Seven male subjects performed 15 x 40m sprints, on three occasions, with rest periods of either 120 s (R120), 60 s (R60) or 30 s (R30) between each sprint. Sprint times were recorded with four photo cells placed at 0, 15, 30 and 40 m. The performance data indicated that whereas running speed over the last 10 m of each sprint decreased in all three protocols (after 11 sprints in R120, 7 sprints in R60 and 3 sprints in R30), performance during the initial acceleration period from 0-15 m was only affected with the shortest rest periods increasing from (mean +/- SEM) 2.58 +/- .03 (sprint 1) to 2.78 +/- .04 s (spring 15) (p < .05). Post-exercise blood lactate concentration was not significantly different in R120 (12.1 +/- 1.3 mmol.l-1) and R60 (13.9 +/- 1.2 mmol.l-1), but a higher concentration was found in R30 (17.2 +/- .7 mmol.l-1) (p < .05). After 6 sprints there was no significant difference in blood lactate concentration with the different recovery durations, however, there were significant differences in sprint times at this point, suggesting that blood lactate is a poor predictor of performance during this type of exercise. Although the work bouts could be classified primarily as anaerobic exercise, oxygen uptake measured during rest periods increased to 52, 57 and 66% of maximum oxygen uptake in R120, R60 and R30, respectively. Evidence of adenine nucleotide degradation was provided by plasma hypoxanthine and uric acid concentrations elevated post-exercise in all three protocols. Post-exercise uric acid concentration was not significantly affected by recovery duration.(ABSTRACT TRUNCATED AT 250 WORDS)