Vertebral height ratios are used to define vertebral deformity in clinical and epidemiologic studies of vertebral osteoporosis. However, few data have been obtained on the variation in these ratios in different populations using standard methods. We examined vertebral morphometric measurements obtained in a population survey from three centers: Malmö (Sweden), Montceau-les-Mines (France), and Graz (Austria), to study the influence of sex and the population center on vertebral height ratios. Radiographs were obtained according to a standardized protocol, and morphometric measurements, anterior height Ha, central height Hc, and posterior height Hp, made in Berlin. The height ratios anterior, Ha/Hp, central, Hc/Hp, posterior I, Hp/Hp', and posterior II, Hp/Hp" (Hp' = posterior height of vertebrae above, Hp" = posterior height of vertebrae below) were calculated for each vertebra from T4 to L4. The mean and standard deviation of these ratios for each sex and each center were derived using a statistical trimming procedure to normalize the distribution. Threshold values for defining grade 1 and grade 2 deformities, wedge, biconcavity, and compression, were calculated using these parameters. Anterior and central vertebral height ratios were smaller in males than females (p < 0.01). There were significant differences between the three centers (p < 0.01) both in the trimmed mean values for anterior and central vertebral height ratios and in the thresholds derived using standard criteria for defining wedge and biconcavity deformity. The data confirm the impression from single-center studies that vertebral height ratios vary between populations and suggest that reference values for vertebral height ratios should be derived separately for males and females within individual populations whenever possible.