Experimental knee-implant infection was induced in the rabbit knee with a strain (1369,phage type 85,a laboratory strain) of Staphyloccus aureus. The experimental model was adapted from an experimental study in rabbits by Gudmund Blomgren, 1981. In these experiment's osteomyelitis was induced with an noculum of 100,000 bacteria in the tibia and femur. In one series we used gentamicin impregnated bone cement and in the other series dicloxacillin was given intravenously. The controlgroup, four rabbits had no supplemental antibiotics. The animals were monitored by clinical observation, microbiological, histological and antibody methods. Only in the controlgroup without supplemental antibiotics we would see manifest inflammation around the prosthesis with increased antibody titre. With supplemental antibiotics no bacteria was found around the prosthesis.The gentamicin concentration was measured in bone from femur, tibia and from the jointfluid. A high level of gentamicin was found in the bone but not in the joint fluid. The animal model was excellent to create a model for human total joint replacement to study the dissemination of localy and intravenously injected bacteria to the artificial joint. The use of either systemic or locally administered antibiotics as prophylaxis avoided development of infection.