Exercise time to exhaustion (TE) is commonly used to evaluate the success or failure of such treatments as endurance training programs or nutritional supplements. The present study determined the variability of TE during submaximal exercise at 80% VO2max. Fifteen males performed cycle exercise to exhaustion on five occasions at the same time of day with a minimum of 72 hrs between sessions. There was no difference in TE (0.1 > p > 0.05) among the trials, with values ranging from 14.4 +/- 1.1 min for Test 1 to 18.2 +/- 2.4 min during the final test. Substantial variability in TE over the five trials was observed among subjects with coefficients of variation (CV) ranging from 2.8 to 31.4%. Subjects were divided into two groups using the median CV for TE. For the low CV group (n = 8), TE was significantly increased during Test 3 (14.9 +/- 1.3 min) compared with Test 1 (12.8 +/- 1.0 min) and Test 5 (12.5 +/- 1.2 min). For the high CV group (n = 7), TE was increased during Test 5 (24.7 +/- 3.7 min) compared with the other tests (18.5 +/- 2.2 min). CV for VO2, VE, pH, Pco2, and rectal temperature were less than 5% and did not differ between groups. Post hoc power calculations revealed that if all subjects were considered as one group, sample size would have to increase to 40 to increase the power to 0.8. Due to the variability in TE that may be observed with males of average fitness, it is concluded that TE should not be the only dependent measure used to evaluate treatment effects during submaximal exercise.