We have developed a suppressive test for the hypothalamic-pituitary-adrenal (HPA) axis using prednisolone, which is similar to endogenous glucocorticoids. We used a single-blind, repeated-measure design in healthy volunteers. In the first phase of the study, we compared placebo or prednisolone 2.5 mg, 5 mg, or 10 mg; in the second phase of the study, we compared placebo or prednisolone 5 mg or dexamethasone.5 mg. On the following day, we collected plasma and salivary cortisol levels from 9 AM to 5 PM. Maximal average prednisolone plasma levels (at 9 AM after the 10-mg dose) were 30 to 35 ng/mL. At all doses, prednisolone caused a larger suppression of salivary cortisol (approximately 20% after 2.5 mg, 30% to 35% after 5 mg, and 70% to 75% after 10 mg) than of plasma cortisol (approximately 5% after 2.5 mg, 10% after 5 mg, and 35% after 10 mg). Dexamethasone.5 mg gave 80% suppression of plasma cortisol and 90% suppression of salivary cortisol. Plasma and salivary cortisol levels were more consistently correlated in each subject after prednisolone than after dexamethasone. We propose that prednisolone at the 5-mg dosage (which gave partial HPA suppression), together with the assessment of salivary cortisol, can be used to investigate both impaired and enhanced glucocorticoid-mediated negative feedback in large samples of patients with psychiatric disorders.