The Li-Fraumeni cancer family syndrome is manifested by susceptibility to breast cancer, sarcomas, and other neoplasms in children and young adults. The present study utilized clinical follow-up data on 545 members of 24 Li-Fraumeni kindreds living and cancer-free at family ascertainment. Two hypotheses were tested based on a model of autosomal dominant genetic predisposition: (a) that syndrome cancers would continue to occur excessively during follow-up compared to the general population, and (b) that the tumors would occur primarily among those family members likely to carry the gene. Population cancer rates were compared with cancer rates in follow-up of the cohort from ascertainment to 1988. Risk of carrying the gene for the syndrome at the time of ascertainment was calculated for each family member under two models with somewhat different definitions of affection with the syndrome. Cancer occurrence after ascertainment was then analyzed according to the risks. Cancer did continue to occur excessively among the entire cohort during follow-up [relative risk (RR 2.1)]. The excess was greatest below age 20 (RR 21.1), declined with increasing age, and was most pronounced for neoplasms featured in the syndrome (RR 18.2). Among persons less than age 45, at least 87% of cancers occurred in those at higher risk of carrying the gene under both genetic models (RR 22.9 and 21.3). The clinical data, therefore, reliably identify individuals likely to carry a dominantly inherited gene conferring susceptibility to a specific constellation of neoplasms. Recent identification of a germ line mutation in the tumor suppressor gene p53 in persons with the syndrome may, if confirmed, have implications for ultimately defining the component tumors of the syndrome and for the causes and prevention of those tumors arising outside these families.