A two-year prospective study was done to assess the prevalence and distribution of various parameters associated with scoliosis in schoolchildren in northwestern and central Greece. A total of 82,901 children (41,939 boys and 40,962 girls) who were nine to fourteen years old were screened for scoliosis. Five thousand eight hundred and three children had clinical signs of scoliosis and, of these, 4185 were referred for posteroanterior radiographs (to be made with the patient standing) because they had a positive result on the forward-bending test (a difference of more than five millimeters between the two sides of the torso as measured in the thoracic or thoracolumbar region with use of a ruler and a level plane) at the time of a second screening.The prevalence of scoliosis (defined as a curve of 10 degrees or more) was 1.7 per cent (1436 of 82,901 children), and most of the curves (1255; prevalence, 1.5 per cent) were small (10 to 19 degrees). The ratio of boys to girls was 1:2.1 over-all but varied according to the magnitude of the curve (1:1.5 for curves of less than 10 degrees, 1:2.7 for curves of 10 to 19 degrees, 1:7.5 for curves of 20 to 29 degrees, 1:5.5 for curves of 30 to 39 degrees, and 1:1.2 for curves of 40 degrees or more). Thoracolumbar curves were the most common type of curve identified, followed by lumbar curves; specifically, of the 1436 children who had a curve of at least 10 degrees, 493 (34.3 per cent) had a thoracolumbar curve, 475 (33.1 per cent) had a lumbar curve, 261 (18.2 per cent) had a thoracic curve, and 207 (14.4 per cent) had a double curve. Although most (753) of these curves were to the left, the left:right ratio varied according to the location of the apex of the curve (1:3.1 for thoracic curves, 2.0:1 for thoracolumbar curves, and 3.2:1 for lumbar curves).The cost of the screening process was negligible (estimated at thirty cents per child); however, the decreased number of operative procedures performed in children from the geographical area of our University Hospital, the identification of a large number of previously undiagnosed curves (eleven of which were treated operatively and 170 of which were treated with a brace), and the identification of children who were at high risk for progression were considered important benefits of the school-screening program.