Summary. In Exp. I, 0·5 mg oestradiol or vehicle (0·5 ml absolute ethanol + 0·5 ml 0·9% NaCl) was injected i.v. at 08:00 h on Day 14 (onset of oestrus = Day 0). Blood samples were obtained via a jugular catheter at 30 and 1 min before oestradiol and every 30 min for 10 h afterwards. Plasma was obtained and assayed for 15-keto-13,14-dihydro-PGF-2α (PGFM) by radioimmunoassay. Before oestradiol, PGFM basal values were higher (P < 0·01) in pregnant (N = 10) than nonpregnant (N = 6) ewes (193 ± 30 vs 67 ± 8 pg/ml). However, at 4–10 h after oestradiol, pregnant ewes (N = 5) had less variable (P < 0·01) PGFM values than did nonpregnant ewes (N = 5). In Exp II, conceptus secretory proteins (CSP) were obtained by pooling medium from cultures of Day-16 sheep conceptuses (N = 40). Ewes received 750 μg CSP + 750 μg plasma protein (N = 6) or 1500 μg plasma protein (N = 6) per uterine horn at 08:00 h and 18:00 h on Days 12–14. All ewes received 0·5 mg oestradiol at 08:00 h on Day 14 and blood samples were collected as in Exp. I and assayed for PGFM. On Day 15, 3 ewes in each group received 10 i.u. oxytocin and 3 received saline i.v. at 08:00 h and blood samples were taken continuously from 10 min before to 60 min after treatment. Mean PGFM response to oestradiol was suppressed (P = 0·05) in CSP- vs plasma protein-treated ewes (371 ± 129 vs 1188 ± 139 pg/ml). Oxytocin induced a greater (P < 0·01) PGFM peak response in plasma protein-treated (3972 ± 2199 pg/ml) than CSP-treated (1669 ± 287 pg/ml) ewes. Uterine production of PGF in response to oestradiol was suppressed in pregnant and CSP-treated ewes and oxytocin-induced PGF production was also suppressed in CSP-treated ewes. These results are consistent with the theory that CSP act to prevent luteolysis by altering either the amount of PGF released by the uterus or the pattern of PGF release.