Since 1923, when Wells challenged the reliability of existing cancer statistics,1 over 100 publications documenting discrepancies (ranging from 20 to 40 per cent) between major clinical and autopsy diagnoses have repeatedly raised questions about the accuracy of published mortality statistics. In 1983, Goldman et al., in a variance study, emphasized the continuing value of the autopsy, even in an era of diagnostic technologic miracles.2 Specious mortality statistics have obvious implications for the biomedical and health services community, the government, funding agencies, and the public at large. The immediate functions of the death certificate are legal: to permit disposal of the . . .