Three experiments were conducted to test the abortifacient effects of PGF2α analogues on mares during midgestation (average gestation length 141.5 days). The progesterone concentration was measured by radioimmunoassay. In experiment I, five mares recieved an injection of PGF2α analogue (fluprostenol: 500 μg intramuscularly) and a second injection either at 24, 48, of 72 h. Although the progesterone concentration decreased (P < 0.05) an average of 44 per cent in 24 h, none of the pregnancies were terminated. In experiment 2, beginning at least 10 days after experiment I, the same five mares were given PGF2α analogue as follows: 250 μg intravaginally and 500 μg intramuscularly. The treatment was repeated 48 h later. Progesterone concentrations had not increased since experiment 1 and dit not decrease during the 48 h following either injection. In experiment 3, six mares (average gestation length 162 days) were treated every 6 or 12 h with PGF2α analogue (cloprostenol: 375 μg) until expulsion of the fetus occurred at 47 ± 25 h after the initial injection; the mares received an average of 5 treatments. The progesterone concentration averaged 22 ± 7 ng/ml before the initial PGF2α treatment, decreased (P<0.05) to 8.4 ±2.7 ng/ml by 12 h before expulsion and 1–8 ±0.4 ng/ml 12 h after fetal expulsion. The progesterone concentration remained below 1.0 ng/ml for the next 4 days. However, only one of six mares exhibited estrual behavior after induced abortion.