A new in vitro bioassay for activin was developed using the mouse plasmacytoma cell line, MPC-11. Human recombinant (hr) activin A dose-dependently inhibited the proliferation of these cells, whereas a range of other factors, including inhibin, follistatin and transforming growth factor-1, -2 and -3 had no effect. Conditioned medium containing activin B induced an inhibition similar to hr-activin A. The inhibitory influence of activin A could be blocked by follistatin, but not by hr-inhibin A. This bioassay had a sensitivity for activin A of around 0·4 ng/ml, an ED 50 response of 3·5 ng/ml, and an intraassay coefficient of variation of <11%. It offers substantial advantages over existing in vitro activin bioassays in terms of ease of use, specificity and throughput. The utility of the MPC-11 bioassay was demonstrated in the purification of activin from amniotic fluid, where an almost identical profile of bioactive activin A was detected compared with the pituitary cell bioassay of activin. Bioactive activin could also be detected in unpurified ovine allantoic and amniotic fluids and bovine follicular fluid. Measuring activin in untreated and heat-treated human sera or seminal plasma was hampered by a non-specific inhibitory effect, so that several serum samples did not run parallel with the hr-activin A standard. This inhibitory effect by serum could not be overcome by addition of follistatin, suggesting it is not activin-like bioactivity. This new bioassay for activin demonstrates widespread applicability for monitoring of purified or partially purified samples during purification procedures, bioactivity measurements, receptor-binding studies and assays of cell culture medium.