1 Plasma cortisol levels were measured before and for 6 h after the intravenous injection of 50 mg cortisol as sodium succinate and oral administration of 50 mg cortisol and 50 mg cortisone acetate in 10 subjects with primary or secondary adrenal failure and in two normal volunteers. 2 Peak cortisol levels of 1518 + 190 nmol l-l (mean + s.e. mean) and 739 + 74 nmol 1-1 were found 1.46 + 0.25 and 1.79 + 0.16 h after oral cortisol and cortisone acetate respectively. The relative bioavailability of oral cortisol and cortisone acetate varied widely (cortisol 26-91%, mean 54 + 6.9%, cortisone acetate 21-95%, mean 44 + 6.5%) but despite this wide variation there was, in individual subjects, a highly significant correlation between the bioavailability of the two steroids (r = 0.870, P < 0.001). 3 This suggests that the wide interindividual variations in plasma cortisol levels seen after oral cortisone acetate are not related to variations in bioconversion of cortisone.