Human erythroid burst-promoting activity (BPA) of recombinant growth factors and crude materials, of media conditioned by omentum tissue (OMCM), and of media conditioned by the bladder carcinoma cell line (HTB9CM) was measured by three different culture methods. Using the two-stage culture method, significant activity was shown in OMCM (137%-329% of the control), HTB9CM (102%-333%), recombinant human (rh) granulocyte-macrophage colony-stimulating factor (rhGM-CSF) (179%-220%), rh interleukin 3 (rhIL-3) (232%-676%), and rh insulin-like growth factor 1 (rh IGF-1) (106%-175%), whereas there was no significant increase in the number of erythroid bursts by the same additives when the one-stage culture or the delayed erythropoietin method was employed. Linear dose-response curves were observed in the tested range of rhIL-3 and rhGM-CSF. We also observed that 1) a larger amount of rhGM-CSF was required for the optimal stimulation of erythroid burst-forming units (BFU-E) than for the optimal stimulation of granulocyte-macrophage colony-forming units (CFU-GM), and 2) even the maximum dose of rhGM-CSF increased erythroid bursts to a lesser extent than was possible by the addition of rhIL-3. The former results implies that BPA is not the major activity of GM-CSF, and the latter result, although it is not conclusive, suggests that the GM-CSF-responsive BFU-E represent only a subset population of BFU-E responsive to IL-3. The two-stage culture is a useful assay method for screening BPA in biological materials with respect to accuracy, dose responsiveness, and reproducibility.