Blood was collected at 20-second intervals from the external carotid artery and from the dorsal longitudinal sagittal sinus (sagittal sinus, SS) of ovariectomized sheep. The point of SS catheterization was very near the point at which diencephalic effluent entered the SS. Concentrations of beta-endorphin (β-EP) immunoreactivity were quantified by radioimmunoassay procedures in blood plasma and in cerebrospinal fluid (CSF) from the cisterna magna. Increases in plasma β-EP concentration were provoked by intracarotid injection of naloxone and by experimental production of bacteremia (i.e., intravascular bacteria), but these procedures failed to increase β-EP in CSF. Quantities of β-EP in plasma samples from the SS were assumed to represent arterial contribution (minus tissue uptake), diencephalic secretion, and retrograde delivery of pituitary β-EP to the diencephalic effluent. The arterial contribution was removed mathematically by subtracting the arterial plasma β-EP concentration from the concurrent SS plasma concentration of β-EP to yield a paired arteriovenous (AV) difference. When this AV difference was consistently positive and satisfied our statistical criterion for being greater than zero, we concluded that either pituitary β-EP was delivered in a retrograde manner to diencephalon or the diencephalon secreted β-EP. However, this situation occurred in only 5 of 31 periods examined. Furthermore, only 2 of these 5 periods occurred during times of increasing arterial concentrations of β-EP. Such concurrence would be expected if both changes were caused by a major discharge of β-EP from the pituitary gland. Therefore, the present results provide little evidence for retrograde delivery of pituitary β-EP to the brain without system