Persons suffering from age psychosis are known to have a high mortality rate. During the last decades there has been a general improvement in the standard of living and the availability of medical resources for the elderly in most western countries. It has been suggested that persons with a diagnosis of age psychosis have benefited from these changes and live longer with their illness than they did before. The Lundby cohort comprises 3,563 persons from a total population followed concerning mental disorders (psychiatrically treated as well as untreated) for 15 or 25 years. In the present Lundby Study we have calculated the changes over time concerning death risk among persons with senile and multi-infarct dementia and overmortality associated with these two main subgroups of age psychosis. We found that the prognosis in terms of mortality had not undergone any statistically significant change during the 25-year period 1947— 1972 among persons in the Lundby cohort with a diagnosis of senile and multi-infarct dementi