This cephalometric study was undertaken to compare and contrast the errors involved in taking linear and angular measurements using three different methods; hand instruments on tracing, digitization of tracings, and direct digitization of the radiographs. Of the three methods direct digitization of the radiographs proved to be the most reproducible particularly with angular measurements, although statistically significant differences were rarely found. Tracings were advantageous only with linear measurements involving constructed points such as Gonion and Articulare. Traditional methods of measurement using a ruler and protractor compared well with computer based methods involving the digitizer.