Polyurethane-coated and uncoated polytetrafluoroethylene (PTFE) (Teflon(R)) and polyester (Dacron(R)) felts were used for repair of full-thickness cartilage defects in the rabbit knee. At 3 months the indentation characteristics and the histological appearance of the repairs were compared with those of a sham-operation, natural repair, and periosteal grafting. Joint compressive load-displacement and indentation characteristics of the cartilage adjacent to the defect were normal for all alternatives. All the repair sites had a higher compliance than had normal cartilage-bone, but synthetic grafting had values closer to normal than periosteal grafting. The adjacent cartilage appeared macroscopically normal, except with periosteal grafting and coated PTFE implants where it showed surface irregularities in some cases. With periosteal and uncoated synthetic implants the defects were completely filled, but not in the natural repair sites. Most of the coated implants failed by separation within the implant. On histological examination, ingrowth of trabecular bone from the base of the defect into the synthetic material was seen in all the specimens. The neocartilage 'score' was equally low with all the repairs reaching only one-third of the maximum points. All the repairs were associated with synovitis. Further, some of the knees with the synthetic materials repair had debris particles in the synovium. We conclude that none of the biological resurfacing techniques described achieved normal articular cartilage characteristics at 3 months, although the compliance of the repair site was closer to normal with synthetic than with periosteal grafting.