OBJECTIVE Different methods of internal and external fixation are used to treat aseptic posttraumatic nonunion of the femur and tibia. The advantages and disadvantages of the different methods will be demonstrated by analysing the clinical course and the outcome of our patients. Utilizing these data, a therapeutic concept tailored to the individual situation is recommended. METHODS Depending on the form of reaction we distinguish between vital and non-vital nonunions. The classification is made according to the clinical course, x-ray-findings and in special cases the results of scintigraphy. Due to anatomic differences in vascularisation and soft tissue coverage nonunion of the femur and the tibia are discussed separately. Stabilisation is achieved by intramedullary nail, plate or external fixator. As new methods the internal plate fixator was used for the femur and the Ilizarov ring fixator for the tibia. If there has been a mistake in the choice of the method of the primary stabilisation a change of method is done. If the indication for the initial method of stabilisation was correct, the therapy of nonunions is limited to the correction of technical mistakes. Additionally, a biologic stimulation is required for the therapy of non-vital nonunion. RESULTS The clinical data of 77 patients treated from 1985-1993 were analysed retrospectively. Vital nonunions of the femur (11) healed after 9.5 months on the average, those of the tibia (49) after 10 months. The duration of treatment of non-vital nonunions was much longer and required 20 and 16 months, respectively. The treatment of two non-vital nonunions of the tibia could not be completed. For the femur only intramedullary nail (4) and plate (8) were used, for the tibia mainly the fixator (43), of these in 18 cases the Ilizarov-apparatus. Differences in the duration of treatment due to the choice of implant could not be recognized. Complications were pin problems (14) and one lesion of the peroneal nerve in the fixator group and superficial wound infection (2), nerve irritation (1) and fracture (1) in the group treated with intramedullary nailing. CONCLUSIONS Due to the good soft tissue coverage and vascularisation internal fixation is favored for the treatment of femoral nonunions. The fixator should only be used if distraction osteogenesis is necessary because of a bony defect. Due to the problematic soft tissue situation and poorer vascularisation on the external fixator is preferred in the treatment of tibial nonunions if a change of method is indicated. For this purpose, we currently use predominantly the Ilizarov-apparatus because of its biomechanical properties and the convincing results. Initial problems with its use could markedly be reduced with growing experience.