After a major disaster, identification of the victims may have to be performed by comparison of the antemortem and post-mortem dental records. Ante-mortem records often contain errors and these must be identified, and compensated for in some way. A computer system for the comparison of large numbers of ante-mortem and post-mortem dental records has been developed, using algorithms which are able to compensate for errors in the records. Data has been collected on the prevalence of errors in dental charts. These were found in 45% of a sample of 50 records. Since matches found by the system may be due to coincidence, a score is computed for each record based on its 'uniqueness'. This allows a confidence limit to be placed on any matches found between the antemortem and post-mortem records.