This study compared the demineralization inhibition properties of fluoride releasing tooth-colored restorative materials. Materials evaluated included a giomer (Reactmer, Shofu [RM]), a conventional glass ionomer (Fuji II, GC [FJ]), a resin modified glass ionomer (Fuji II LC, GC [FL]) and a compomer (Dyract AP, Dentsply [DY]). A non-fluoride releasing composite (Spectrum TPH, Dentsply [SP]) was used for comparison. Class V preparations on buccal and palatal/lingual were made at the CEJ of 75 freshly extracted molars. The teeth were randomly divided into five groups of 15 and restored with the various materials. The occlusal half of each restoration was in enamel, while the gingival half was in dentin. The restored teeth were stored in distilled water at 37 degrees C for two weeks and subjected to artificial caries challenge (18 hours demineralization [pH 5.0] followed by six hours of remineralization [pH 7.0]) for three days. Sections of 130 +/- 20 microm were examined with a polarized light microscope, and outer lesion depth [OLD] and wall area [WA] lesion/inhibition measurements were made using image analysis software. All data were subjected to statistical analyses at 0.05 significance level. For the various materials, OLD ranged from 54.55 to 65.86 microm and 124.68 to 145.97 microm in enamel and dentin, respectively. WA ranged from -2356.13 to 1398.20 microm2 and -3011.73 to 5095.80 microm2 (positive values indicate wall inhibition, negative values indicate wall lesion) in enamel and dentin, respectively. Results of ANOVA/Scheffe's post-hoc test (p FJ, FL & RM; Enamel WA inhibition--FJ, FL & RM > DY & SP and Dentin WA inhibition--FJ > FL > RM > DY > SP. The demineralization inhibition effect of giomers, conventional and resin-modified glass ionomer cements appear to be more evident at the margins of restorations.