Twenty eight subjects (mean age 64 years) who had been treated for tuberculosis by thoracoplasty in the past performed an increasing work rate exercise test, from which maximum oxygen consumption (VO2max), ventilation and heart rate were measured. VO2max was significantly lower than predicted, being 0.75 l/min in 17 subjects, 1.0 l/min in 10, and 1.5 l/min in one. Only one subject achieved a heart rate of 85% of the predicted maximum. The ratio of heart rate to oxygen consumption (HR/VO2) and heart rate at standard interpolated submaximal levels of oxygen uptake at 0.75 l/min (heart rate 0.75) and 1.0 l/min (heart rate 1.0) were normal. VO2max correlated with ventilation at maximal exercise (VE max) (r = 0.87) and FEV1 (r = 0.47). It did not correlate with resting arterial oxygen or carbon dioxide tensions, FEV1, maximum inspiratory pressure, angle of scoliosis, or number of ribs resected. The relation between ventilation and oxygen consumption (VE/VO2) and VE at the submaximal levels of oxygen consumption of 0.75 l/min (VE 0.75) and 1.0 l/min (VE 1.0) were normal. In 10 subjects a plateau of breathing frequency (fmax) was reached, after which the increase in ventilation was achieved by a further increase in tidal volume (VT). These subjects showed significantly lower values for the forced expiratory ratio, VO2max, and VEmax than those with a normal relation between tidal volume and breathing frequency. VEmax was correlated with FEV1 (r = 0.61), FVC (r = 0.46), maximum VT (r = 0.55), change in VT (r = 0.52), fmax (r = 0.56), and change in breathing frequency (r = 0.72). These results indicate that exercise in patients treated for tuberculosis by thoracoplasty is limited by ventilatory capacity and that this is due to a reduction in both dynamic lung volumes and respiratory frequency.