Rats were fed a diet containing 1% beta-guanidinopropionic acid (GPA) for 6-10 weeks to deplete their skeletal muscle of creatine. 31P-NMR was used to monitor metabolic changes in the gastrocnemius muscle at rest, during stimulated steady-state isometric contraction at 4 Hz and during recovery from stimulation. In resting muscles, the [creatine phosphate] was reduced to 10% (2.8 mumol X g-1) and the [ATP] to 50% (3.3 mumol X g-1) of those found in rats fed a control diet. The concentration of the phosphorylated form of the analogue (PGPA) was 23 mumol X g-1. There was no significant difference in muscle performance or in the relative changes in the [ATP] during stimulation. Intracellular pH decreased rapidly on stimulation and recovered during the stimulation period to near resting values in both groups. In control rats, the initial decrease in pH was greater and the time to recovery was longer than in GPA-fed rats. The rate at which PGPA supplied energy to the contracting muscle (0.027 mM X s-1) was insignificant relative to the minimum estimated rate of ATP turnover (1 mM X s-1). The rate of PGPA resynthesis during recovery (0.018 mM X s-1) is enzyme-limited and provides an independent estimate of creatine kinase flux during this period (18.9 mM X s-1). The creatine kinase flux (creatine phosphate----ATP) in the resting muscle of GPA-fed rats was 12-fold less than in control animals, 1.3 vs. 15.7 mM X s-1. These results demonstrate that neither the [creatine phosphate] nor the activity of creatine kinase is critical for aerobic metabolism. Skeletal muscle appears to adapt to a diminished creatine pool by enhancing its aerobic capacity.