We tested the response characteristics of a dual transcutaneous (tc) PO2/PCO2 monitoring system in healthy subjects who breathed various gas mixtures, and we compared steady-state tc readings to simultaneous arterial blood gas analysis in 20 stable respiratory outpatients. The electrodes were simple to apply, required very little skin preparation, and had trivial signal drift. In healthy subjects, tcPCO2 lag time during CO2 rebreathing was 16.8 seconds, with a 90 percent response time of 77.9 seconds after CO2 breathing was discontinued. The 90 percent response times of the O2 electrode when subjects breathed a hypoxic mixture was 257 seconds after a lag of 31 seconds. When inhaled gas mixtures were changed from hypoxia to room air, the lag time was shorter (12.5 seconds), but 90 percent response time exceeded 5 minutes. In stable patients with respiratory disease, tcPCO2 and tcPO2 were linearly related to PaCO2 (range, 19 to 53 mm Hg) and PaO2 (range, 45 to 99 mm Hg), respectively (tcPCO2 = 1.4 PaCO2-9.44, with r = 0.90 and SEE = 5.35 mm Hg; tcPO2 = 0.56 PaO2 + 20.4, with r = 0.53 and SEE = 11.7 mm Hg). We conclude that the response of the dual transcutaneous monitoring system is more rapid for the CO2 than the O2 electrode and may be rapid enough to be useful in some clinical settings; however, the O2 system fails to offer the response characteristics and accuracy that would allow it to be substituted for arterial gas tensions in unstable clinical situations.