To understand the complex biochemistry and physiology of activin and inhibin requires the development and validation of a number of specific immunoassays. The importance of this endeavour is signified by the tremendous effort spent by a number of laboratories to develop such assays over the past decade. Based on the deduced amino acid sequence obtained when the inhibin-\g=a\and -\g=b\A/\g=b\Bgenes were cloned (1), RIAs using synthetic peptides were developed that were quite useful in animal studies (2). Human applications required the further development of a heterologous RIA using purified bovine inhibin as antigen, the result of which was an assay that allowed the charting of inhibin's physiology in a number of settings (3). It was soon discovered that multiple forms of inhibin and its free \g=a\-subunitexist in physiological fluids (4\p=n-\7) and these free \g=a\-subunitscross-react in the existing RIAs, thereby complicating interpretation of results (7). This observation spurred development of second-generation assays using synthetic peptides to target antibody development to specific sequences on a-or ß-subunits, allowing the formulation of two-site immunoassays with required specificity for dimeric inhibin (8,9). During the development phase, it was discovered that binding proteins for inhibin and activin existed in physiological fluids that interfered with analytical recovery (10-12), thereby requiring adaptation of assay formats to circumvent these problems (13-15).Assays for activin, the homodimeric ß-subunit relative of inhibin, have been even more problematic for a variety of reasons. Antibody development has been complicated by the relatively high degree of primary sequence conservation among the different species, as well as the difficulty in choosing appropriate sequences for synthetic peptide immunogen strategies. Never¬ theless, a number of assays have been developed recently that can detect activin in various physiological fluids, employing strategies ranging from specific antibodies (16, 17) to use of binding proteins such as follistatin (FS) (18) in competitive binding formats.These activin assays have all suffered from one major complication, however. Follistatin, a 32-35kD singlechain polypeptide, binds and neutralizes activin (19) with high affinity and low reversibility under physiolo¬ gical conditions, including those in serum and human follicular fluid (hFF) (10, 20). Furthermore, the binding sites of many peptide-directed anti-activin antibodies overlap with at least one of follistatin's putative activin binding sites (20). Thus, without sample pretreatment to dissociate the FS-activin complex, these activin assays measure only free activin, as manifest by compromised analytical recovery of activin in the presence of pure FS, serum or hFF. This complication makes interpretation of changes in activin concentra¬ tions with physiological state or pathophysiological condition difficult because levels of either activin or FS could be changing to produce a change in free activin concentration....