The response of the hexose monophosphate shunt in erythrocytes was studied with the ionization chamber-electrometer apparatus to measure continuously 14CO2 derived from 14C-labeled substrates. The effect of methylene blue at high (0.1 mM) and low (1 muM) concentrations was evaluated under different gas mixtures; air, carbon monoxide, and 6% carbon monoxide in air. The latter gas mixture results in nearly 100% carboxyhemoglobin but provides a physiologic partial pressure of oxygen. The extent to which pentose is recycled through the shunt in response to methylene blue stimulation was examined with radioactive glucose substrates labeled on the first, second, and third carbon positions. Generation of hydrogen peroxide after stimulation of erythrocytes with methylene blue was evaluated by the catalase-aminotriazole trapping technique, [14C]formate oxidation, and oxidation of reduced glutatione. Stimulation of the shunt with 1 muM methylene blue was markedly impaired in the absence of oxyhemoglobin, but stimulation with 0.1 mM methylene blue was only slightly impaired under the carbon monoxide-air mixture. The higher concentration of methylene blue produced evidence of hydrogen peroxide generation of all three techniques. Despite the evidence for the involvement of oxygen, oxyhemoglobin, and hydrogen peroxide in the response to methylene blue, cells containing methemoglobin induced by sodium nitrite or from a patient with congenital methemoglobinemia responded normally to methylene blue in the absence of oxygen. These experiments indicate that the reactions induced by methylene blue in erythrocytes are more complex than generally thought and that high concentrations are associated with production of peroxide.