Aim: To investigate the effect of task familiarisation on the spontaneous pattern of energy expenditure during a series of 2000 m cycling time trials (TTs). Method: Nine trained males completed three 2000 m TTs on a Velotron cycling ergometer. To examine pacing strategy, the data were assigned to 250 m ?bins,? with the pattern of aerobic and anaerobic energy expenditure calculated from total work accomplished and gas-exchange data. Results: There were no significant differences between trials in performance times (191.4 (SD 4.3), 189.4 (4.6), 190.1 (5.6) s), total aerobic (58.3 (2.7), 58.4 (3.1), 58.0 (3.4) kJ) and total anaerobic energy expenditure (16.4 (3.3), 17.3 (2.8), 16.5 (3.1) kJ). Pacing strategy in the second and third TT differed from the first TT in that a lower power output was adopted during the first 500 m, enabling a higher power output during the final 750 m of the TT. This adjustment in the pattern of energy expenditure was mediated by an alteration in the pattern of anaerobic energy expenditure, which paralleled changes in total energy expenditure. Furthermore, participants retained an anaerobic energy ?reserve? enabling an end-spurt during the second and third trials. Conclusion: Small modifications to the pacing strategy are made following a single bout of exercise, primarily by altering the rate of anaerobic energy expenditure. This may have served to prevent critical metabolic disturbances. The alteration in pacing strategy following the first exercise bout is compatible with a complex intelligent regulatory system.