SUMMARY A radioimmunoassay method for the measurement of arginine-vasopressin (AVP) in human plasma has been developed which requires 5 ml of plasma and has a lower limit of detection of 1·8 pg/ml plasma. Arginine-vasopressin was found to be stable in whole blood for up to 1 h at room temperature and for at least 4 h at 4 °C, while in plasma stored at − 20 °C no loss was seen over 10 days. Dehydration and rehydration in normal subjects produced appropriate changes in AVP concentration but there was considerable variability in the levels attained by individual subjects and no obvious correlation with plasma osmolality. No consistent increase in plasma AVP concentration was seen on change of posture from the recumbent to the upright position. Vigorous exercise produced a marked rise in plasma AVP concentrations in most subjects which could not be attributed simply to an increase in plasma osmolality. Infusion studies with Pitressin in normal subjects showed a mean halflife of 6·4 min with an overall plasma clearance rate of 8·5 ml/min/kg body weight and a mean volume of distribution of 5·33 1. In patients with a biochemical picture suggestive of inappropriate antidiuretic hormone secretion, markedly raised plasma AVP concentrations were found only in patients with bronchial carcinoma.