A comparative study was undertaken in order to assess the hematotoxic effects of hydroquinone (HQ), 1,4-benzoquinone (BQ), and doxorubicin (DX) on mouse and human bone marrow (BM) cells. Initial experiments indicated that the inhibitory effects of near-ambient pO2 and HQ on granulocyte/macrophage colony-stimulating factor (GM-CSF)-induced colony formation were additive. Thus, subsequent experiments were done under conditions of continuous toxicant exposure in complete medium at physiological temperature and O2 partial pressure. Viability was measured 24 hr after exposure, and at the concentrations tested, HQ was less cytotoxic than BQ. DX did not exhibit significant cytotoxicity at the concentrations used. Both HQ and BQ were slightly more cytotoxic to mouse BM cells than to human BM cells. Dose-response analyses of HQ, BQ, or DX inhibition of GM-CSF-induced proliferative and colony-forming responses indicated that murine GM progenitors were significantly less sensitive to HQ than to the majority of myeloid BM cells that proliferated in response to GM-CSF. This preferential resistance of GM progenitors to HQ was not observed when human BM cells were used. HQ was somewhat more inhibitory to human than to mouse GM-CSF responses. Inhibition of GM-CSF-induced responses by BQ correlated closely with cytotoxicity, and DX was 1000-fold more inhibitory to GM-CSF-induced proliferative and colony-forming responses than either HQ or BQ. Again, DX appeared to be slightly more inhibitory to human BM cells than to mouse BM cells. Purified human hematopoietic progenitor cells (HPCs) were also used in the dose-response analyses of HQ, BQ, or DX inhibition of GM-CSF-induced proliferative and colony-forming responses. Inhibition of GM-CSF-induced HPC responses by HQ, BQ, and DX was very similar to that obtained when BM mononuclear cells were used, suggesting that the human HPC is a target for the direct effects of HQ, BQ, and DX.