A pelvic radiographic examination was performed on 15 large breed dogs without history and clinical sign of hip dysplasia. The effect of anaesthesia and of two stress-radiographic methods on the coxofemoral subluxation was evaluated. With anaesthesia a mild coxofemoral subluxation was seen in 31 per cent of the dogs, which all appeared radiographically normal when sedated. The Norberg angle was significantly decreased in anaesthetised dogs (P < 0–05). When dogs with a radiographically normal coxofemoral joint conformation were submitted to a stress-radiographic examination there was a shift towards an increased coxofemoral subluxation. Subluxation was seen in 70 per cent of the dogs when the knee fixation method of positioning was applied and in 100 per cent of the dogs when the wedge method was used. Correspondingly the Norberg angle decreased significantly when the coxofermoral joints were stressed (Pkneefix < 0–00003; Pwedge= 0–000008). The study concludes that the currently used procedure for evaluation of the hip joints does not necessarily disclose whether or not a dog has hip dysplasia.