The aim of this study was to compare the effects of bioabsorbable and non-resorb-able membranes on experimental guided bone augmentation in 8 Japanese white rabbits. A cutaneous flap was demarcated and raised from the forehead of each animal, the periosteum was lifted, and the calvarial bone on both sides of the midline was exposed. A titanium screw was inserted into the bone on each side of the midline and one screw was covered with a bioabsorbable (polylactic acid) membrane and the other with a non-resorbable (expanded polytetrafluoroethylene) membrane. The implanted screws and membranes were then covered with the periosteum and cutaneous flap. After healing for 6 months, the animals were euthanized and the experimental area was prepared for histological investigation. New bone had formed under both membranes with no sign of infection or membrane exposure. The amount of newly generated bone (89.0 ± 17.3% versus 54.7 ± 14.0%, P <0.05) and the percentage of newly generated bone height (81.5 ± 6.3% versus 58.9 ± 7.8%, P <0.05) in the space beneath the non-resorbable membrane was greater than that beneath the bioabsorbable membrane. However, there were no statistically significant differences between the bioabsorbable and non-resorbable membranes with respect to the percentage areas of mineralized bone (52.3 ± 11.3% versus 47.1 ± 6.7%, P= 0.8658) and bone marrow (47.7 ± 11.3% versus 52.9 ± 6.7%, P = 0.4838) and bone contact with the screw (88.3 ± 6.9% versus 89.2 ± 7.3%, P = 0.9999). In conclusion, at least within the limitations of this rabbit model, we suggest that non-resorbable membranes with sufficient stiffness should be used to obtain greater bone volume and height instead of bioabsorbable membranes for the GBR procedure, and that this will facilitate predictable bone augmentation in spaces beyond the bone surface. Therefore, the bioabsorbable membrane could not replace the non-resorbable membrane used in this model. J Periodontol 1998;69:1229–1237.