Salivary cortisol represents a simple, noninvasive, stress-free measure that can greatly facilitate the longitudinal study of hypothalamic-pituitary-adrenal axis activity in patients with psychiatric disorders. By means of a slight modification of a commercially available radioimmunoassay kit, we studied the stability of salivary cortisol under different conditions, as well as the relationship between plasma and salivary cortisol under basal circadian conditions and following stimulation (CRH) and suppression (dexamethasone). We observed that salivary cortisol was quite stable at room temperature without centrifugation and that salivary and plasma cortisol values were highly correlated. Additionally, we observed a close correspondence in circadian and ultradian fluctuations in salivary and plasma cortisol. The salivary cortisol response to ovine and human CRH was similar to that observed with plasma cortisol, but was greater in magnitude. Finally, employing a plasma criterion as the standard, salivary measures identified 48% of the nonsuppressed Dexamethasone Suppression Tests (DSTs) and 97% of the suppressed DSTs.