The repeatability of radiographic measurements has been investigated frequently. Errors of Cobb angles of 5 degrees or more are very common when evaluating the same radiograph. When evaluating two different radiographs, further errors arise. The error margin in surface meas- urements of the trunk is normally estimated to be low, as evaluation of measurement is automated and as in repeat measurements of a plastic model, no significant differences have been found. When measuring living and moving objects, however, certain measurement errors are naturally to be expected. Purpose of this study is to evaluate the size of measurement error. 95 scoliosis pa- tients between 7 and 65 years of age, presenting a Cobb angle between 16 and 95 degrees were measured twice within 5 minutes in a standardized way with the Formetric System. The cor- relation coefficient for the maximal and the average lateral deviation was 0.961 respectively 0.987, the correlation coefficient for the maximal and average surface rotation was 0.967 respectively 0.935. The differences were not significant in the t-test. The relatively high technical error in the comparative analysis of the surface changes of the trunk due to positioning and movement of the living object has to be considered during the control follow-up. For the maximal respectively average lateral de- viation an error of 5.1 respectively 2.6 mm was evaluated, for the maximal respectively average ro- tation 2.3 respectively 0.9 degrees were evaluated.