Twenty-three subjects isokinetically trained the right and left quadriceps femoris, three times per week for 16 weeks; one group (n = 13) trained at an angular velocity of 4.19 rad.s-1 and a second group (n = 10), at 1.05 rad.s-1. A control group (n = 10) performed no training. Isometric endurance time at 60% quadriceps maximum voluntary contraction (MVC), mean power output and work done (W) during all-out cycling, and the muscle buffer value (B) and carnosine concentration of biopsy samples from the vastus lateralis, were all assessed before and after training. The two training groups did not differ significantly from each other in their training response to any of these variables (P < 0.05). No significant difference in either 60% MVC endurance time or impulse [(endurance time x force) at 60% MVC] was observed for any group after the 16 week period (P > 0.05). However, the post-training increase (9%) in W during high-intensity cycling was greater in the training group than in the control group (P = 0.04). Neither B nor carnosine concentration showed any significant change following training (P = 0.56 and P = 0.37, respectively). It is concluded that 16 weeks of isokinetic training of the knee extensors enables subjects to do more work during high-intensity cycling. Although the precise adaptations responsible for the improved performance have yet to be identified, they are unlikely to include an increase in B.Twenty-three subjects isokinetically trained the right and left quadriceps femoris, three times per week for 16 weeks; one group (n = 13) trained at an angular velocity of 4.19 rad.s-1 and a second group (n = 10), at 1.05 rad.s-1. A control group (n = 10) performed no training. Isometric endurance time at 60% quadriceps maximum voluntary contraction (MVC), mean power output and work done (W) during all-out cycling, and the muscle buffer value (B) and carnosine concentration of biopsy samples from the vastus lateralis, were all assessed before and after training. The two training groups did not differ significantly from each other in their training response to any of these variables (P < 0.05). No significant difference in either 60% MVC endurance time or impulse [(endurance time x force) at 60% MVC] was observed for any group after the 16 week period (P > 0.05). However, the post-training increase (9%) in W during high-intensity cycling was greater in the training group than in the control group (P = 0.04). Neither B nor carnosine concentration showed any significant change following training (P = 0.56 and P = 0.37, respectively). It is concluded that 16 weeks of isokinetic training of the knee extensors enables subjects to do more work during high-intensity cycling. Although the precise adaptations responsible for the improved performance have yet to be identified, they are unlikely to include an increase in B.