Purpose: The aim of this in vitro trial was to evaluate and compare the external and internal adaptation of Class II composite and ceramic inlays with scanning electron microscopy after mechanical loading and thermal cycling (fatigue). Materials and Methods: Standardized cavities were prepared on human third molars with margins extending proximally 1 mm above or below the cement-enamel junction (CEJ). A fine hybrid composite (Tetric), a hydrothermal glass (Ducera LFC), and a sintered spinell ceramic (In-Ceram) were used to fabricate the restorations. Combinations of Syntac classic - Variolink, ED primer - Panavia and Scotchbond MP - Variolink were used for cementation. Results: For marginal adaptation in enamel, percentages of continuity varied from 85.4% to 99.6% and 85.1% to 98.2% for mesial and distal sites, respectively. Marginal adaptation in dentin showed continuity percentages varying from 63.6% (In-Ceram - Panavia) to 96.1% (Ducera LFC - Panavia). As regards internal adaptation, overall interface results varied from 32.1% (Tetric - Variolink) to 87.3% {Ducera LFC - Panavia) Gingival enamel showed the higher proportions of continuity (94.0%, all group-pooled data) compared to dentin locations (46.0% occlusally to 60.1% In gingival dentin). In all groups. debonding occurred mainly within the dentin - restoration interface and also, but to a lesser extent, between In-Ceram ceramic and both luting composites. When it occurred, adhesive failure to denim consistently took place at or close to the hybrid layer surface. Conclusion: Despite important differences in their physical properties, no difference could be found under the present experimental conditions between Class II composite and ceramic inlays m terms of their marginal and internal adaptation.