Genetic epidemiology provides data on cancer etiology, familial risks and genotype-specific risks. These data are useful for clinical counselling and gene identification. The studies require large, unbiased sample sizes and collaboration between research teams, nationally and internationally. A recent study on Nordic twins suggests that in colorectal, breast and prostate cancer, the inherited component ranges between 27 and 42%, far in excess of the known susceptibility genes. The data from the Swedish Family-Cancer Database, particularly on second cancers, also suggest that a main genetic component in cancer is polygenic. The results have implications for design of genetic studies and for clinical counselling.