Background. Each year, about 90 million new restorations are placed in the United States and 200 million are replaced. Controversy surrounds the pulpal reactions and frequency of bacterial microleakage associated with common restorative materials. The authors investigated and compared pulpal reactions to different types of restorative materials. Methods. Two hundred seventy-two teeth with standardized rectangular Class V unexposed cavities were restored with resin-based composite bonded to dentin; resin-based composite bonded to enamel; resin-modified glass ionomers, or RMGI; amalgam lined with zinc polycarboxylate, or ZnPC; amalgam lined with calcium hydroxide, or Ca(OH)(2); or zinc oxide-eugenol, or ZnOE. Teeth were extracted for orthodontic reasons between 20 and 381 days later. The authors categorized pulpal responses according to standards set by the Federation Dentaire Internationale and the International Organization for Standardization. Bacteria were detected using Brown-Brenn-stained sections. Pulpal responses were evaluated using histomorphometric analysis and analysis of variance statistics. Results. The results showed that RMGI was the best material for preventing bacterial microleakage, and resin-based composite bonded to enamel was the worst. In regard to minimizing pulpal inflammatory activity, ZnOE was the best material and resin-based composite bonded to enamel was the worst. In terms of maximizing odontoblast survival beneath deep cavity preparations, Ca(OH)(2), was the best material and RMGI was the worst. Conclusions. The results show that bacterial microleakage, pulpal injury and repair responses varied widely with different restorative materials. Clinical Implications. The authors recommend that RMGI be used to restore teeth with cavities that are shallow to moderate in depth, with the floor of deep cavities being lined with Ca(OH)(2) before the teeth are restored with RMGI.