This study evaluated the predictive ability of the modified Canadian Aerobic Fitness Test (mCAFT) on an independent sample of subjects and compared it to the original Canadian Aerobic Fitness Test (CAFT). Male and female subjects (n = 154), 15-69 yrs of age, performed the mCAFT and a maximal treadmill test. VO(2)max scores predicted from the mCAFT equation did not differ significantly from those measured during the treadmill lest, whether the sample was analysed overall or categorized by sex, whereas the CAFT (Jette) equation resulted in VO(2)max scores that were significantly lower than either the measured values or those predicted using the mCAFT. The strength of the linear relationship between predicted and measured VO(2)max scores is the same for both the mCAFT (r = 0.88) and CAFT (r = 0.89). However, the mCAFT results in a lower mean square error (37.0 for mCAFT vs. 63.3 for CAFT) and thus is an improvement over the CAFT prediction.