Cephalometric analysis conventionally requires radiographic exposure which may not be compatible with the growing concern over radiation hazards. Recently, the Dolphin Workstation Imaging System introduced to the dental profession a non-radiographic system, called the DigiGraph Workstation which may be an alternative to cephalometric radiography. The aims of this study were to compare the validity and reproducibility of cephalometric measurements obtained from the DigiGraph Workstation with conventional cephalometric radiographs. The sample consisted of 30 human dry skulls. Two replicated sets of lateral cephalograms were obtained with steel ball markers placed at the majority of the cephalometric landmarks. Duplicate tracings prepared from each radiograph were digitized to obtain cephalometric measurements using the computer software, Dentofacial Planner. For the DigiGraph Workstation, double sonic digitizations were repeated twice for each skull, on two occasions. Fifteen angular and one linear measurements were obtained from both methods and these findings compared using ANOVA, paired t-tests and F-tests. All, except one, cephalometric measurement showed significant differences between the two methods (P < 0.0001). The DigiGraph Workstation consistently produced higher values in 11 measurements (mean differences +0.5 to +15.7 degrees or mm) and lower values in four measurements (mean differences -0.2 to -3.5 degrees). The standard deviations of the differences between readings of both methods were large (0.4-5.8 degrees or mm). The reproducibility of the DigiGraph Workstation measurements was lower than that of the radiographic measurements. The method error of the DigiGraph Workstation ranged from 7 to 70 per cent, while that of radiographic tracings was less than 2 per cent. It was concluded that measurements obtained with the DigiGraph Workstation should be interpreted with caution.