The effects of active recovery on metabolic and cardiorespiratory responses and power output were examined during repeated sprints. Male subjects (n = 13) performed two maximal 30-s cycle ergometer sprints, 4 min apart, on two separate occasions with either an active [cycling at 40 (1)% of maximal oxygen uptake; mean (SEM)] or passive recovery. Active recovery resulted in a significantly higher mean power output (W) during sprint 2, compared with passive recovery [W] 603 (17) W and 589 (15) W, P < 0.05]. This improvement was totally attributed to a 3.1 (1.0)% higher power generation during the initial 10 s of sprint 2 following the active recovery (P < 0.05), since power output during the last 20 s sprint 2 was the same after both recoveries. Despite the higher power output during sprint 2 after active recovery, no differences were observed between conditions in venous blood lactate and pH, but peak plasma ammonia was significantly higher in the active recovery condition [205 (23) vs 170 (20) mumol .l-1; P < 0.05]. No differences were found between active and passive recovery in terms of changes in plasma volume or arterial blood pressure throughout the test. However, heart rate between the two 30-s sprints and oxygen uptake during the second sprint were higher for the active compared with passive recovery [148 (3) vs 130 (4) beats.min-1; P < 0.01) and 3.3 (0.1) vs 2.8 (0.1) l.min-1; P < 0.01]. These data suggest that recovery of power output during repeated sprint exercise is enhanced when low-intensity exercise is performed between sprints. The beneficial effects of an active recovery are possibly mediated by an increased blood flow to the previously exercised muscle.