The effect of active and passive recovery on repeated-sprint swimming bouts was studied in eight elite swimmers. Participants performed three trials of two sets of front crawl swims with 5 min rest between sets. Set A consisted of four 30-s bouts of high-intensity tethered swimming separated by 30 s passive rest, whereas Set B consisted of four 50-yard maximal-sprint swimming repetitions at intervals of 2 min. Recovery was active only between sets (AP trial), between sets and repetitions of Set B (AA trial) or passive throughout (PP trial). Performance during and metabolic responses after Set A were similar between trials. Blood lactate concentration after Set B was higher and blood pH was lower in the PP (18.29 +/- 1.31 mmol x l(-1) and 7.12 +/- 0.11 respectively) and AP (17.56 +/- 1.22 mmol x l(-1) and 7.14 +/- 0.11 respectively) trials compared with the AA (14.13 +/- 1.56 mmol x l(-1) and 7.23 +/- 0.10 respectively) trial (P < 0.01). Performance time during Set B was not different between trials (P > 0.05), but the decline in performance during Set B of the AP trial was less marked than in the AA or PP trials (main effect of sprints, P < 0.05). Results suggest that active recovery (60% of the 100-m pace) could be beneficial between training sets, and may compromise swimming performance between repetitions when recovery durations are short (< 2 min).