We analyzed the influence of prior exercise designed to reduce predominantly muscle glycogen in either type I or II fibers on pacing and performance during a 4-km cycling time trial (TT). After preliminary and familiarization trials, in a randomized, repeated-measures crossover design, ten amateur cyclists performed: 1) an exercise designed to reduce glycogen of type I muscle fibers, followed by a 4-km TT (EX-FIB I); 2) an exercise designed to reduce glycogen of type II muscle fibers, followed by a 4-km TT (EX-FIB II) and; 3) a 4-km TT, without the prior exercise (CONT). The muscle-glycogen-reducing exercise in both EX-FIB I and EX-FIB II was performed in the evening, ∼12 h before the 4-km TT. Performance time was increased and power output (PO) was reduced in EX-FIB I (432.8±8.3 s and 204.9±10.9 W) and EX-FIB II (428.7±6.7 s and 207.5±9.1 W) compared to CONT (420.8±6.4 s and 218.4±9.3 W; P<0.01), without a difference between EX-FIB I and EX-FIB II (P>0.05). The PO was lower in EX-FIB I than in CONT at the beginning and middle of the trial (P<0.05). The mean aerobic contribution during EX-FIB I was also significantly lower than in CONT (P<0.05), but there was no difference between CONT and EX-FIB II or between EX-FIB I and EX-FIB II (P>0.05). The integrated electromyography was unchanged between conditions (P>0.05). Performance may have been impaired in EX-FIB I due a more conservative pacing at the beginning and middle, which was associated with a reduced aerobic contribution. In turn, the PO profile adopted in EX-FIB II was also reduced throughout the trial, but the impairment in performance may be attributed to a reduced glycolytic contribution (i.e. reduced lactate accumulation).