The effects of three root-end filling materials on healing following endodontic surgery were assessed radiologically and correlated with histological findings reported elsewhere. The materials compared were a light-cured glass ionomer cement (Vitrebond), a reinforced zinc oxide-eugenol cement (Kalzinol) and amalgam. The root canals of 27 two-rooted mandibular premolar teeth of six beagle dogs were inoculated with endodontic pathogenic bacteria to induce periradicular lesions. The roots were apicected and root-end cavities filled with the tested filling materials. The teeth and surrounding jaw were removed after 4 weeks (30 roots) or 8 weeks (24 roots). Radiographs were taken of each jaw section and subjected to image analysis. Healing was evaluated based on measurements of the size of the periradicular radiolucent areas. ANOVA disclosed no statistically significant differences in the size of the periradicular areas either between time periods or between materials. These results did not correlate with the tissue responses in the same material as assessed histologically and previously reported. The use of radiographs alone to assess healing after endodontic surgery in the dog mandible is unsatisfactory, and should not be regarded as a substitute for histological examination for the determination of healing.