The mortality in a series of 6,032 homeless men in Stockholm was investigated. The observation period was up to 3 years. The observed number of deaths amounted to 327 compared to the expected 86.7, corresponding to an excess mortality ratio of about 4. Mortality was raised for all the causes of death studied. Excess mortality from suicide was lower than that found in some earlier studies. This discrepancy is discussed. (10 refs)