Cancer mortality from 1986 to 1992 was examined in a cohort of 4419 individuals who had been residing in an area of the municipality of Reggio Emilia, northern Italy, where tap water with unusually high selenium content was accidentally supplied. Mortality for all cancers was not significantly different, both in males and in females, from that expected using death rates in the remaining municipal population as standard rates. No significant difference in mortality for site-specific cancers was observed in males, while in females a higher mortality for malignancies of the lymphatic-hematopoietic tissue overall considered and for non-Hodgkin's lymphoma was detected. Even if evaluation of the results is hampered by the low number of cancer deaths on which the analysis is based, findings of the study do not support the hypothesis of a strong inverse independent relationship between dietary intake of selenium and cancer mortality in humans.