This study evaluated the validity of the desktop CardioCoach metabolic system to measure VO2max and VEmax. Sixteen subjects (mean age = 19.5 +/- 3.2 years) completed 2 maximal graded exercise tests following the same protocol before and after 7 and 14 weeks of endurance training. Subjects' VO2max and VEmax were measured by either the CardioCoach or the ParvoMedics TrueOne 2400 metabolic measurement system (TrueOne). An alpha level of significance of p < 0.05 was maintained for all statistical analyses. The time to test completion and the final treadmill grade of the exercise tests performed by both the CardioCoach and the TrueOne increased over the 3 testing periods, confirming an improvement in cardiorespiratory fitness resulting from the 14 weeks of training. A linear growth curve analysis indicated that there were statistically significant differences between VO2max (ml x kg(-1) x min(-1)) as measured by the TrueOne and the CardioCoach before (44.4 +/- 5.0 and 49.3 +/- 5.4) and after 7 weeks (46.0 +/- 5.2 and 48.2 +/- 5.4) of training but not after 14 weeks of training (47.8 +/- 5.6 and 48.4 +/- 5.2). Significant differences also existed in VEmax (L x min(-1)) as measured by the TrueOne and the CardioCoach before (76.8 +/- 17.7 and 71.9 +/- 13.7), after 7 weeks (81.4 +/- 16.2 and 72.8 +/- 14.1), and after 14 weeks (86.8 +/- 19.4 and 74.2 +/- 13.1) of training. Although significant growth of VO2max (0.24 ml x kg(-1) x min(-1) x wk(-1)) and VEmax (0.71 L x min(-1) x wk(-1)) was measured by the TrueOne over 14 weeks of training, the CardioCoach was unable to detect growth in VO2max (-0.02 ml x kg(-1) x min(-1) x wk(-1)) or VEmax (0.17 L x min(-1) x wk(-1)). This study indicates that the CardioCoach did not accurately measure or monitor changes in VO2max or VEmax resulting from training.