The purpose of this in vitro study was to evaluate the enamel surface structure subjected to various techniques of debonding orthodontic attachments and to develop a technique for residual adhesive removal that restores the enamel surface as closely as possible to its pretreatment condition without introducing iatrogenic damage. Enamel surface structure was examined with a scanning electron microscope before bonding of twin metal brackets to 60 previously extracted premolars with two heavily filled composite resins. Two groups, each consisting of 30 teeth, were equally subdivided into 10 subgroups. The first three subgroups were used to compare the efficacy of three bracket removing instruments. Since there were no differences in the debonding properties between the two resins, the two groups were combined. In this way, each subgroup (n = 6) could be used for a more meaningful comparison. On the basis of the results of this comparison, the bracket removing instrument that produced the most consistent separation at the bracket-adhesive interface was used in the remainder of the study. After appliance removal, the teeth were again examined microscopically and photographed, and seven different procedures for residual resin removal were compared. After resin removal, the final polished enamel surface was followed by microscopic evaluation. Results of this study show the bracket removing plier produced the most consistent separation at the bracket-adhesive interface, leaving the enamel surface intact. Carbide burs at high speed and air coolant proved to be efficient in residual resin removal, but when used alone, failed to produce a satisfactory enamel surface. After the removal of residual resin, graded medium, fine, and superfine Sof-Lex finishing disks (Unitek Corp., Monrovia, Calif.) produced surfaces that could be readily restored satisfactorily after receiving a final polish with a rubber cup and Zircate paste.