The purpose of this study was to test the CFK equation for its prediction of the rate of formation of carboxyhemoglobin (HbCO) in exercising humans by use of measured values of the respiratory variables and to characterize the rate of appearance of HbCO with frequent blood sampling. Ten nonsmoking male subjects were exposed to carbon monoxide (CO) on two separate occasions distinguished by the level of activity. Steady-state exercise was conducted on a cycle ergometer at either a low (approximately 45 W) or moderate (approximately 90 W) power output. Each experiment began with an exposure of 3,000 ppm CO for 3 min during a rest period followed by three intermittent exposures ranging from 3,000 ppm CO for 1 min at low exercise to 667 ppm CO for 3 min at moderate exercise. Increases in HbCO were normalized against predicted values to account for individual differences in the variables that govern CO uptake. No difference in the normalized uptake of CO was found between the low- and moderate-exercise trials. However, the CFK equation underpredicted the increase in HbCO for the exposures at rest and the first exposure at exercise, whereas it overpredicted for the latter two exposures at exercise. The net increase in HbCO after all exposures (approximately 10% HbCO) deviated by less than 1% HbCO between the measured and predicted values. The rate of appearance of HbCO fits a sigmoidal shape with considerable overshoot at the end of exposure. This can be explained by delays in the delivery of CO to the blood sampling point (dorsal hand vein) and by a relatively small blood circulation time compared with other regions of the body. A simple circulation model is used to demonstrate the overshoot phenomenon.