Homogeneous Silastic rods containing ethinylestradiol (EE) (1.5 or 4 g), estradiol-17 beta (E) (4 g) or progesterone (P) (6 g) were implanted into feral mares (Equus caballus) between 4- and 10-yr-old. Six treatment groups (greater than or equal to 10 mares/group) of non-pregnant mares received 36 g P and 12 g E (P+E), 36 g P and 8 g EE (P+HEE), 1.5 g EE (LEE), 3 g EE (MEE, 8 g EE (HEE) or control-implanted mares (CI). CI received implants containing no steroid. Two groups of pregnant mares received P+HEE or HEE. Stallions were placed with the mares 15 to 26 mo after implanting. Blood was collected biweekly for up to 28 mo after implanting and serum analyzed for P by radioimmunoassay. A single P value greater than or equal to 2.5 ng/ml indicated ovulation and 2 consecutive values greater than or equal to 2.5 ng/ml indicated pregnancy. Serum from blood collected before and at 4, 12, 24, 50, 64 and 89 wk after implanting was analyzed for EE concentrations. All animals pregnant at the time of contraceptive placement delivered normal foals. Contraceptive efficacy for groups LEE, MEE, HEE and P+HEE were 75, 75, 100, and 100%, respectively after two breeding seasons. Suppression of ovulation appeared to be inversely related to the concentration of EE used in the implant. The percent of animals ovulating after 2 yr of contraception in each group was 100, 100, 88, 62, 20, and 12 for groups CI, P+E, LEE, MEE, HEE and P+HEE, respectively. The pregnancy rate for the same groups was 100, 78, 25, 25, 0 and 0%, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)