IGF-I was measured by RIA in plasma samples collected 8-hourly for 24 days which included two consecutive preovulatory surges of LH. In a separate study, ovarian venous blood was collected from animals undergoing ovariectomy on day 10 of the oestrous cycle, or 36 h later after being treated with prostaglandin with or without steroid-free bovine follicular fluid. Jugular venous blood samples were collected before, during and after surgery. Follicles were dissected from ovaries of these animals and sorted into categories of small, intermediate and large, non-atretic or atretic, and the follicular fluid was pooled and assayed for IGF-I. From another population of ovaries recovered from the slaughterhouse, granulosa, theca and corpora lutea were isolated, homogenized and assayed for IGF-I. Finally ovarian corpora lutea and granulosa cells were each incubated with tritiated amino acids overnight at 37 degrees C. Thereafter the tissues and media were sonicated, IGF-I extracted from the supernatant and tritiated IGF-I precipitated using a specific IGF-I antibody. The absence of any significant change in peripheral IGF-I concentrations following ovariectomy and the finding that the ovarian venous IGF-I concentrations (161 +/- 10 micrograms/l) were not significantly different from levels seen in peripheral blood (157 +/- 10 microgram/l) indicated that the ovary is not a net exporter of IGF-I. However, the ovary does synthesize IGF-I, as evidenced by granulosa and luteal synthesis, but probably not in quantities in excess of that utilized by ovarian tissues per se. Although the plasma IGF-I levels increased around the second preovulatory LH surge, the results overall indicated that the IGF-I concentrations in plasma are not strictly related to any major ovarian event during the oestrous cycle in the sheep. This view is based on the findings that the concentration of IGF-I in follicular fluid was not related to follicular health but correlated with those in peripheral plasma and that the ovarian venous concentrations did not vary between left and right ovaries irrespective of whether the ovaries contained a corpus luteum, dominant follicle or neither. Collectively, these results are consistent with the notion that IGF-I of ovarian origin fulfils an autocrine/paracrine function and does not have an endocrine role. Moreover, the results show that the concentrations of IGF-I in follicular fluid reflect those in peripheral plasma.