ABSTRACT Previous controlled reliability studies of cup to disc (C/D) ratio estimations may have been biased by lack of criteria for cup determination, small sample size; and/or large interval grouping of collected data. In our study four examiners independently estimated the C/D ratios of 40 patients by direct ophthalmoscopy using a contour criterion for cup determination. Results indicated that 75% of the estimate pairs differed by 0.10 or less. Fewer than 14% of the estimate pairs differed by 0.2 or greater. The results of the study suggest that when examiners use a specified criterion of contour cupping, interexaminer assessment of the C/D ratio estimation is clinically and statistically reliable.