To determine the sites of action whereby estradiol (E2) induces the preovulatory luteinizing hormone (LH) surge in pigs, hypophysial-stalk-transected (HST) ovariectomized (OVX) gilts (mean ± SE: 118 ± 12 kg) received intramuscular injections of E2 benzoate (10 µg E2B/kg body weight) on day 0, plus intravenous pulses of 1 µg gonadotropin-releasing hormone (GnRH) every 45 min either from days -5 to 4, days -5 to 0 or days -5 to 0 then days 2 to 4. A fourth and fifth group received steroid vehicle on day 0 and pulses of GnRH on either days -5 to 4 or days -5 to 0 then days 2 to 4. In stalk-intact OVX gilts, E2B inhibited LH release for 48 h, then induced an LH surge that peaked 60–84 h after steroid treatment. In HST OVX gilts, serum LH increased from undetectable concentrations (<0.17 ng/ml) to 0.42 ± 0.03 ng/ml after 5 days of pulsatile GnRH replacement. In the presence of pulsatile GnRH stimulation, serum LH concentrations were inhibited for 12 h after E2B but then returned to values that were similar to those for HST gilts given GnRH in the absence of steroid treatment. Discontinuing GnRH pulses at the time of E2B injection caused serum LH concentrations to drop to 0.18 ng/ml or less for the remaining 96 h. When GnRH pulses were interrupted for 48 h, beginning immediately after injecting E2B or vehicle to mimic the secretory hiatus of LH and presumably GnRH observed in stalk-intact pigs prior to the surge, resumption of GnRH pulses caused serum LH concentrations to increase progressively over 3 h in E2B-treated gilts but to peak abruptly in controls given vehicle. During this period when GnRH administration was resumed, E2B successively increased LH pulse amplitudes and apparently prolonged the duration of LH release within each pulse, but did not augment the peak serum LH concentration. Before treatments were initiated, serum prolactin concentrations were higher in HST gilts than stalk-intact controls. The transection of the hypophysial stalk did not attenuate the E2B-induced release of prolactin that accompanied the LH surge in stalk-intact gilts. Based on these results, we submit that E2 induces the LH surge in gilts primarily by stimulating GnRH secretion, whereas changes in gonadotrope responsiveness to GnRH may influence only the surge profile. In contrast, E2 acts primarily on the pituitary gland to stimulate prolactin release.