This in vitro study compared three different methods to evaluate detectable levels of microleakage in Class II restorations placed with five commercially available packable resin composites: Alert, Glacier, Pyramid, Solitaire 2 and SureFil. A flowable resin composite, Flow-It, was used to line all packable composites. The hybrid resin composite Z100 was also included. The adhesive system used with all groups was Scotchbond MultiPurpose Plus. Standard Class II cavities were prepared on the mesial (enamel margins) and distal (dentin margins) sides of the teeth with no communication between them. Based on a power analysis, 180 permanent human molars were randomly assign to each of six groups with 30 specimens per group. All restorative materials were placed following manufacturers' recommendations. Following restoration and thermocycling, the specimens were stored at room temperature in solutions of 45Ca, methylene blue and rhodamine B, sequentially. Microleakage was ordinal scored as 1 (no penetration), 2 (penetration up to one-third of the cervical floor), 3 (penetration beyond one-third of the cervical floor to the axial wall) and 4 (penetration along the axial wall) by two independent evaluators. Analysis of the occlusal surfaces was also accomplished following the same scheme. In this study, tracers/dyes were evaluated for differences in penetration using generalized estimating equation methodology applied to cumulative logistic regression models. The results showed that Rhodamine B detected more microleakage than 45Ca or methylene blue, and 45Ca generally detected more microleakage than methylene blue.