Musculoskeletal injuries are commonly associated with muscle atrophy as a function of immobilization or change of normal function. For example, injuries to the anterior cruciate ligament (ACL) which may involve ligament reconstruction, results in the "quadriceps avoidance" gait which leads to atrophy of the knee extensormuscles. In these situations it is not clear whether or not the atrophy is associated with loss of specific muscle fibre types with accompanying functional deficits. Such knowledge would be helpful in implementing exercise regimes designed to compensate for loss of particular fibre types. It is believed that isokinetic exercise performed at speeds below 180° per second strengthens type I muscle fibres, and type II fibres at fast speeds. However, there is no evidence to indicate the specific muscle fibre response to different rates of muscle contraction. Identification of muscle fibre type is most directly determined by biopsy technique but is too invasive for a routine measurement. Electromyography median frequency has been used as a non-invasive measure to infer muscle fibre composition in various studies. However, the reliability and accuracy of this technique has been questioned and improvement is necessary. This research was designed to provide a more accurate and reliable protocol for the determination of EMG median frequency which may be used, after validation against more direct biopsy techniques, as a routine method for inferring muscle fibre composition. The investigation also explored the muscular response as measured by EMG median frequency to varying speeds of muscle contraction, fatiguing exercise and atrophy following ACL reconstruction. The ultimate aim of this research was to improve the reliability of the determination of EMG median frequency to enhance its application as a predictor of muscle fibre composition. This provides information which may improve ACL rehabilitation programs designed to restore and prevent specific muscle fibre types loss that have not previously been targeted by current rehabilitation programs. This research was conducted in three studies. Study one determined the stability of the EMG median frequency bilaterally for the quadriceps and hamstrings muscles and identified the mode of contraction associated with the greatest reliability. The strength and EMG median frequency of the vastus lateralis, medial hamstrings and vastus medialis of 55 subjects was determined across 5 speeds from 0° to 240° per second using a Kin-Com isokinetic dynamometer and an EMG data acquisition system. Isometric contraction was found to have the least bilateral discrepancy (4.01% ±3.06) and between trials standard deviation (4.50) in the vastus lateralis, medial hamstrings and vastus medialis. Study two investigated the EMG median frequency changes in the vastus lateralis which occur immediately following different speeds of isokinetic exercise to the point of fatigue in normal subjects. Thirty-four subjects participated in the study, and performed a 90-second period of isokinetic exercise to activate the knee extensors at either 30° or 300° per second. EMG median frequency of the vastus lateralis was determined before, immediately after and 7 minutes after the fatiguing exercise. The percentage drop in EMG median frequency of the vastus medialis was gnificantly (p<0.05) greater after slow speed (27.9%) than fast speed (20.25%) exercise, while no significant difference was found for the percentage drop in extension torque. Full recovery was found 7 minutes after the fatiguing exercise. By reference to previous research showing a relationship between EMG median frequency and muscle fibre type, an increase in activation of type I muscle fibres with slow speed exercise and an increase in type II muscle fibres with fast speed exercise was observed. Study three identified the changes in EMG median frequency following ACL reconstruction and evaluated the bilateral differences in EMG median frequency of the knee muscles. The relationships between EMG median frequency and the measures of knee functional ability, knee muscle strength, age and time since surgery were also investigated. Twelve subjects who had undergone ACL reconstruction using a semitendinosus and gracilis graft 6 to 12 months earlier, participated in the study. EMG median frequency was determined from an 8-second isometric contraction and knee functional ability was assessed using the Cincinnati Rating Scale. Bilateral EMG median frequency shifts were inconsistent among subjects. On the basis of previous research which indicated a relationship between EMG median frequency and fibre type, no consistent pattern of muscular fibre type atrophy subsequent to ACL reconstruction occurred within 6 to 12 months (ranged from -43 to 57 Hz). Additionally, no significant correlations were found between the EMG median frequency and the knee functional score and knee extension and flexion torques, age, time since operation and the bilateral differences in EMG median frequency. The results of this investigation will serve to improve the reliability of EMG median frequency across a range of conditions in which it has been evaluated. Further research is needed to confirm the relationship between EMG median frequency and direct observations of muscle fibre composition to improve the predictive value of this measure. Following this validation it will be possible to evaluate the bilateral EMG median frequency shift to infer the type of muscle fibre atrophy, and use this measure in determining the efficacy of specific rehabilitation programs. In conclusion * An 8-second isometric contraction is recommended for determining EMG median frequency. * EMG median frequency of a muscle decreases significantly more after slow fatiguing exercise than after fast speed fatiguing exercise. * There was no generalised bilateral EMG median frequency shift found in a group of subjects 6 to 12 months following semitendinosus and gracilis graft ACL reconstruction. * The results of this study will serve to improve the reliability of procedures used to determine EMG median frequency under a range of different contractile conditions. The EMG median frequency changes in response to these conditions require further validations with muscle biopsy in future.