The purpose of this investigation was to assess the accuracy of the COSMED K4 b2 portable metabolic measurement system against the criterion Douglas bag (DB) method. During cycle ergometry on consecutive days, oxygen consumption (VO2), carbon dioxide production (VCO2), minute ventilation (VE), and respiratory exchange ratio (R) were measured at rest and during power outputs of 50, 100, 150, 200, and 250W. No significant differences (P > 0.05) were observed in VO2 between the K4 b2 and DB at rest and at 250W. Though the K4 b2 values were significantly higher (P<0.05) than DB values at 50, 100, 150, and 200 W, the magnitude of these differences was small (0.088, 0.092, 0.096, and 0.088 L x min(-1), respectively). VCO2 and VE values from the K4 b2 were significantly lower than the DB at 200 and 250 W, while no significant differences were observed from rest through 150W. The slight overestimation of VO2 (50-200 W) combined with the underestimation of VCO2 (200 and 250W) by the K4 b2 resulted in significantly lower R values at every stage. These findings suggest the COSMED K4 b2 portable metabolic measurement system is acceptable for measuring oxygen uptake over a fairly wide range of exercise intensities.