Both ovaries of 31 rabbits were perfused with a chemically defined medium in vitro in a recirculation system. In one series of experiments, hCG (100 IU) was injected iv 5-6 h prior to anaesthesia and surgery. Approximately 1 h later the perfusion was started. One ovary was perfused as control while the other ovary was perfused with 5 micrograms/ml indomethacin or with indomethacin and 1 micrograms/ml PGF2 alpha. In another series of experiments the rabbits received no pretreatment prior to operation. Instead, bovine LH was added to the perfusion medium of both control and experimental ovaries. The experimental side also received either indomethacin or indomethacin and PGF2 alpha. Finally, the effect of PGF2 alpha in the absence of LH was compared to the control ovary receiving only LH. After injection of hCG in vivo, ovulations occurred in 4 of 5 control ovaries. Indomethacin completely blocked ovulation in 4 of the 5 ovaries treated, while PGF2 alpha restored ovulations in all the experimental ovaries. In the group of experiments where LH was added in vitro, ovulations were induced in all ovaries treated with varying LH doses. Furthermore, indomethacin blocked ovulation in 5 out of 7 ovaries, and PGF2 alpha restored ovulation in all ovaries. Fifty per cent of the ovaries treated only with PGF2 alpha (in the absence of LH) also ovulated. The pattern of steroid release did not differ between control ovaries, indomethacin treated ovaries, and indomethacin + PGF2 alpha treated ovaries. Ovaries treated in perfusion with PGF2 alpha alone had very low steroid levels compared to the ovaries treated with LH.(ABSTRACT TRUNCATED AT 250 WORDS)