In this study we describe a principle for the accomplishment of bone regeneration based on the hypothesis that different cellular components in the tissue have varying rates of migration into a wound area during healing. By a mechanical hindrance, using a membrane technique, fibroblasts and other soft connective-tissue cells are prevented from entering the bone defect so that the presumably slower-migrating cells with osteogenic potential are allowed to repopulate the defect. Defects of standard size were created bilaterally through the mandibular angles of rats. On one side of the jaw the defect was covered with Teflon membranes, whereas the defect on the other side served as control. Histologic analysis after healing demonstrated that on the test (membrane) side, half the number of animals showed complete bone healing after 3 weeks and all animals showed complete healing after 6 weeks. Little or no sign of healing was evident on the control side even after an observation period of 22 weeks.