This paper describes the episodic release and response to adrenal stimulation of cortisol and dehydroepiandrosterone (DHEA) in cows. Observations made in samples taken every 10 min for 8 h (experiment 1) showed that plasma DHEA was significantly greater (P < 0.001) than DHEA-S, and release of these steroids was episodic and variable between animals (P < 0.01). No relationship was found between DHEA and cortisol. Significant (P < 0.001) DHEA-sulphate (DHEA-S) versus cortisol (R = -0.264) and DHEA-S versus DHEA (R = 0.200) correlations were found. DHEA and DHEA-S were not affected by a single ACTH challenge (experiment 2). In experiment 3, cortisol and DHEA secretions in response to prolonged ACTH administration (every 12 h for 6 days) were studied. On day 7, the episodic cortisol and DHEA release and response to the opioid antagonist naloxone were studied in blood samples taken every 10 min for 8 h. Animals were injected with naloxone after 4 h. A significant increase (P < 0.05) in mean circulating DHEA and DHEA pulse amplitude was observed during frequent sampling following ACTH treatment. DHEA and DHEA-S plasma concentrations were not affected following luteal regression (experiment 4). The effect of milk secretion around parturition on DHEA secretion was studied in dry and continuously milked cows (experiment 5). Plasma DHEA was significantly lower (P < 0.05) in milked cows. In the cow, ACTH is not an important DHEA secretagogue. Adrenal contribution to plasma DHEA is scarce. Likely, the placenta is the most important source of DHEA, and the lactating mammary gland can affect circulating DHEA levels. Investigation about the DHEA biological role in cows should be focused around parturition.