During the last decade two types of energy balance discrepancies were reported: 1) enthalpy production (heat + work) during contraction is greater than that expected on the basis of the known molar enthalpies and phosphagen breakdown; 2) recovery oxygen consumption is greater than that predicted when using standard biochemical stoichiometry and the phosphagen breakdown during contraction (delta approximately P). To test whether these phenomena were causally related, measurements of delta approximately P, oxygen consumption (Jo2), and enthalpy production during contraction (QI) and recovery (QR) were made on frog sartorius muscle at O C. To achieve equal precision among these diverse measurements, a steady-state protocol involving 3-s isometric tetani at 256-s intervals was employed. delta approximately P during the first 3-s tetanus was not different from that during a contraction in the steady state, and averaged 1.1 mumol/g. Steady-state Jo2 was 0.11 mumol-min-1.g-1, approximately 70% of the maximum rate. QI and QR were 88.2 and 93.3 mJ/g (QR/QI = 1.06). Neither the enthalpy during contraction nor recovery oxygen consumption could be accounted for in terms of delta approximately P. However, a total energy balance was achieved, i.e., the total enthalpy production could be accounted for solely in terms of the measured Jo2 and the molar enthalpy of carbohydrate oxidation. As the oxidation of glycogen is implicated as the only net reaction, the unknown reactions producing the unexplained enthalpy during contraction must be reversed during the recovery period. This reversal would require hydrolysis during recovery of mumol ATP per 35.2 mJ of unexplained enthalpy if the theoretical ADP/O ratio of 3.25 is to be attained. Thus a major portion of the phosphagen breakdown associated with muscle contraction is likely to occur during recovery.