This study verified the relationship between the volume and microleakage of conventional and bonded amalgam restorations. Also, the microleakage influence of intermediate materials, substrates and the direction of sectioning was investigated. Fifty-six bovine incisors were selected. Standard Class V cavities were prepared in buccal and lingual surfaces. For each tooth, two cavity sizes were prepared, corresponding to two cavity volumes: one larger (A) and the other smaller (B). The cervical wall was located in cementum/dentin and the incisal wall in enamel. The teeth were distributed in four groups (n=28) according to the intermediate material employed (glass-ionomer cement, resin cement, adhesive system and copal varnish-control). The materials were applied following manufacturers' directions. After restoration, the teeth were submitted to thermal cycling. They were then immersed in a dye solution and sectioned in two directions inciso-cervical (IC) and mesio-distal (MD) sections to evaluate the microleakage. Data were subjected to non-parametric statistical analysis (Wilcoxon's paired test and Kruskal-Wallis test). No significant difference was found between the two cavity sizes. Leakage in enamel was statistically lower than in the cementum/dentin interface (p < 0.05). In some situations, glass-ionomer or resin cement lined amalgam restorations presented less dye leakage than copal varnish lined restorations (p < 0.05). No significant difference was observed in microleakage between IC or MD sectioning. Within the limitations of this study, it was concluded that cavity size and direction of section were not significant factors for microleakage, while substrate and intermediate materials had a significant effect on the sealing ability in amalgam restorations.31377-83