The aim of this study was to investigate the visibility of secondary caries in the gingivobuccal and gingivolingual corne rs of teeth restored with amalgam restorations. Standard Class II cavities were created in 15 orthodontically extracted mandibular premolar teeth, and the teeth were randomly divided into five groups of three teeth each. In four of the gruops, a 1.0- or 1.5-mm cavity was prepared in the gingivolingual or gingivobuccal corner of the restoration. No lesions were created in group 5, the control group. The teeth were restored with amalgam. The teeth were adapted in the actual tooth space of 15 volunteers with one mandibular premolar missing. Radiographs o f each patient were taken with the bisecting-angle technique and the bite wing technique. The radiographs were sorted at random and given to 15 members of the professoriate who were often involved in detecting caries and and to 17 members who were not normally involved in detecting caries. The bitewing technique was found to be more reliable than the bisecting-angle technique in detecting secondary caries in gingivobuccal approximal corners (P .05). It was also found that, in group 1, the bisecting-angle technique was more reliable than the bitewing technique in detecting caries in gingivolingual corners (P .05). No significant differences were found in the correct evaluation of radiographs between the faculty who were normally involved in the detection of caries and those who where not.