The purpose of the present study was to examine, in highly trained cyclists, the reproducibility of cycling time to exhaustion (T(max)) at the power output equal to that attained at peak oxygen uptake (.VO2peak) during a progressive exercise test. Forty-three highly trained male cyclists (M +/- SD; age = 25 +/- 6 yrs; weight = 75 +/- 7 kg; .VO2peak = 64.8 +/- 5.2 ml.kg-1.min-1) performed two T(max) tests one week apart. While the two measures of T(max) were strongly related (r = 0.884; p < 0.001), T(max) from the second test (245 +/- 57 s) was significantly higher than that of the first (237 +/- 57 s; p = 0.047; two-tailed). Within-subject variability in the present study was calculated to be 6 +/- 6%, which was lower than that previously reported for T(max) in sub-elite runners (25%). The mean T(max) was significantly (p < 0.05) related to both the second ventilatory turnpoint (VT(2); r = 0.38) and to .VO2peak (r = 0.34). Despite a relatively low within-subject coefficient of variation, these data demonstrate that the second score in a series of two T(max) tests may be significantly greater than the first. Moreover, the present data show that T(max) in highly trained cyclists is moderately related to VT(2) and .VO2peak.