Ovine and rat pituitary bioassays for gonadotrophin surgeattenuating factor (GnSAF) were utilized to determine whether the level of GnSAF bioactivity in pooled human follicular fluid (hFF) from superovulated women varied according to follicle diameter (≤11 mm, 12–15 mm and 16–21 mm follicles examined using the ovine bioassay, or ≤10 mm, 11–13 mm, 14–17 mm, 18–20 mm, 21–24 mm and ≥ 25 mm follicles examined using the rat bioassay). When tested using dispersed ovine pituitary cells, GnSAF bioactivity, expressed in terms of the reduction in gonadotrophin-releasing hormone (GnRH)-induced LH secretion, was inversely related to follicle diameter (P<0·01). In response to 5 μl hFF/well from follicles of ≤ 11, 12–15 and 16–21 mm diameter, GnRH-induced LH secretion was reduced to 40·5±6·6.9%, 65·2±6·6% and 83·7±7·9% of control respectively. A similar inverse relationship was observed when a second batch of hFF samples from different sized follicles was tested using rat pituitary cell monolayers. Expressing GnSAF bioactivity in terms of the dose required to suppress GnRH-induced LH secretion by rat pituitary cells to 50% of the maximal suppression observed (ED50), the three smallest follicle size pools contained the most GnSAF (ED50 values of 0·13, 2·79 and 5·36 μl hFF/well from follicles of ≤ 10, 11–13 and 14–17 mm respectively). The ED50 values for follicles of 18–20, 21–24 and ≥25 mm were 8·81, 27·1 and 60·0 μl hFF/well respectively. Thus hFF from follicles ≤ 11 mm was over 450 times more potent than hFF from follicles ≥ 25 mm in suppressing GnRH-induced LH release. The ED50 values for inhibin bioactivity (measured as the suppression of basal FSH secretion from rat pituitary monolayers) were much less variable than those for GnSAF bioactivity (between 0·85 and 0·13 μl hFF/well). Inhibin immunoreactivity, measured by a two-site immunoradiometric assay, followed the same pattern as inhibin bioactivity with lowest concentrations in the smallest follicles (41·96 ng/ml) and highest concentrations in the three largest follicle size groups (56·48–64·48 ng/ml). The specific effects of inhibin on GnRH-induced LH and basal FSH release in these pituitary bioassays were determined by incubating culture dishes with pure recombinant human inhibin at doses of 0·025–25 ng/well. In both the sheep and rat pituitary monolayers, basal FSH was suppressed (ED50=0·02 and 0·16 ng/well respectively). However, while inhibin markedly stimulated GnRH-induced LH secretion from ovine pituitary monolayers (ED50=0·04 ng/well), it suppressed GnRH-induced LH secretion from rat pituitary monolayers (ED50=0·31 ng/well) by 13%. The divergent effects of inhibin on GnRH-induced LH secretion in the two culture systems, and the relative insensitivity of GnRH-induced LH secretion to recombinant human inhibin in the rat system, indicates that the inverse relationship between GnSAF concentrations and follicular diameter cannot be an artefact of inhibin bioactivity. In addition, when hFF was fractionated by hydrophobic interaction chromatography using phenyl Sepharose, fractions which contained the greatest amounts of GnSAF bioactivity differed from those which contained peak levels of bioactive or immunoreactive inhibin. These results support in vivo observations that small follicles are important regulators of gonadotrophin secretion in superovulated women. Concentrations of GnSAF fall as the follicles approach an ovulatory size which enables positive steroid feedback on pituitary responses to hypothalamic GnRH, leading to the preovulatory LH surge.