Measurements were made of the physiological responses to exercise in patients who had had one leg immobilised following bone fracture. As a consequence of immobilisation they had muscular weakness and atrophy of the affected limb. Data were also collected for comparative purposes on normal healthy male subjects. Exercise was performed pedalling a stationary bicycle ergometer with each leg separately and both legs together and a system was developed to enable the pattern or force exerted on the crank by each leg to be measured. Anthropometric data including estimates of total leg and component tissue volumes were also obtained. Analysis of the pattern of force exerted in cycling established the comparability of 1-leg exercise involving patients' injured or uninjured legs. In 2-leg cycling the patients showed a disproportionate sharing of work between the legs, although the actual 'pattern* of force remained the same in both legs and the same as in 1-leg cycling. In submaximal exercise with the patients’ injured leg the oxygen uptake (iO2) for a given work load and cardiac frequency for a given iO2 were higher than with the uninjured leg. Rated perceived exertion was also higher for a given iO2 during exercise with the patients' Injured leg, but this difference was removed when iO2 was expressed in relative terms (% iO2 max). Maximum oxygen uptake (iO2 max) was reduced by -11% in the injured (cf. uninjured) leg, and was associated with the degree of muscle atrophy estimated anthropometrically. In 2-leg exercise there was a greater reduction of i02 max for a given leg muscle (plus bone) volume when comparison was made with normal subjects. The effect of rehabilitation therapy undertaken by the patients was to restore 1— and 2—leg exercises toward normal. The patient data were interpreted in relation to normal data including consideration of the effect of habitual limb preference.