Migrant studies have taken advantage of the wide geographical variation in cancer risk. Cancer rates in migrants, obtained from routinely collected incidence or mortality statistics, are compared with those in the host country and in the country of origin; the rate of change with time since migration (or age at migration) and in subsequent generations is assessed; and the results are interpreted in the light of differences in socio-economic status and the degree of cultural assimilation. Rapid changes in cancer risk following migration imply that life-style or environmental factors are of overriding importance in aetiology. The susceptibility of fair-skinned races to ultraviolet (UV)-associated skin cancers is an example of racial differences based on inherited factors, but the long-term excess or deficit of other cancers in migrants has not yet been attributed definitively to genetic rather than persisting life-style factors. Are there racial differences in metabolism, DNA repair mechanisms or altered expression of oncogenes or tumour suppressor genes? Several genetic polymorphisms affecting the metabolism of known occupational carcinogens or hormonal factors do vary by race. While classical epidemiology has shown that the environment predominates in determining cancer incidence, molecular epidemiology has identified several examples of genetically determined differences between races.