The response to an osmolar load (750 ml 2.5% NaCl solution iv preceded by 500 ml water by mouth) was studied in 20 patients with Sheehan's syndrome and 12 normal women. Sodium and osmolality were determined in plasma and urine and arginine-vasopressin (AVP) was measured by RIA in urine. The test was performed in each patient when untreated (group P), after hydrocortisone replacement alone (group F), and combined hydrocortisone and thyroid hormone replacement (group F+T). After the osmolar loading, maximum urinary osmolality in the patients was lower than in the normal women and remained unaffected by both hydrocortisone alone and hydrocortisone and thyroid hormone. Comparison of the mean hourly urinary volume before and after NaCl infusion demonstrated an increase in group P, a decrease in group C, and no change in groups F and F+T. Although free water clearance became negative in all groups, values in groups P, F, and F+T were constantly above that of group C. None of the patients in groups P and F had a significant rise in urine AVP excretion during or after NaCl infusion. Those in group F+T had a slight AVP response which was less than in normal women. Impaired response of AVP to an osmolar load appears to be a constant feature of Sheehan's syndrome even without overt diabetes insipidus.