The purpose of this study was to compare the effect of 2 training programs differing in the relative contribution of training volume, clearly below vs. within the lactate threshold/maximal lactate steady state region on performance in endurance runners. Twelve subelite endurance runners (who are specialists in track events, mostly the 5,000-m race usually held during spring-summer months and who also participate in cross-country races [9-12 km] during fall and winter months) were randomly assigned to a training program emphasizing low-intensity (subthreshold) (Z1) or moderately high-intensity (between thresholds) (Z2) training intensities. At the start of the study, the subjects performed a maximal exercise test to determine ventilatory (VT) and respiratory compensation thresholds (RCT), which allowed training to be controlled based on heart rate during each training session over a 5-month training period. Subjects performed a simulated 10.4-km cross-country race before and after the training period. Training was quantified based on the cumulative time spent in 3 intensity zones: zone 1 (low intensity; <VT), zone 2 (moderate intensity; between VT and RCT), and zone 3 (high intensity; >RCT). The contribution of total training time spent in zones 1 and 2 was controlled to have relatively more low-intensity training in Z1 (80.5 +/- 1.8% and 11.8 +/- 2.0%, respectively) than in Z2 (66.8 +/- 1.1% and 24.7 +/- 1.5%, respectively), whereas the contribution of high-intensity (zone 3) training was similar (8.3 +/- 0.7% [Z1] and 8.5 +/- 1.0% [Z2]). The magnitude of the improvement in running performance was significantly greater (p = 0.03) in Z1 (-157 +/- 13 seconds) than in Z2 (-121.5 +/- 7.1 seconds). These results provide experimental evidence supporting the value of a relatively large percentage of low-intensity training over a long period ( approximately 5 months), provided that the contribution of high-intensity training remains sufficient.