The purpose of this study was to evaluate the effects of 6 weeks of low-intensity continuous exercise training (CE; 40 min at 50% VO2peak, 3 days/week) and high-intensity interval exercise training (IE: 10 x 2 min at VO2peak, 3 days/week) on the parameters of the power-endurance time relationship for cycle ergometry. The hyperbolic relationship between power and endurance time was linearized by expressing the power against the inverse of time, as described by Whipp et al. (22). This model consists of two parameters: theta f, a fatigue threshold reflecting the capability for sustained aerobic power, and W', a constant postulated to reflect a finite energy store (i.e., those factors comprising the O2 deficit: Phosphagen stores, anaerobic glycogenolysis, myoglobin O2 stores). Prior to training, test-retest reliability coefficients (r2) for theta f and W' were 0.92 and 0.62, respectively (P less than 0.01). Training resulted in significant (P less than 0.01) increases in theta f for both CE [27 +/- 3 W (13.4%) increase] and IE [33 +/- 5 W (15.0%) increase], with no difference between groups. Increases in theta f were not dependent upon improvements in VO2peak. W' was not changed significantly in either group after training. However, a significant negative correlation between the training-induced changes in theta f and W' (R = 0.76; P less than 0.01) was obtained. The minimum intensity threshold for exercise training necessary to elicit increases in theta f has yet to be identified, but is at least as low as 50% of VO2peak.