The difference between injury scaling performed in the same patients on the basis of clinical information only and post-mortem examination only is largely unknown. We compared scores in all 279 trauma patients who died in the Department of Critical Care Medicine at Auckland Hospital from 1982 through 1987 (93% blunt trauma, 4% penetrating trauma, 3% burns, median time until death; 2 days) using the 1985 revision of the Abbreviated Injury Scale (AIS-85) and derived Injury Severity Scores where such scoring was based on clinical information only (CLAIS, CLISS) or post-mortem findings only (PMAIS, PMISS). For the group as a whole there was little difference in the distribution of scores between CLAIS and PMAIS or between CLISS and PMISS. However CLISS was different from PMISS in 69% of individual patients. Most major differences between CLAIS and PMAIS (two AIS grades or more) occurred in the Head region, where injury scoring based on physiological features (e.g. coma) occurred without an anatomic injury of similar AIS grade, or in the Thorax region where therapy had either abolished the evidence of injury (e.g. pneumothorax) or injuries were discovered at postmortem examination which were not appreciated clinically. Injury scaling data derived only from postmortem examination is not equivalent to that derived clinically. For maximum accuracy post-mortem data must be derived from an examination specifically guided by the needs of injury scaling and in full cognizance of injuries recognised and treated clinically.