The aims of this study were to document the effect terrain has on the physiological responses and work demands (power output) of riding a typical mountain bike cross-country course under race conditions. We were particularly interested in determining whether physiological measures relative to mass were better predictors of race performance than absolute measures. Eleven A-grade male cross-country mountain bike riders (VO2max 67.1 +/- 3.6 ml x kg(-1) x min(-1)) performed 2 tests: a laboratory-based maximum progressive exercise test, and a 15.5-km (six 2.58-km laps) mountain bike cross-country time trial. There were significant differences among the speed, cadence, and power output measured in each of 8 different terrain types found in the cross-country time trial course. The highest average speed was measured during the 10-15% downhill section (22.7 +/- 2.6 km x h(-1)), whereas the cadence was highest in the posttechnical flat sections (74.3 +/- 5.6 rpm) and lowest on the 15-20% downhill sections (6.4 +/- 12.1 rpm). The highest mean heart rate (HR) was obtained during the steepest (15-20% incline) section of the course (179 +/- 8 b x min(-1)), when the power output was greatest (419.8 +/- 39.7 W). However, HR remained elevated relative to power output in the downhill sections of the course. Physiological measures relative to total rider mass correlated more strongly to average course speed than did absolute measures (peak power relative to mass r = 0.93, p < 0.01, vs. peak power r = 0.64, p < 0.05; relative VO2max r = 0.80, p < 0.05, vs. VO2max r = 0.66, p < 0.05; power at anaerobic threshold relative to mass r = 0.78, p < 0.05, vs. power at anaerobic threshold r = 0.5, p < 0.05). This suggests that mountain bike cross-country training programs should focus upon improving relative physiological values rather than focusing upon maximizing absolute values to improve performance.