Submaximal cycle ergometry is routinely used for the prediction of mode specific maximal oxygen consumption (VO2peak). Such testing is almost universally performed at a cycling cadence of 50 rpm despite the finding that a cadence of approximately 80 rpm yields greater economy of effort among individuals with cycling experience. We sought to determine if 50 or 80 rpm were superior cadences for predicting VO2peak in individuals with or without cycling experience. Thirty experienced (EXP: 16 male and 14 female) and 28 nonexperienced (NEXP: 15 male and 13 female) subjects between the ages of 18-40 yr completed two incremental exercise tests on a cycle ergometer, one at 50 rpm and one at 80 rpm. in random order. VO2peak was predicted from submaximal data according to the method of the ACSM. There was no difference between actual VO2peak at 50 rpm and at 80 rpm. Correlations between predicted and actual VO2peak were 0.79 at 50 rpm and 0.81 at 80 rpm for all subjects. No significant differences in correlations were observed between groups or between cadences. However, the predicted values of VO2peak overestimated the actual values in both groups. Therefore, we conclude that cadences of 50 and 80 rpm have similar validity in cycle ergometry prediction of VO2peak, but the ACSM protocol overestimates VO2peak.