This paper presents examples of epidemiological evaluation of exposure- and cluster-type mutations in human populations. The self-poisoning model did not show that offspring born from mothers after a semi-lethal self-poisoning had higher rates of prenatal selection (fetal death) or abnormalities due to germinal mutations; however, an intrauterine growth retardation was found. The surveillance function of the Hungarian Congenital Abnormality Registry has detected many cluster-type situations, one example of which is an extreme increase of Down's syndrome in a small Hungarian village in 1989 to 1990. Environmental investigations have pointed to the excessive use of trichlorfon at local fish farms as the cause.