Seven experts drew the rectal axes of 18 representative proctographic images on two occasions, with a 1-year interval, in order to assess intraobserver variation in the determination of the anorectal angle (ARA). Intraobserver variation (6%) and interobserver variation (17%) were smallest when the central rectal axis was used to determine the ARA. A strong relation was found between inter- and intraobserver variation (r = 0.77). Intraobserver variation tended to be rather small for pictures made during straining, but a relation with the magnitude of the ARA was not found. Although none of the seven experts could reproduce the rectal axes with less than or equal to 10% variation in all 18 pictures, redrawing of the central rectal axis delivered less than or equal to 10% variation in 86% of determinations. It is concluded that intraobserver variation is influenced by the expertise of the investigator, the method of analysis, and the anorectal configuration to be analyzed. Radiologic assessment of the ARA may yield reliable data on the dynamics of the anorectum if performed by a single investigator on x-ray films that allow confident analysis.