Intraperitoneal doses of 4-dimethylaminophenol hydrochloride (DMAP), hydroxylamine hydrochloride (H2NOH) and sodium nitrite (NaNO2) were found where each converted a maximum of about 37% of the total circulating hemoglobin in mice to methemoglobin. Those doses in mmol/kg were: 0.29 for DMAP, 1.1 for HENOH, and 1.1 for NaNO2. For DMAP and H2NOH the peak was sharp and at about 7 min after injection whereas for NaNO2 the peak was much broader and at about 40 min. The i.p. LDs0's in mmol/kg were: 0.48 for DMAP, 1.8 for HENOH and 2.3 for NaNO2. When mice pretreated with each of the methemoglobin-generating agents were challenged with sodium cyanide, the ratios of the LDs0's in protected mice to those in control mice (protection index, PI) were 1.5 for H2NOH, 2.0 for DMAP and 3.1 for NaNO2. When sodium thiosulfate was also given in combination with each of the three methemoglobin-generating agents, the protective effect was at least additive. The PI against sodium sulfide was also significantly greater in mice pretreated with NaNO2 than in mice given H2NOH. Methemoglobins generated from human and mouse hemoglobins by either NaNO2 or by HENOH had identical binding affinities (dissociation constants) for cyanide. When human red cells containing methemoglobin generated by exposure to either NaNO2 or HENOH were injected into the peritoneal cavity of mice and then followed by cyanide challenges, there was no difference in the PI for the two kinds of methemoglobin. Not only was the PI the same in each case with human cells, but it was also identical with that in mice given NaNO2 systemically to generate the same total amount of methemoglobin. The difference in PI between NaNO2 and H2NOH (or DMAP) in mice appears to be related to the high rate of methemoglobin reductase activity in mouse RBC. It appears likely that cyanmethemoglobin is a substrate for mouse methemoglobin reductase activity, and that NaNO2 is an inhibitor of mouse methemoglobin reductase. No differences in cyanide antagonism between NaNO2 and H2NOH would be anticipated in humans because of the slow rates of methemoglobin reduction in human red cells.