This in vitro study compared the marginal and internal adaptation of bulk-filled Class I and cuspal coverage direct resin composite restorations filled with different types of adhesive restorative systems and different thicknesses of bonding agent. Seventy-two intact, caries-free, freshly extracted human molars were randomly divided into 12 groups of six teeth each, according to the type of cavity (Class I [I] or Cuspal Coverage [C]), adhesive restorative system (SE Bond/Clearfil AP-X [SE] or Prime&Bond NT/Spectrum TPH [PB]) and thickness of bonding agent (normal or thick layer) in Class I restorations. Standardized Class I and Cuspal coverage cavities with enamel outer margins were prepared and restored with the corresponding type and thickness of bonding agent and respective resin composite. The resin composite was placed and polymerized in one increment (bulk filling). Dentinal fluid was simulated using 1:3 diluted horse serum and fed into the pulp chamber both during restoration and stressing. In six of the 12 groups, the restorations were subjected to 1.2 million mechanical occlusal cycles (maximum force 49 N; frequency 1.7Hz) and 3,000 simultaneous thermal cycles (5-50-5 degrees C). Marginal adaptation before and after mechanical and thermal stressing was assessed by using the replica technique and quantitative evaluation under SEM at 200x magnification. The teeth were dissected in a mesio-distal direction with a slow rotating diamond disc under water cooling, and the internal adaptation was also assessed by using the replica technique under the conditions described. Statistical evaluation of the continuous margin at the external and internal interface was performed with one-way analysis of variance (ANOVA) and Tukey's Studentized Range (HSD) test (p = 0.05). Even though Cuspal coverage restorations (SE- C: 96.89 +/- 1.83 and PB- C: 97.15 +/- 2.93) exhibited statistically significantly better external adaptation than Class I restorations (SE- I: 63.95 +/- 12.82 and PB- I 64.74 +/- 14.62) before stressing, there was no statistically significant difference after mechanical and thermal stressing (SE- C: 76.35 +/- 18.53 and PB- C: 76.02 +/- 12.49 SE- I: 54.67 +/- 10.82 and PB- I: 59.94 +/- 15.20). After stressing, SE Bond Cuspal coverage restorations (SE- C: 96,72 +/- 3,26) exhibited superior internal adaptation compared to SE Bond Class I restorations (SE- I: 57.83 +/- 12.91). No difference was observed in internal adaptation between Prime&Bond NT Cuspal coverage and Class I restorations (PB- C:36.46 +/- 21.82, PB- I: 38.71 +/- 6.76). In Class I restorations, the increased thickness in bonding did not improve the marginal and internal adaptation either before or after stressing. Bulk-filled direct resin composite Cuspal coverage restorations exhibited marginal adaptation similar to bulk-filled direct resin composite Class I restorations. The internal adaptation of Cuspal coverage SE Bond/Clearfil AP-X restorations was superior to all the other groups tested.