This study examined the effect of three exercise-diet regimens on muscle glycogen supercompensation and subsequent performance during a 20.9-km run. A diet containing 15% carbohydrate (CHO,L), 50% CHO (M), or 70% (CHO (H) was arranged in three trials as follows: trial A = 3 days L, 3 days H; trial M = 3 days M, 3 days H; trial C = 6 days M. For each trial a 5-day depletion-taper exercise sequence was conducted on the treadmill at 73% V̇O2max. The runs were 90, 40, 40, 20, and 20 min, respectively. A day of rest preceded the 20.9-km performance run. Muscle biopsies were obtained from the gastrocnemius on days 4 and 7 (both prior to and after the performance run). Trials A, B, and C elevated muscle glycogen to 207, 203, and 159 mmol glucosyl units/kg wet tissue ImtnGI, respectively. The performance run in both trials A and B utilized significantly more glycogen than in trial C: 5.0 and 5.1 mmG/km vs. 3.1 mmG/km. There were, however, no differences in either performance run times or post-performance run glycogen levels between the trials. These data demonstrate that (1) muscle glycogen can be elevated to high levels with a moderate exercise-diet regimen; (2) initial muscle glycogen levels influence the amount subsequently utilized during exercise; (3) carbohydrate loading is of no benefit to performance for trained runners during a 20.9-km run.