The effect of oral dexamethasone on the plasma content of cortisol and dexamethasone was investigated in 175 patients suspected of having Cushing's syndrome. Plasma concentrations of cortisol and dexamethasone were measured by specific radioimmunoassays at 08.00 h following administration of either a low-low (0.5 mg), low (1.0 mg), high (4.0 mg) or high-high (8.0 or more mg) dose of dexamethasone at midnight. All seventeen patients with documented Cushing's syndrome exhibited resistance to the action of low-low and/or low dose dexamethasone on suppression of 08.00 h plasma cortisol content. Nine of twelve patients with pituitary dependent Cushing's syndrome had plasma cortisol values of less than 166 nmol/l following high-high dose testing. In 157 patients with suspected Cushing's syndrome, standard dexamethasone testing was considered unsatisfactory in at least 20%. After low-low or low dose tests 11% had supranormal cortisol values, but plasma cortisol content overlapped with values observed in patients with Cushing's syndrome only when plasma dexamethasone content was less than 5.6 nmol/l. Twelve per cent of patients suspected of having Cushing's syndrome had sufficient elevation of plasma dexamethasone values after low dose testing so that marked reduction of plasma cortisol might be expected even in patients with pituitary dependent Cushing's syndrome. Four patients receiving anticonvulsants had subnormal plasma levels of dexamethasone for the dose administered, but all exhibited normal suppression when plasma levels of dexamethasone and cortisol were correlated simultaneously. In summary, there is considerable variation in the plasma content of dexamethasone following oral doses. Simultaneous measurement of both plasma levels of dexamethasone and cortisol has proved most useful in identifying patients with unsatisfactory dexamethasone suppression tests.