Pre- and post-autopsy diagnoses of underlying cause of death were compared in consecutive autopsies on persons aged 30 to 80 years; 1000 from each of two pathology departments in Budapest. Data on admission diagnoses and on contributory causes of death were also analysed. At autopsy, the percentages of deaths by underlying cause were neoplasms (any site) 34.9%, diseases of the circulatory system 40.2%, digestive system 13.8%, endocrine, nutritional, metabolic or immune systems 2.7%, and respiratory system 2.2%. For these five disease groupings, the percentages of cases diagnosed clinically as the underlying cause of death which were confirmed at autopsy were, respectively, 90.9%, 84.0%, 82.9%, 55.2% and 32.5%. Although, out of 697 cases with an autopsy diagnosis of neoplasia as the underlying cause, there were only 61 (8.8%) where neoplasms were not diagnosed clinically as the underlying cause, this conceals the fact that in 130 (18.7%) the two diagnoses differed as to the site of the primary neoplasm (ICD 3 digit code). The fact that 43% of post-mortem diagnoses (ICD major category) of underlying cause are missed on admission, and that 19% are missed clinically, indicates that improved clinical diagnostic procedures have not diminished the need for high autopsy rates. Morbid anatomy needs to be better resourced.