At present, one of the most sensitive assays for human erythropoietin (Ep) is a bioassay that measures the Ep-dependent proliferation of spleen cells from phenylhydrazine-treated mice after 24 hours in culture. We describe how this assay can be used as the basis of a very sensitive method for detecting mouse antibodies to biologically active human Ep. In this procedure, microtiter wells are first coated with goat anti-mouse Ig antibody, then treated with mouse antibodies (serum or hybridoma culture supernatants), and finally incubated with a fixed amount of pure human Ep. Specific binding of anti-Ep antibodies is detected by adding spleen cells from phenylhydrazine-treated mice to the wells and measuring the ability of the cells to incorporate 3H-thymidine 24 hours later. This bioimmunosorbent assay (BISA) revealed the presence of anti-EP antibodies in sera from mice immunized with either pure human urinary Ep or a synthetic dodecapeptide corresponding to the aminoterminal region of Ep and in the culture supernatants from three of eight stable anti-Ep antibody-producing hybridoma cell lines that we have isolated. The three monoclonal antibodies showed similar reactivities in the BISA, but showed different affinities for Ep, with Kd values of approximately 0.7, 8, and 240 nmol/L, respectively. Further studies showed that all antibodies were capable of neutralizing Ep bioactivity and of binding 125I-labeled Ep in a radioimmunosorbent assay (RIA) but were virtually unreactive to Ep adsorbed to the bottom of enzyme-linked immunosorbent assay (ELISA) wells. Our results suggest that the BISA strategy may be an important complement to conventional RIA and ELISA techniques for identification of monoclonal antibodies specific for biologically active growth factors.