In the early 1980s, Doll and Peto estimated that about 35% of all deaths from cancer in the United States were attributable to dietary factors, with a margin of uncertainty ranging from 10 to 70%. Since then, several dietary factors, e.g. fat and meat, have been suggested to increase the risk for cancer, while other factors, e.g. fibre, fruit and vegetables, have been suggested to decrease the risk. The case-control and cohort studies have, however, given ambiguous results, and the overall evidence is far from conclusive. The major findings on dietary factors that increase risk have been reported from case-control studies, but have not been confirmed in large population-based cohort studies. Although the research in this area indicates that diet is important in cancer prevention, current knowledge does not allow reliable estimates of the numbers and proportions of cancers that could be avoided through well-described modifications of dietary habits. During the last 10 years, low physical activity has been pinpointed as a risk factor for cancers at various sites, especially the colon; however, the causal mechanism is still unknown. Obesity, defined as a body mass index of 30 or more, is consistently associated with endometrial and gall-bladder cancers in women and renal-cell cancer in both men and women. As the prevalence of obesity was between 5 and almost 20% in the Nordic populations in 1995, 625 cancer cases (310 endometrial cancers, 270 renal-cell cancers in men and women and 45 gall-bladder and bile-duct cancers among women) can be predicted in the Nordic countries around the year 2000 to be caused by obesity. This implies that about 1% of all cancers in Nordic women and less than 1% of those in Nordic men could be avoided around the year 2000 if a healthy body weight could be maintained by all inhabitants.