We designed two experiments to investigate the relationship between ventilation (VE) and CO2 output (VCO2) during exercise under the conditions of exercising different limbs, the arms as opposed to the legs (experiment 1), and of different physical training states after undergoing standard exercise training for 90 d (experiment 2). Six healthy young subjects underwent submaximal ramp exercise at an incremental work rate of 15 W/min for the arm and leg, and 11 healthy middle-aged subjects underwent an incremental exercise test at the rate of 30 W/3 min before and after exercise training. We measured pulmonary breath-by-breath VE, VCO2, oxygen uptake (VO2), tidal volume (VT), breathing frequency (bf), and end-tidal O2 and CO2 pressures (PETO2, PETCO2) via a computerized metabolic cart. In experiment 1, arm exercise produced significantly greater VE than did leg exercise at the same work rates, as well as significantly higher VO2, VCO2, and bf. The slopes of the regression lines in the VE-VCO2 relationship were not significantly different: the values were 27.8 +/- 2.1 (SD) during the arm exercise, and 25.3 +/- 3.9 during the leg exercise, with no differences in their intercepts. In experiment 2, the VO2, VCO2, and VE responses at the same work rates were similar in both before and after the 90-d exercise training, whereas the heart rate (HR) and mean blood pressure (MBP) were significantly reduced after training. Exercise training did not alter the VE-VCO2 relationship, the slope of which was 31.9 +/- 4.9 before exercise training and 34.2 +/- 4.4 after exercise training. We concluded that the VE-VCO2 relationship during exercise is unaltered, independent of not only working muscle regions but also exercise training states.