Recently, some investigators have established a blunted natural killer cell activity (NKCA) in severely depressed patients. In order to replicate these findings NKC cytotoxicity assays--on fresh cell suspensions in human plasma and fetal calf serum--were performed in healthy controls and depressed inpatients. Instead of the commonly used 51Cr-release assay we have used a fluorescent NKC cytotoxicity assay, which allows a greater sensitivity. We observed a significantly blunted NKCA in melancholic patients as compared with healthy controls and minor depressives, whilst simple major depressives exhibited an intermediate position. NKC cytotoxicity assays in fetal calf serum were significantly and negatively correlated with the severity of illness. We were unable to establish any relationship between NKCA and measures of hypothalamic-pituitary-adrenal-axis function, such as baseline, postdexamethasone plasma cortisol and 24 hr urinary cortisol secretion. In addition, we did not find any effects of dexamethasone administration (1 mg orally) on NKCA.