A comparison of serum inhibin levels in men and women was undertaken using a sensitive sheep pituitary cell in vitro bioassay and a newly developed heterologous RIA. The RIA was based on an antiserum raised to bovine 31K inhibin using [125I]31K inhibin as tracer. Bovine inhibin alpha- and beta-subunits, bovine activin-A, transforming growth factor-beta, and Mullerian inhibitory substance did not cross-react in the RIA. In both assays, dilutions of serum gave response lines parallel to that of the partially purified human follicular fluid inhibin preparation used as standard. Negligible levels of both bio (B)- and immuno (I) activities were found in serum from women with premature ovarian failure or castrated men. In ovulation-induced cycles, serum B inhibin levels increased progressively from the early to the late follicular phase and remained at the late follicular phase level during the early and midluteal phases. Serum I inhibin levels also rose during the follicular phase, but declined during the early luteal phase before increasing again in the midluteal phase. As a consequence, inhibin B:I ratios varied during the treatment cycle, with high ratios in early follicular (2.86) and early luteal (2.25) phases and a low ratio in the midluteal phase (1.09). Similar changes in serum B:I ratios also occurred during the midcycle and midluteal phases of normal cycles. The B:I ratio was lower (0.35) in normal men. We conclude that the largely similar pattern of inhibin biological and immunological activities in serum obtained during a variety of physiological conditions support the validity of the RIA procedure, and the B:I ratio of serum inhibin varies during the follicular and luteal phases of the cycle and is low in men. Potential reasons for these changes in B:I ratio include the presence of interfering substances in either the bioassay or the RIA, the presence of inhibin isoforms, and/or modulation of secreted forms by sex steroids.