Methods. Mortality rates for four types of cancer (stomach, colorectal
, lung, and breast cancer) in migrant populations were compared to tho
se of individuals who still resided in the political region in which t
hey were born. The effects on mortality rates of place of birth and of
place of residence were studied, comparing different regression model
s. Results. Overall, people who were born in the South and who later m
igrated had significantly higher mortality rates than the southern pop
ulation, but lower than the population in the area of residence, for m
ost cancers. Place of birth and place of residence showed different po
wer in explaining the observed mortality rates for different cancer si
tes: place of birth was a stronger predictor for stomach and breast ca
ncers, while residence was a stronger predictor for lung and colorecta
l cancers. The status of 'migrant' was found to be an overall risk fac
tor. The compatibility of the results obtained with different aetiolog
ical hypotheses is discussed.