This investigation evaluated the performance of a resin-modified glass ionomer, a compomer, and a bonding system/composite combination for the restoration of cervical erosion lesions without cavity preparation. Forty-eight lesions (11 patients) were restored with a bonding agent/composite combination [Prime & Bond 2.1/PrismaTPH (P & B/TPH); DeTrey/Dentsply], a compomer (Dyract; DeTrey/Dentsply), or a light-curing glass ionomer (Fuji II LC; Fuji). The materials were randomly assigned to the patients in triplets. No cavity preparation was performed. The procedures strictly followed the manufacturers' instructions. The restorations were evaluated clinically, using modified USPHS criteria, and by quantitative scanning electron microscope (SEM) analysis, at baseline and 12 months. The clinical data were statistically evaluated with the Pearson chi-square test, the SEM data (criterion gap formation) were analyzed with the Mann-Whitney U-test and error rates method. Clinically, two restorations could not be evaluated. One Dyract restoration failed. With respect to marginal discoloration, recurrent caries and contour, no significant differences could be found between the materials. The surface texture of P & B/TPH and Dyract was significantly better than that of Fuji II LC at baseline and 12 months. Compared to P & B/TPH and Fuji II LC, Dyract revealed a significant decrease in marginal integrity between baseline and 12 months. In SEM analysis, gap formation was determined as follows: baseline, enamel interface: 4% Dyract= 4% Fuji >2% P & B/TPH and dentin interface: 11% Dyract >9% P & B/TPH >2% Fuji; 12 months, enamel interface: 15% Dyract >4% Fuji >3% P & B/TPH and dentin interface: 11% P & B/TPH >6% Fuji >5% Dyract. The error rates method revealed no significant differences, in general, between the three materials with regard to gap formation. In conclusion, the restorations of erosion lesions with different classes of adhesive materials were well retained after 12 months. None of the materials studied revealed superiority over the other materials. All materials revealed shortcomings with respect to either surface texture, marginal integrity or color stability clinically and for all materials gap formation was recorded in the SEM evaluation.