The purpose of this study was to investigate the effects of training and substrate manipulation upon the physiological and metabolic response to graded exercise for both trained and non-trained subjects. A major interest of this work was the concept of an "anaerobic" threshold for exercise of increasing intensity, and the relevance of the anaerobic threshold to endurance performance. Procedures were designed and validated to measure the physiological responses to maximal and submaximal work in an attempt to link physiological measures of performance to running performance for middle distance events. The potential of these measures to detect changes in running performance was investigated using cross-sectional studies of trained athletes and longitudinal studies of the effects of training. Biochemical analyses required the development of high performance liquid chromatographic procedures for the assay of hormones in blood and urine", and gas liquid chromatographic procedures for the assay of plasma fats. Performance times for middle distance running events were found to be highly correlated to selected physiological measures. The sensitivity of these measures were such that they were capable of detecting small differences in running performance between athletes of similar ability. In this respect?, the anaerobic threshold was found to be the most sensitive measure of performance. Changes in the anaerobic threshold were observed for both training and substrate manipulation. The corresponding changes in the metabolic responses to graded exercise suggest that the change in the anaerobic threshold is closely linked to overall metabolic adaptation to training. The data support the role of the anaerobic threshold as a sensitive indicator of the metabolic and cardiovascular adaptation to exercise and training.