.7, p less than 0.001) and absent in group B (r = 0.2, p greater than 0.3) who tended to have lower chlorpropamide levels and cheek temperature rise than group A.