Immediate placement of implants into fresh extraction sockets would have the principal advantage of decreasing the recommended period of healing. It also would result in a guided placement of the implant, and it could reduce the resorption of the alveolar bone in the extraction area. However, when an implant is placed immediately into an extraction socket, it may not engage the walls of the socket near the crest of the alveolar ridge. With the presence of a bone defect around an implant, ingrowth of soft tissue could compromise the achievement of osseointegration in the crestal bone area. The objective of this study was to evaluate the crestal bone healing response adjacent to implants placed immediately into fresh extraction sockets with and without covering membranes. Eight adult mongrel dogs had the third and fourth mandibular premolars extracted bilaterally. Thirty-two submerged titanium hollow-screw implants were inserted immediately into the extraction sockets. On the right side, the implants were covered with an expanded polytetrafluorethylene membrane, whereas the left side served as a control. One dog was killed after 2 weeks, one after 4 weeks, and six after 12 weeks. Soft tissue dehiscence developed over 10 implants (12-week dogs) covered with membranes. Dehiscence was noted histologically over three contralateral control implants. When soft tissue dehiscence occurred and the membrane was left exposed without oral hygiene during healing, the degree of bone integration was significantly less than in the control sites without membranes.(ABSTRACT TRUNCATED AT 250 WORDS)