When a muscle contracts it converts chemical energy into heat and mechanical work. The total amount of this energy 'liberation' can be found, for a cycle of contraction and relaxation, by adding any net external work done to the heat produced. Both these quantities can be directly determined. To find the time course of the energy liberation during a contraction is much harder because not only is the recorded heat production, particularly during relaxation, greatly affected by any inequalities in the muscle (Hill & Howarth, 1957;Hill, 1961), but also two other factors enter into the calculation: (1) the 'internal' work, that is work done by the contractile part of the muscle in stretching the series elastic component and any compliance in the apparatus to which it is connected; to find out how much energy is being liberated by a muscle when the tension is rising (or falling) it is necessary to add (or subtract) the internal work which is being done (or absorbed); (2) the thermoelastic heat; it was shown by Hill (1953 a) that, when the tension falls in a contracting muscle, heat is produced proportional to the fall of tension. Woledge (1961) showed this to be a reversible process, heat being absorbed when the tension rises. Although the molecular nature of the thermoelastic effect is not known it is believed to be the result of some physical process occurring in a muscle when the tension changes; that is, it is not a direct result of the chemical reactions supplying the energy for the contraction. Therefore to find the heat or energy which is being produced by these reactions when the tension is changing it is necessary to correct the observed heat production for the thermoelastic effect: adding to it the heat absorbed by the thermoelastic effect when the tension is rising, and subtracting the heat produced by a fall of tension. The heat production so corrected will be referred to, for convenience, as the 'true' heat production.Neither the internal work nor the thermoelastic heat change can be directly recorded. Both are zero over a complete cycle of contraction and relaxation. The rate of change of each is zero when the tension is constant, therefore the rate of energy liberation can be found without these complications when the tension is constant; for example, (a) during the 'plateau'