Since 1985, we have applied our nonsurgical technique for collecting pituitary venous (PitVen) blood from ambulatory horses to investigate the regulation of adrenocorticotropic hormone (ACTH) secretion. This method offers particular advantages for studying the hypothalamo-pituitary-adrenal axis since its benign nature enables hypothalamic and pituitary interactions to be monitored without disturbing the animal, and the horse's large blood volume allows 3- to 4-ml samples to be collected as frequently as every 20s for prolonged periods so that the secretion patterns of ACTH and its secretagogues can be precisely defined. When PitVen blood was sampled every 20 or 30s during the circadian maximum, arginine vasopressin (AVP) and ACTH secretion patterns were complex and irregular, with mean interpeak intervals of approximately 5 min. Despite their erratic patterns, AVP and ACTH secretions were closely coupled on cross-correlation analysis. By contrast, PitVen corticotropin-releasing hormone (CRH) concentrations were low, relatively stable, and not consistently related to ACTH secretion. However, when cortisol negative feedback was reduced acutely by metyrapone infusion, CRH and AVP secretion were stimulated. Mathematical modeling suggested that CRH had become the more effective secretagogue and that much of the ACTH response was mediated by increased pituitary responsiveness to CRH. Elevated blood osmolality triggered synchronous AVP and ACTH secretion, without altering PitVen CRH. In this case, the source of PitVen AVP was presumably the magnocellular/neurohypophysial pathway, which is thought to respond primarily to changes in blood osmolality and pressure. Our results suggest that this pathway also participates in ACTH regulation. We have studied the effect of several perturbations and found, as have others, that the secretagogues released vary with the stimulus given. For example, vigorous exercise promptly raised PitVen AVP and ACTH, but not PitVen CRH. Hypoglycemia provoked both CRH and AVP secretions, with the CRH increment being inversely proportional to the glucose nadir. Administration of the opioid antagonist, naloxone, increased PitVen ACTH; however, changes in AVP and CRH were variable and overall could not account for the ACTH response. This suggests that endogenous opioids inhibit a third ACTH secretagogue, stimulate an inhibitory factor, or also act at the pituitary. Chronic social stress, induced by confining newcomers with aggressive, resident mares, caused most introduced horses to become submissive. In such horses, plasma cortisol declined to levels similar to those during metyrapone infusion. Despite hypocortisolemia, PitVen ACTH was low, whereas PitVen CRH tended to be elevated. Moreover, chronically stressed horses did not respond to exogenous CRH. We conclude that at rest and during some perturbations AVP is the immediate stimulus for ACTH release. Even ACTH micropulses, previously thought to occur spontaneously, appear to be regulated by AVP in horses. On the other hand, CRH secretion and pituitary responsiveness to CRH rise when cortisol falls, suggesting that a major role for CRH is to fix the cortisol setpoint. However, during chronic stress, these relationships become disturbed, with results to date pointing toward the existence of an ACTH-release inhibiting factor.