Aim of this research was to test the hypothesis that marginal gap measurements by a digital microscopic image processing computer are reproducible. Light microscopic images of the margins of Artglass partial crowns on stone dies were obtained with a 3CCD color video camera, digitized and processed. Ten measuring points were painted on a 0.5 mm section of the finish line of the preparation. The opposite corresponding measuring points on the margin of the partial crown were painted as well. The distance (marginal gap) between two corresponding points was painted and calculated by the computer. Reproducibility (precision) is reflected in the ability to reproduce the same measurement results in repeated measurements (n = 5). The reproducibility error was expressed as the coefficient of variation (CV) in percent (%). The localization of the corresponding measurement points and the calculation of the marginal gap were performed without errors. For a marginal gap between 15-50 microns the average measurement error was 3% or 0.45-1.5 microns. It was concluded that the measurement of the gap between two points by the digital image processing computer was reproducible. The measurement error was neglectable given that a clinically acceptable marginal gap is 100 microns.