The antidiuretic action of chlorpropamide was studied in 10 diabetes insipidus patients sensitive to vasopressin and 15 normal subjects. After an oral load of water followed by sustained hydration through water ingestion in a volume equivalent to the urinary flow, chlorpropamide (4 mg/kg body weight) was intravenously administered as a single dose, either alone (34 experiments) or simultaneously with pitressin infusion (6 experiments). During the experiments, creatinine clearance as well as osmolal and plasma and urine electrolyte concentrations were measured. Several urine collection periods were made before and after the administration of chlorpropamide. In thes experimental conditions, chlorpropamide promoted a marked antidiuresis (p less than 0.01) in patients with diabetes insipidus and presented no antidiuretic effect (p greater than 0.01) in normal subjects, even with a double dose. However, if previously to the experiment, the normal individuals were dehydrated (mean weight loss of 4.8 per cent) or the diabetes insipidus patients were hydrated (mean weight gain of 6.4 per cent) a reverse behavior was observed: that is, the normal subjects did present (p less than 0.01) and the diabetes insipidus patients did not present (p greater than 0.01) antidiuresis to chlorpropamide. The results suggest that the state of hydration modulates the action of chlorpropamide in a way apparently independent of the antiduretic hormone.