The intensity of active recovery may be important for performance because it is related to the energy spent within the interval period between sprints. Different "ideal" recovery exercise intensities have been reported for cycling (Belcastro and Bonen, 1975, Bonen et al., 1978), running (Hermansen and Stensvold, 1972, Gisolfi et al., 1966) or swimming (Cazorla et al., 1983, McMaster et al., 1989). When comparing the different modes of exercise, it is likely that lactate removal during active recovery may be faster after swimming compared to running following exercise that had increased the blood lactate to similar concentrations (Denadai et al., 2000). Lactate removal rate after active recovery was higher during swimming (5.3%min-1;Cazorla et al., 1983) compared to cycling (2.9%min-1 at 29% of VO2max; McGrail et al., 1978, 3.2%min-1 at 32% of VO2max; Belcastro and Bonen, 1975) or running (4.5%min-1 at 63% of VO2max; Hermansen and Stensvold, 1972). It is suggested that the ideal intensity should not exceed the individual "anaerobic threshold" (Stamford et al., 1981). It has been reported that the most effective intensity of active recovery for lactate removal is related to the individual "anaerobic threshold", suggesting that an intensity of 10% of VO2 max below the "anaerobic threshold" is the most efficient (McLellan and Skinner 1982).

However, there is evidence that athletes are able to self-select the intensity of active recovery, and no difference was observed in the lactate removal between the self-selected and the "ideal" active recovery intensity (Bonen and Belcastro, 1976; Cazorla et al., 1983).
Even though the reported intensities of active recovery are very useful in making comparisons in the scientific literature, they offer no help to the coaches, since they usually have no data that allow them to express swimming, running or cycling speed during a training session as a percentage of VO2max.
Expression of active recovery as percentage of the speed attained in a race distance may be more helpful to coaches. For example, swimming speed corresponding to 60-70 % of the 100 m speed (55 to 73 % of VO2max) was effective in faster lactate removal than passive rest (Cazorla et al., 1983). It was reported that 65% of maximum velocity of 200 yd swimming was the most efficient recovery intensity; however, the velocity of 55 or 75% was equally effective for lactate removal (McMaster et al., 1989). The self-selected pace of active recovery in the study of Reaburn and Mackinnon (1990) corresponded to 63 % of the 100 m swimming speed and significantly improved the half time of lactate removal compared to passive recovery. The faster lactate removal during running has been reported to correspond to velocity at the ventilatory threshold or below the ventilatory threshold in triathletes (Baltari et al., 2005) and soccer players (Baltari et al., 2004).
Unfortunately, besides swimming there are no data to report the running or cycling intensities as a percentage of performance time or speed. In summary, the intensity of active recovery should be below the intensity that increases the lactate production within the muscle. A question which arises however is whether the intensity below the lactate or ventilatory threshold that maximizes blood lactate removal during active recovery is also the most appropriate for performance recovery during a subsequent exercise bout.