We estimated risks for concordant and discordant cancers in spouses in order to quantify cancer risks from the shared environment. The study was restricted to spouses who had one or more children in common and who lived together for at least 15 years after the first child's birth. The nation-wide Family-Cancer Database was used as the source of family and cancer data. Standardised incidence ratios were calculated for concordant and discordant cancers in spouses after 50 years of age. Among the 18 cancer sites considered, only three sites, stomach, lung and bladder, showed concordant increases of cancer among spouses, standardised incidence ratios ranging only from 1.19 to 1.38. Additionally, gastric and pancreatic cancer were associated among spouses, as did many cancers which were related to tobacco smoking or human papilloma virus infection. By contrast, standardised incidence ratios of colon, rectal, renal and skin cancers showed no increases among spouses. Shared lifestyle among family members seems to explain only a small proportion of familial cancer susceptibility. Because lifestyles are likely to differ more between parents and offspring than between spouses, familial cancer risks between parents and offspring are even more likely to be due to heritable than environmental effects. (Doll and Peto, 1981; IARC, 1990;Lichtenstein et al, 2000;Peto, 2001). During the past decade it has become increasingly clear that overweight and lack of physical activity convey a risk of cancer, which may account for 5% of all cancers in Europe (Bergstrom et al, 2001; IARC, 2002). Moreover, the risks at the population level caused by various infections have become better understood, and the known infections have been estimated to account for 15% of cancer worldwide, though less in Europe (Pisani et al, 1997;Zur Hausen, 1999). In spite of the enormous research effort on diet and cancer, the proportion of cancer attributable to diet or to any specific dietary component remains speculative. It has been estimated that at least 50%, and probably as much as 70% of cancer deaths are unavoidable among non-smokers mainly because their aetiology remains unknown (Peto, 2001).Decades long cohabitation by spouses should tend to result in many habits and carcinogenic exposures being similar. Interest in disease among spouses earlier focused on sexually transmitted diseases and the effects of passive smoking (IARC, 1995;Hackshaw, 1998;Hemminki et al, 2000a;Hemminki and Dong, 2000a). Besides assessing life-style factors and cancer risks, they can point to the environmental contribution to the familial aggregation of cancer, and they thus help to apportion heritable effects (Hemminki et al, 2001a,d). The studies from the Swedish Family-Cancer Database have shown limited spouse concordance, affecting mainly the sites of known environmental carcinogens (Hemminki and Dong, 2000b;Hemminki et al, 2001a). However, in the previous studies the length of cohabitation between the spouses was not considered, nor were any adjustments for socio-economi...