The first commercial high-copper amalgam was developed in the mid-1960's. In 1975, a study to evaluate in vivo/in vitro performance of nine of these new alloys was initiated at Indiana University. These nine high-copper amalgams plus a traditional amalgam were tested for four laboratory properties and their gamma-2 content, and were clinically assessed for fracture at the margins. Originally, 60 restorations of each material were placed in patients and then recalled at six, 12, and 24 months. Recently, 13 years post-operatively, approximately one-third of the original patients were located. These patients were given a clinical exam, and photographs were taken of their restorations. From these recall data, the mean loss of all restorations at 13 years was 11.3%. The alloys Cupralloy, Tytin, and Dispersalloy had the lowest loss rates. The evaluation of the fracture at the margins (ridit means) indicates that there are five groups of alloys with some overlap between groups. The alloys Cupralloy, Indiloy, Dispersalloy, and Phasealloy exhibited the least fracture at the margins. The gamma-2-containing alloys--Optaloy II, Micro II, Velvalloy, and Aristaloy CR--showed the greatest rate of fracture at the margins. A correlation matrix of the ridit means at different years, the loss of restorations at 13 years, and the mechanical properties indicate that the early (one- and two-year) fracture at the margins predicts the 13-year results. Mechanical properties, however, do not predict the clinical performance of the materials at 13 years.