Objective: The aims were, first, to detect and quantify the release of ATP from human erythrocytes in response to a brief exposure to a hypoxic/hypercapnic environment, similar to that found in vigorously exercising skeletal or cardiac muscle; and second, to explore the mechanism of ATP release in response to hypoxia. Methods: Washed human erythrocytes suspended in Krebs-Henseleit solution were exposed for 50 s to an atmosphere of ∼ 8.0 kPa Pco₂ and 2.7 kPa Po₂; ATP released into the suspension was assayed using the firefly luminescence technique. Samples of human blood were obtained by venepuncture of the median cubital vein from male and female volunteers ranging in age from 21 to 74 years. Anticoagulation was with EDTA. Results: A background of 0.49 × 10⁶ (SEM 0.037 × 10⁶) ATP molecules·cell⁻¹ was attributed to spontaneous haemolysis of 1% of the cell population, as estimated by levels of haemoglobin measured in the suspension fluid. When the erythrocytes were exposed to the hypoxic/hypercapnic gas mixture at 37°C the ATP concentration in the suspension fluid rose to 2.67 × 10⁶ (0.27 × 10⁶) molecules·cell⁻¹. An efflux rate of 276(37) molecules·μ⁻²·s⁻¹ was calculated. The hypoxia induced ATP release was blocked in three different ways: first, by application of 50 μM of the specific band 3 anion channel blocking agents niflumic acid (a translocation inhibitor), DIDS (a transport site inhibitor), or dipyridamole (a channel blocker); secondly, by replacement of extracellular chloride and bicarbonate with the impermeant anion methanesulphonate; and thirdly, by application of 5 nM of the nucleoside transport blocker nitrobenzylthioinosine. None of these blocking techniques affected the background levels of ATP attributed to haemolysis. Conclusions: A situation of hypoxia/hypercapnia, such as would be found in exercising muscle, induces release of ATP from the erythrocyte via the plasma membrane protein moiety known as band 4.5 (a nucleoside transporter) and electrical balance across the erythrocyte membrane is maintained by the simultaneous influx of extracellular chloride and/or bicarbonate via the plasma membrane protein known as band 3 (anion channel). The circulation of erythrocytes into a region of hypoxia in vivo could promote an increase in local blood flow through release of endothelium dependent relaxing factor in response to released ATP.