The study aimed to validate the scoring of the degree of marginal degradation of amalgam restorations by using impressions, as an alternative to other indirect scoring methods using photographs or casts. Ten-year-old condensation silicone elastomer impressions and epoxy replicas made in 1979 were compared in a scanning electron microscope at 5 kV with different magnifications up to x200. The impre~sion material was not distorted or degraded, and the dimensional stability was good after 10 years of storage in a dry environment. The interexaminer agreement of the scorings of impressions and a six-point scale reference set was satisfactory as evaluated by kappa statistics, demonstrating that degrees of marginal degradation can be distinguished on impressions with relatively high accuracy. The rating distribution of the scorings of impressions showed good correlation to the rating distributions obtained with the clinical USPHS rating method and with photographs for recording marginal degradation. A slight difference between the photographic and impression ratings at the upper and lower levels of the six-point rating scale was observed. The difference varied with the type of alloy, possibly due to a bias depending on the surface quality-that is, whether the restoration kept the glossiness of high polishing or became heavily tarnished. 0 Amalgam degradation; clinical rating; clinical study; kappa statistics; silicone elastomer AsbjomJokstad,