Lateral flexion–extension radiographs of 72 patients with Down syndrome were used to assess the interobserver reliability and intraobserver reproducibility of the atlanto–dens interval, Wiesel–Rothman measurement, occiput atlas angle, and Power's ratio in flexion and extension. The radiographs were reviewed by three blinded observers on three different occasions with at least a 1-month interval between assessments. The intraclass correlation coefficient was used to measure the reproducibility of the measurements from a given observer and the reliability between different observers. With the exception of observer one, the atlanto–dens interval had a statistically significant intraobserver agreement compared with any of the other measurements (p < 0.05). The atlanto–dens interval and the Wiesel–Rothman measurements tended to have better correlation between observers, although there was only fair agreement. The agreement, however, was statistically significant (p < 0.05) compared with Power's ratio. The degree of intraobserver reproducibility and interobserver reliability may make it difficult to base treatment protocols on these measurements.