We have carried out a retrospective study, over a 2.5 year period, of the results of compression-fixation for diaphyseal long bone fractures (forearm 49, tibia 18, others 9). The method was used in forearm fractures mainly for fresh fractures, whilst in fractures of the tibia it was predominantly for delayed union. In forearm fractures, although 10 per cent showed failure of fixation, only 2 per cent developed infection, and 90 per cent recovered a full range of movement. Conversely, plating of fractures of the tibia with delayed union resulted in a 37 per cent deep infection rate. It is concluded that while compression-fixation gives excellent results when applied to displaced forearm fractures, alternative methods should be used in fractures of the tibia with delayed union. This unexpected finding illustrates the value of periodic audit of surgical results.