The results of treatment of idiopathic scoliosis with a Milwaukee brace were studied in 133 patients (127 girls and six boys) whose ages ranged from eight years and five months to sixteen years and two months at the beginning of treatment. These 133 patients had 192 separate curves (119 right thoracic, thirty-nine high lumbar, twelve thoracolumbar, and twenty-two high left thoracic). Of these patients, seventy-four with 109 curves were followed for five years or more after the brace was discontinued (average, eight years; range, five to thirteen years); twenty-nine patients were treated surgically because of a poor response to the brace or progression of the curve; and thirty patients were lost to follow-up. More than 80 per cent of the seventy-four patients followed for five years or more showed some increase of their curves after the brace was discontinued. The average correction at follow-up compared with the original curve was 2 degrees for thoracic curves (range, -18 to 24 degrees) and 4 degrees for the thoracolumbar and lumbar curves (range, -11 to 17 degrees). The brace was more effective for curves of less than 40 degrees. More than one-third of the patients with curves of 40 degrees or more eventually required surgical treatment. Age, curve pattern, and status of the iliac and ring epiphyses did not correlate withe response to brace treatment. The best guideline for prediction of the results of brace treatment was the response of the curve to the brace, especially during the first year of treatment. If the curve is reduced in the brace to less than 50 per cent of its initial measurement, there is a good chance of obtaining significant permanent correction.