Marginal integrity is important in the evaluation of a restoration. The fit of a crown is important in long-term success of a cast restoration. However, there is no agreeable definition of a clinically acceptable margin. Purpose: The purpose of this project is to develop a test apparatus that allows students and prosthodontists to evaluate and/or develop skills in non-visible and visible margin detection. Methods: The margin detection apparatus used three extracted caries-free teeth. Teeth were prepared for full cast (gold) crown using a chamfer finish line configuration. Die spacer was applied to the marginal area before waxing to vary the marginal opening. After casting, the axial marginal openings were measured. Pre-doctoral students (n=10) and prosthodontists (n=9) were asked to evaluate for “clinical acceptability” or “unacceptability” of each marginal area. Each subject repeated the evaluation 6 months later. Results: Marginal gaps ranged from 40-615ìm. Cohen’s kappa based upon all nineteen raters = 0.39, the nine prosthodontists = 0.51 and the ten predoctoral students = 0.30. Spearman rank correlation of margins reported “clinical unacceptable” = 0.81, p=0.0001. Wilcoxon signed rank test for proportions rated “clinically unacceptable” by the two groups (p>0.9), for supragingival surfaces (p>0.35) and subgingival surfaces (p=0.16). Upon re-evaluation the number of surfaces rated differently by the prosthodontists ranged from 1 to 6, with a median of 1; and the predoctoral students ranged from 1 to 6, with a median of 3. Wilcoxon Rank Sum test or the number of inconsistent ratings was 0.05