The reliability and validity of a continuous progressive arm test, in which maximal 02 consumption (V02 max arm) is determined, were analyzed. Forty-one men (28.2 +/- 8.8 yr) performed the test twice. Eighteen additional men (22.6 +/- 5.6 yr) performed the arm test, as well as the treadmill run, in which maximal O2 consumption VO2max leg) was determined. The validity of the VO2 max arm test was computed, using VO2 max leg as a criterion for the individual's aerobic capacity. The reliability coefficients of VO2 max arm, VEmax arm, and HRmax arm were 0.94, 0.98, and 0.76, respectively, indicating a high reliability of the testmthe validity coefficient of VO2max arm was only 0.74. The regression equation of VO2max leg on VO2max arm was y = 24.4 + 0.9 +/- 4.4 (Syx). These findings indicate that, following the suggested protocol, the individual repeatedly uses the same muscles and does reach an all-out stage. However, different individuals apparently are aided by their trunk and leg muscles to different degrees, which lowers the validity of this test as a predictor of aerobic capacity.