The effectiveness of the prostaglandin F analogue fluprostenol in inducing labour in the mare was examined by giving sequential injections over the last 50 days of gestation. The behavioural and endocrine changes elicited by the drug in pregnant and non-pregnant animals and in foals were also studied. Fluprostenol (250 or 500 micrograms intramuscularly) failed to induce labour before 320 days gestation; thereafter its effect was capricious. Twelve mares foaled 1 to 36 h after the last test; eight delivered normal, viable, apparently 'term' foals and four produced stillborn/premature animals. Eight of the deliveries (five term and three pre-term foals) could be ascribed to the action of fluprostenol because they occurred 1 to 6 h after its administration, at a time when spontaneous foaling would have been unlikely. The other four mares foaled between 12 and 36 h after the fluprostenol injection and it is therefore doubtful whether there was a causal relationship between the two events. In the mares which delivered viable foals the pre-partum milk samples were characteristic of full term samples with respect to calcium, sodium and potassium. Those which delivered premature/stillborn foals had low calcium and a high sodium/potassium ratio in the pre-partum milk. Behavioural changes (sweating, increased respiration, defaecation etc), which varied in intensity between tests and individuals, were seen in all three groups of animals following the administration of fluprostenol. These changes were accompanied by rises in plasma cortisol and adrenocorticotrophic hormone concentration during the 2 h sampling period, suggesting a centrally mediated response to the drug.(ABSTRACT TRUNCATED AT 250 WORDS)