A comprehensive clinical investigation of amalgam restorations was undertaken to evaluate the performance of "high copper' amalgams and the methods used to examine the marginal integrity of restorations. Six alloys, Aristaloy CR (A), Dispersalloy (D), Indiloy (I), an experimental "high copper' amalgam (E), Tytin (T), and New True Dentalloy (N) as a control, were randomly placed by four operators in 435 Class 1 and Class II restorations. Black and white 1 x photographs were taken at baseline, 6, 12 and 18 months and evaluated using the method described by Mahler & Marantz (1979). In addition, the U.S. Public Health Service system (Cvar & Ryge, 1971) was used at each recall. An analysis of variance based on photographic scoring indicated significant differences (P less than 0.05) among alloys at each recall. The use of Duncan's multiple range test resulted in two homogeneous subsets at 18 months: D, I, E and T, N, A. Six and 12 months recall data showed a close similarity in the rank order of the six alloys. The U.S.P.H.S. system showed no statistically significant differences between the alloys using the chi-square test, however, at each recall the rank order was comparable to that obtained with the photographic method.