The ability of interrupted photoperiods to induce early estrus and ovulation was examined. Horse mares were exposed to long (16 h light) or short (10 h light), noninterrupted photoperiods, ambient light, or various interrupted photoperiod treatments from December 1 to April 15 (135 d). Follicular development was assessed by rectal palpation and estrous behavior was determined by teasing with a stallion. Serum concentrations of progesterone were used as an indicator of corpus luteum function. Differences among the light treatment groups were compared for the following behavioral and ovarian characteristics: days to first detectable 3-cm follicle, days to first estrous behavior, days to first ovulation, the number of mares ovulating within the treatment period, and the number of ovulations within the treatment period per mare. Compared with the ambient and 10L:14D (L = h of light and D = h of darkness) photoperiod treatments, ovulation was advanced to the greatest extent by a photoperiod of 16L:8D and the interrupted photoperiod 10L:8D:2L:4D. These two stimulatory photoperiod treatments were characterized by the presence of light 8 to 10 h after dusk. Therefore, the present data are consistent with an external coincidence model for the induction of seasonal breeding in horses, with the photoinducible phase occurring within the period 8 to 10 h after dusk.