In 18 healthy children three to 13 years of age, the transcutaneous partial pressure of carbon dioxide (PtcCO2) (Radiometer electrode) and the alveolar partial pressure of carbon dioxide (PACO2) (Beckman analyzer) were measured simultaneously during the breathing of room air and 5 percent carbon dioxide. The PtcCO2 electrode was placed on the anterior thorax and heated to 42 degrees C. The PACO2 was calculated on the 4/5 part of the carbon dioxide expired trace. Minute ventilation (VE) was measured in 11 cases. There was a significant correlation between PtcCO2 (in millimeters of mercury) and PACO2 (in millimeters of mercury) while breathing room air (PtcCO2 = 0.82 PACO2 + 19.7; r = 0.55; p less than 0.02) and while breathing 5 percent carbon dioxide (PtcCO2 = 0.77 PACO2 + 22.5; r = 0.61; p less than 0.01); however, the ratio of PtcCO2 over PACO2 was significantly lower while breathing 5 percent carbon dioxide (p less than 0.01) than while breathing room air. When considering the relationship between the increase in VE (delta VE while breathing 5 percent carbon dioxide and the changes in PACO2 (delta PACO2) or in PtcCO2 (delta PtcCO2), a significant correlation was found only between delta VE and delta PACO2, ie, delta VE = 0.41 delta PACO2 + 0.44 (r = 0.63; p less than 0.01). These results suggest that breathing carbon dioxide modified the factors acting on PtcCO2, possibly by changes in the vasomotor tone of cutaneous blood vessels. These modifications appeared to be variable from subject to subject. Therefore, we conclude that PtcCO2 does not appear to be an accurate quantitative index to assess ventilatory response to carbon dioxide.