The efficiency of energy transduction is defined as the ratio of the work done by a muscle to the free energy change of the chemical processes driving contraction. Two examples of the experimental measurement of muscle efficiency are: (1) the classical method of Hill which measures the value during a steady state of shortening, (2) measuring the overall efficiency during a complete cycle of a sinusoidal process, which comes closer to the situation during natural locomotion. The reasons why fatigue might lower efficiency are the following. (1) The reduction in PCr concentration and increase in Pi and Cr concentration which are characteristic of fatigued muscle, reduce the free energy of PCr splitting. This will reduce the efficiency of the recovery process. It is not known whether the efficiency of the initial process is increased to compensate. (2) There is a general conflict between efficiency and power output when motor units are chosen for a task or when the timing of activation is decided. During fatigue more powerful units have to be used to achieve a task which is no longer within the scope of less powerful units. (3) The slowing of relaxation that is sometimes found with fatigue may make it impossible to achieve the short periods of activity required for optimum efficiency during rapid cyclical movements. A reason why fatigue might increase efficiency is that muscles are thought to be more efficient energy converters when not fully activated than when fully active. Full activation is often not achieved in muscle which is considerably fatigued. Available observations do not allow us to find where the balance between these factors lies. The conclusion is thus that experiments of both the types discussed here should be performed.