The aim of this in vitro study was to evaluate marginal leakage in class V restorations in primary teeth restored with amalgam, using three different techniques. Thirty maxillary anterior primary teeth, clinically sound and naturally exfoliated, were used. In group 1 (n = 10), two thin layers of a copal varnish (Cavitine) were applied. In group 2 (n = 10), Scotchbond Multi-Purpose Plus, a dual adhesive system, was used according to manufacturer instructions. In group 3 (n = 10), One-Step adhesive system in combination with a low-viscosity resin (Resinomer) were used according to manufacturer instructions. All samples were restored with a high-copper dental amalgam alloy (GS 80, SDI). After restoration, the samples were stored in normal saline at 37 degrees C for 72 h. The specimens were polished, thermocycled (500 cycles, 5 degrees and 55 degrees C, 30-s dwell time) and impermeabilized with fingernail polish to within 1.0 mm of the restoration margins. The teeth were then placed in 0.5% methylene blue for 4 h. Finally, the samples were sectioned and evaluated for marginal leakage. The Kruskal-Wallis test showed that the filled adhesive resin (group 3) had the least microleakage. There was no significant difference between groups 1 and 2.