A reduction in endogenous opioid inhibition (disinhibition) of GnRH secretion is thought to be permissive for the preovulatory GnRH/LH surge. There are no published studies of the effects of highly specific opioid receptor agonists on the LH surge in any species, and the relative importance of the opioid receptor subtypes μ, δ and κ in the mechanism of disinhibition is unknown. In sheep, attempts to block the LH surge with opiates have been largely unsuccessful, and there is little evidence for reduced opioid inhibition during the GnRH/LH surge. The opioid receptor subtypes regulating PRL secretion in sheep are also unknown. Conscious, ovariectomized ewes with permanent third ventricular cannulae were injected with estradiol benzoate (EB) 50 μg or oil im (t = 0 h). In this model, EB elicits a time-delayed surge in LH secretion after 13-18 h. Jugular venous blood was sampled at half hourly intervals between -2 and 0 h and 10 and 26 h. From 12-20 h, infusions were made into the third ventricle of either the highly specific μ-agonist DAGO (10, 20 or 40 nmol/h), the δ-agonist DPDPE (40 nmol/h), the κ-agonist U50488 (40 nmol/h) or saline (vehicle). In oil- treated animals (n = 4-6), DAGO infusion at 20 and 40 nmol/h reduced plasma LH whereas DPDPE or U50488 had no effect. In EB-treated animals (n = 6), DAGO (40 nmol/h) delayed the LH surge (mean ± SEM time to surge onset 21.4 ± 0.3 h vs. 14.0 ± 0.4 h in controls, P < 0.0001). DAGO at 10 nmol/h did not alter surge onset and at 20 nmol/h had variable effects. DPDPE or U50488 did not affect LH surge timing or amplitude. All doses of DAGO increased plasma PRL, whereas DPDPE and U50488 had no effect. We conclude that, in ovariectomized ewes, activation of opioid μ-receptors, but not δ- or κ-receptors, inhibits GnRH secretion, can block the estrogen-induced GnRH/LH surge and increases PRL secretion. The results are consistent with the disinhibition hypothesis.