Studies of pulmonary function were performed on seventy-four patients with thoracic scoliosis prior to corrective surgery by the Harrington method. The angle of the curvature, vital capacity, maximum breathing capacity, and arterial oxygen saturation were recorded at intervals up to five years after surgery.

Before operation in the patients with idiopathic scoliosis, there was a significant negative correlation when vital capacity, maximum breathing capacity, and arterial saturation were compared with the angle of the curvature. No such negative correlation was found in the patients with poliomyelitic and with congential scoliosis.

One year after operation in the over-all series, there was a decrease in the average vital capacity from 57 to 48 per cent of the predicted capacity and a decrease in the maximum breathing capacity from a preoperative average of 55 to 51 per cent of the predicted value. The patients with congenital scoliosis showed the greatest decrease postoperatively in both vital capacity and maximum breathing capacity; the patients with idiopathic scoliosis showed a small decrease in vital capacity and a slight increase in the maximum breathing capacity; and the patients with poliomyelitic scoliosis showed the same pattern as the combined group.

Two years and five years after operation, results were essentially unchanged from those obtained at one year. An over-all decrease in the average vital capacity occurred from the preoperative 56 per cent to 50 per cent at two years. The maximum breathing capacity remained essentially the same—it was 55 per cent of the predicted value before operation and 56 per cent two years after operation. Arterial oxygen saturation increased from 93.5 to 95.2 per cent after operation while there was a decrease in the spinal curvature from the preoperative average of 87 degrees to an average of 53 degrees.

There was no correlation between the change in the angle of the curvature and the changes in the vital capacity of maximum breathing capacity except in the patients with poliomyclitic scoliosis. In these patients, a significant correlation was found between the amount of correction of the curvature and the decrease in vital capacity.

The Harrington method for the correction of kyphoscoliosis does not improve the average vital capacity or maximum breathing capacity but may improve arterial oxygen saturation. Since there is a positive correlation between the decrease in the spinal curvature and the decrease in vital capacity in the poliomyelitis group, over-enthusiastic correction of the curve may lead to a further decrease in lung volumes, but adequate treatment should be instituted as soon as a curve is noted in order to preserve that lung function is left. Since before operation there was a correlation between the amount of curvature and the decrease invital capacity in the patients with idiopathic scoliosis, treatment of idiophathcic scoliosis should be instituted if there is any sudden increase in the curvature or as early as possible if the curve is slowly progressing. Treatment then will preserve what respiratory function is present at the time of surgery.