1. This study tested the hypothesis that moderate fatigue of skeletal muscle arises from a mismatch between energy demand and energy supply. Fatigue was defined as the decline in isometric force. Energy supply and demand were assessed from measurements of muscle heat production. 2. Experiments were performed in vitro (21 degrees C) with bundles of muscle fibres from mouse fast-twitch extensor digitorum longus muscle and slow-twitch soleus muscle. Fibre bundles were fatigued using a series of thirty isometric tetani. Cycle duration (time between successive tetani) was 5 s. The amount of fatigue that occurred during a series of tetani was varied by varying contraction duty cycle (tetanus duration/cycle duration) by varying tetanus duration. 3. Peak isometric force and total heat production in each cycle were measured. For each cycle, the amounts of initial heat (H(i)) and recovery heat (Hr) produced were calculated and used as indices of energy use and supply, respectively. H(i) and Hr were used to estimate the net initial chemical breakdown (in energy units) in each cycle (H(i,net)). 4. The magnitude of H(i,net) was greatest in the early stages of the contraction protocol when Hr was still increasing towards a steady value. The magnitude of decline in force between successive tetani was proportional to H(i,net) for both muscles. 5. The results are consistent with the idea that the development of moderate levels of fatigue at the start of a series of contractions is due to the rate of energy supply being inadequate to match the rate of energy use.