This study examined the test-retest reliability of peak physiological responses during wheelchair ergometry (WE) in individuals with spinal cord injury (SCI). Seven wheelchair dependent subjects, two with paraplegia (T10-11 and T11-12 lesions) and five with quadriplegia (all with C6-7 lesions), were given two incremental exercise tests to volitional fatigue on separate occasions within a one-week period. Each subject wheeled his or her personal wheelchair, which was mounted on a set of frictionless rollers with side-mounted flywheels. Metabolic and cardiorespiratory responses were continuously monitored by means of an automated metabolic measurement cart interfaced with an electrocardiograph. Statistical analysis revealed no significant differences (p greater than .05) between the mean values of the two test trials for six peak values. Reliability coefficients (p less than .01) were: oxygen uptake (0.98), heart rate (0.97), ventilation volume (0.96), respiratory exchange ratio (0.91), oxygen pulse (0.96), and ventilatory equivalent for oxygen (0.88). The investigators concluded that these six physiological responses in subjects with SCI undergoing WE are highly reliable, and that these variables can be used in the objective prescription, monitoring, and evaluation of exercise rehabilitation programs for individuals with SCI.