Skeletal muscle cyclic AMP (cAMP) content and glycogen synthesis were investigated in male rats subjected to exhaustive exercise, alloxan diabetes, and combinations of these conditions. After an exhaustive swim or control treatment of wading, randomly selected animals were administered 500 mg glucose via stomach tube. Two hours after glucose administration, gastrocnemius glycogen levels rose from 1.31 to 10.67 mg/g wet wt in fatigued nondiabetics (FND), producing a 94% supercompensation above control values. Glycogen of fatigued diabetics (FD) increased from 0.88 to 4.21 mg/g wet wt during the first 2 hr after glucose administration and did not reach control values for 24 h. In conjunction with these glycogen changes, cAMP increased from 1.23 to 2.59 and 1.47 to 2.81 pmol/mg wet wt for FND and FD, respectively (P less than 0.05). No difference in cAMP levels between diabetics and nondiabetics was found. These in vivo data suggest that insulin may not be essential for muscle glycogen synthesis, but that after glycogen depletion it plays a prominent role in supercompensation. Also, this hormone's mechanism of action in skeletal muscle does not appear to be mediated through alteration in the tissue cAMP concentration.