Natural GnRH and its analog have potential for hastening ovulation in mares. A study was conducted to evaluate the efficacy of a GnRH agonist given either as an injectable or s.c. implant for induction of ovulation in mares. Forty-five seasonally anestrous mares (March) were assigned to one of three groups (n = 15/group): 1) untreated controls; 2) i.m. injection of the GnRH agonist buserelin at 12-h intervals (40 micrograms/injection for 28 d or until ovulation) and 3) GnRH agonist administered as a s.c. implant (approximately 100 micrograms/24 h for 28 d). Six mares per group were bled on d 0, 7, 14 and 21 after injection or insertion of implant. Samples were taken at -1, -.5 and 0 h and at .5, 1, 1.5, 2, 4, 6 and 8 h after GnRH. Additional daily samples were drawn for 28 d after injection or until ovulation. Samples were assayed for concentration of LH and FSH. Progesterone concentrations were determined in samples collected on d 4, 6 and 10 after ovulation. Number and size of follicles and detection of ovulation were determined by ultrasonography. Number of mares induced to ovulate within 30 d was 0 of 15, 7 of 15 and 9 of 15 for groups 1, 2 and 3, respectively. During treatment, follicle sizes were smaller for mares in group 3 (implant). The LH response to GnRH agonist (area under curve) was similar among groups at d 0 but was greater (P less than .05) for mares in group 3 on d 7 and 14 and groups 2 and 3 on d 21 than for controls. A similar pattern was detected for peak concentrations of LH after GnRH on d 0, 7, 14 and 21. Daily concentrations of LH remained low in untreated control mares compared with GnRH-treated mares throughout the sampling period. Concentrations of LH for mares in group 3 that ovulated were elevated greatly above those for group 2 mares, whereas concentrations of FSH were similar in both treatment groups prior to ovulation.