OBJECTIVE To investigate the marginal adaptation of resin-modified glass ionomer cements in dentin cavities placed with or without additional application of resin bonding systems. METHODS Three resin-modified materials (Fuji II LC, Photac-Fil, Vitremer), one compomer (Dyract) and as reference an adhesively bonded resin composite system (Gluma CPS-Pekafill) were used. Flat peripheral dentin surfaces on human molar teeth were produced by wet grinding on SiC paper. Cylindrical cavities, 3.5 mm wide, were prepared in these dentin areas and restored with the individual materials. Sixty cavities were pretreated and restored as requested by the respective manufacturers. Following water storage of the specimens for 15 minutes or 24 hours, excess was gently removed by wet grinding for microscopic inspection of the marginal area. Additionally, in 30 cavities an experimental one-component adhesive resin system, a proprietary dimethacrylate and HEMA mixture dissolved in acetone, was combined with each of the restoratives for evaluation after 15-minutes water storage. Finally, in six cavities each, Dyract was combined with Prime and Bond 2.0, and Vitremer with Scotchbond Multi-Purpose Plus for assessment after 15 minutes. Maximum marginal gap widths (MGW) were measured. One-way ANOVA by ranks (Kruskal-Wallis-Test) followed by Wilcoxon's Two-Sample test were used to study the statistical difference of MGW among the treatment groups at a rejection level P = 0.05. RESULTS Neither the conventionally placed material systems nor the restorations in combination with adhesives showed consistently gap-free margins after 15-minute water storage. After 24-hour storage with Vitremer 4 of the 6 restorations were gap-free, whereas with all other materials only perfect margins were registered. There was, however, no significant difference between the groups. Application of the experimental and/or the specific resin bonding agents had no effect on early MGW except for the Photac-Fil group, which was significantly reduced.