Three experiments were performed to test the following hypotheses: 1) stallions and/or progesterone-estradiol-treated geldings could serve as models for the effects of a single implant of the GnRH analog, deslorelin acetate, on LH and FSH secretion by mares; and 2) multiple implants of deslorelin acetate could be used as a means of inducing ovarian atrophy in mares for future study of the mechanisms involved in the atrophy observed in some mares after a single implant. In Exp. 1, nine light horse stallions received either a single deslorelin implant (n = 5) or a sham injection (n = 4) on d 0. In Exp. 2, 12 geldings received daily injections of progesterone on d -20 through -4, followed by twice-daily injections of estradiol on d -2 to 0. On the morning of d 0, geldings received either a single deslorelin implant (n = 6) or a sham injection (n = 6). Daily injections of progesterone were resumed on d 2 through 15. In Exp. 1, plasma LH and FSH were elevated (P < 0.05) in the treatment group relative to controls at 4, 8, and 12 h after implant insertion. In the treated stallions, FSH was decreased (P < 0.05) on d 3 to 13, and LH was decreased on d 6 to 13. In Exp. 2, plasma LH and FSH were elevated (P < 0.05) at 4,8, and 12 h after deslorelin implant insertion. Plasma LH was suppressed (P < 0.05) below controls on d 2 to 7, 9, and 11 to 15; plasma FSH was suppressed (P < 0.05) on d 4 to 15. In Exp. 3, 21 mares were used to determine whether multiple doses of deslorelin would cause ovarian atrophy. Mares received one of three treatments: 1) sham injections; 2) three implants on the first day; or 3) one implant per day for 3 d (n = 7 per group). Treatment with multiple implants increased (P < 0.05) the interovulatory interval by 14.8 d and suppressed (P < 0.01) LH and FSH concentrations for approximately 25 d; no mare exhibited ovarian atrophy. In conclusion, after an initial short-term increase in LH and FSH secretion, deslorelin implants caused long-term suppression of both gonadotropins in stallions as well as in geldings treated with progesterone and estradiol to mimic the estrous cycle. It is likely that either of these models may be useful for further study of this suppression in horses. Although multiple implants in mares suppressed gonadotropin secretion longer than a single implant, the lack of ovarian atrophy indicates that the atrophy observed after a single implant in previous experiments was likely due to the susceptibility of individual mares.