The aim of this study was to compare the results from a Cooper walk run test, a multistage shuttle run test, and a submaximal cycle test with the direct measurement of maximum oxygen uptake on a treadmill. Three predictive tests of maximum oxygen uptake--linear extrapolation of heart rate of VO2 collected from a submaximal cycle ergometer test (predicted L/E), the Cooper 12 min walk, run test, and a multi-stage progressive shuttle run test (MST)--were performed by 22 young healthy males (mean(s.d.) age 22.1 (2.4) years; body mass 72.4(8.9kg)) and the values compared to those obtained by direct measurement on a maximal treadmill test. All of the subjects were regular exercisers. The mean(s.d.) from the various tests in ml.kg-1.min-1 were as follows: treadmill 60.1(8.0), Cooper 60.6(10.3), MST 55.6(8.0), and predictedL/E 52.0(8.4). The Cooper test had a correlation with the treadmill test of 0.92, while the MST and the predictedL/E had correlations of 0.86 and 0.76 respectively. Both the MST and predictedL/E showed systematic underprediction of the treadmill value. On average, the MST was 4.5 ml.kg-1.min-1 (s.e. 0.9) lower than the treadmill VO2max while the predictedL/E was 7.8 ml.kg-1. min-1 (s.e. 1.4) lower than the treadmill VO2max. These findings indicate that, for the population assessed, the Cooper walk run test is the best predictor of VO2max among the three tests.