Spinal curvatures alter measured stature and may influence the evaluation of skeletal maturity and growth based on stature measurements. A dataset of calibrated measurements of vertebral positions of 407 radiographs in the frontal plane, together with clinically measured Cobb angles was used to determine the difference between spinal length and spinal height ('height loss') as a function of Cobb angles for radiographs indicating both single (N=182) and double (N=225) curves. An apparently quadratic relationship: Height loss (mm)=1.0+0.066*Cobb+0.0084*Cobb*Cobb was found between height loss and each patient's mean Cobb angle for double curves. There was close agreement of the regression coefficients for single and double curves, and the present findings were very similar to the relationship reported by Ylikoski (Eur Spine J, 2003, 12:288-291). The relationships differed substantially from those proposed by Bjure (Clin Orthop, 1973 93:44-52) and by Brookenthal (SRS Exhibit 15, 2002). The findings of the present study indicate that height loss (in mm) occurring with a 10 degrees increase in mean Cobb angle (for two curves) would be 1.1+0.16 times the mean Cobb angle (in degrees). For example, for a Cobb angle change from 30 to 40 degrees, the expected height loss would be 1.1+35*0.16 mm=6.7 mm. This assumes that height loss occurs only as a result of altered curvature, without alteration in disc height associated with an increase in scoliosis.