To determine the degree of surface staining of resin-based composites (RBCs) and glass-ionomer cements (GICs) after immersion in various stains and food-simulating solutions (FSS). Six tooth-coloured restorative materials were used: a light-cured microfilled RBC (Durafil, Kulzer), a light-cured microglass RBC (Charisma, Kulzer), a polyacid-modified RBC (F2000, 3M/ESPE), a conventional GIC (Fuji IX, GC) and two resin-modified GICs (Fuji II LC, GC; Photac Fil, 3M/ESPE). Disk-shaped specimens were prepared and tested with either a matrix finish or polished using wet silicon carbide papers up to 2000 grit. All specimens were immersed in 37 degrees C distilled water for 1 week, followed by three different FSS (water, 10% ethanol, Crodamol GTCC) and five stains (red wine, coffee, tea, soy sauce and cola) for a further 2 weeks. Three specimens of each material for each stain were tested. Colour coefficients (CIE L* a* b*) were measured by a spectrophotometer after each treatment. The change in colour (DeltaEn) was calculated using the formula: DeltaEn=[(DeltaLn+(Deltaa(n))2+(Deltab(n))2]1/2. Distilled water caused no perceptible colour change as tested by ANOVA and Tukey's tests. The effect of surface finish on staining was not statistically significant (P>0.05). There was no strong interaction between FSS and stains or between FSS and materials. There was a strong interaction between surface and material, and stain and material (P<0.001). All materials were susceptible to staining by all stains especially coffee, red wine and tea; Fuji IX showed the least susceptibility and F2000 the greatest.