Several dietary exposures may increase the risk of cancer, while others may have a protective effect. The degrees of evidence for a casual relationship between these exposures and human cancer vary considerably, as do their suitability for prevention. Investigations on the mechanisms by which food components contribute to either increase or decrease cancer risks therefore deserve priority, as their results would allow a quantitative evaluation of risks and of their preventability. Dietary factors are likely to contribute directly or indirectly to the induction of cancer in a variety of organs--namely, oesophagus, stomach, colon and rectum, liver, breast and endometrium, as well as the oral cavity and larynx. In addition, certain dietary factors may contribute to the prevention of cancer at other sites, as, for instance, the lung and prostate. Dietary interventions may therefore have a very considerable impact on prevention. This is certainly not the least reason for the attraction that intervention studies exert on scientists. Attractive as they may be, however, they should not encourage short-cuts, in the belief that understanding of the means for the prevention of cancer might be easier than understanding of the mechanisms of its induction. This very timely symposium has exposed the controversies that still exist about certain basic assumptions on the role of dietary factors in human cancer, and has underlined the importance of a multidisciplinary approach to understanding of the underlying mechanisms.