The aim of this study was to quantify and compare the reproducibility (intraexaminer method error) of cephalometric measurements in two groups of adult patients: one group with minor skeletal disharmonies and the other with severe skeletal disharmonies. It was hypothesized that errors were greater for subjects displaying severe morphologic deviations. The samples comprised two groups: 30 pretreatment cephalograms of patients planned for comprehensive orthodontic treatment and 30 cephalograms of patients additionally planned for orthognathic surgery. Two reference points and 28 cephalometric landmarks were defined and digitized by one examiner on two occasions 1 week apart. In total, 30 angles and 9 distances were studied. Reproducibility was expressed as the angular or linear differences between the two series of recordings. Two sample t tests were performed. In general, no significant differences in landmark location and measurement were found between the orthodontic and surgical patient groups. It was concluded that extreme variations in skeletal morphology do not effect the accuracy of cephalometric evaluation.