This study examined the effect of thermal cycling on the microleakage of bonded amalgam restorations. Three dental amalgam alloys and a gallium alloy were tested with two adhesive resin systems and copal varnish as a control. Class V cavity preparations were prepared on 168 freshly extracted premolars or molars. The preparations were placed parallel to and 1.0 mm occlusal to the cementoenamel junction (CEJ). Four groups of 42 teeth each were treated with one of the following adhesive dentin systems: Bond-It, All-Bond 2/Resinomer or a copal varnish (Copalite). The four groups of 42 teeth each were then restored with one of three dental amalgams: Orosphere Plus, Indiloy, Oralloy or a Gallium alloy (Galloy), resulting in 12 test groups of 14 teeth each. The specimens were stored in double distilled water at 37 degrees C for 24 hours. Final contouring and polishing of the restorations were performed under water spray. Half of the restorations in each group were thermocycled for 3000 cycles (5 degrees C-37 degrees C-55 degrees C-37 degrees C) with a dwell time of 15 sec at each temperature. The other half were stored in double distilled water at 37 degrees C for 24 hours. Then all 168 restorations were stained with dye, sectioned and scored for microleakage. Results showed that the adhesive dentin systems reduced microleakage in amalgam restorations compared to copal varnish only in non-thermocycled specimens. Statistical analysis of the results showed that there was an extremely significant difference (p < 0.001) in microleakage between the non-thermocycled and the thermocycled specimens in all test groups, whereas, there was no significant difference (p > 0.05) among thermocycled specimens. The reduction of microleakage was not significantly different between Bond-It and All-Bond 2/Resinomer in non-thermocycled specimens. Oralloy showed the most microleakage in the non-thermocycled groups when compared to the other alloys using the same adhesive liner.