Twelve pony geldings with (n = 6) and without (n = 6) testosterone replacement (200 micrograms/kg testosterone propionate in oil, SC every 48 hours) received either gonadotropin releasing hormone (GnRH; 25 micrograms SC every 3 hours) or control treatment. Sexual behavior was recorded during 4-minute exposure to an estrous mare, 3 times weekly for 2 weeks before treatment, 3 weeks during treatment, and 3 weeks after treatment had been discontinued. The group receiving testosterone and GnRH (n = 3) exhibited significantly greater flehmen response frequency and attention duration and significantly lower vocalization frequency and erection duration than the group receiving only testosterone (n = 3). GnRH had no apparent effect on sexual behavior in geldings without testosterone replacement (n = 3). These results support an hypothesis of testosterone-dependent, CNS-mediated effects of GnRH on precopulatory behavior in the stallion.