An investigation was undertaken to determine the inter- and intraexaminer reliability of the Gonstead pelvic radiographic marking system. This methodology analyzes the relative dimensions and spatial positions of the two innominate bones and sacrum and measures leg length inequality through an evaluation of femur head height. Two examiners marked 71 full spine radiographs twice to provide data from 284 analyses. Reliability was ascertained with the Pearson r, Spearman, intraclass correlation coefficient (ANOVA) and Kappa statistics. All results were statistically significant (<0.001) and indicated high levels of concordance. In every case, intraexaminer agreement was superior to interexaminer concordance.