Epidemiological studies have identified several factors that are likely to have a major effect on reducing rates of cancer: reduction of smoking, increased consumption of fruits and vegetables, and control of infections. Other factors include avoidance of intense sun exposure, increased physical activity, and reduced consumption of alcohol and possibly red meat. Risks of many types of cancer can already be reduced, and the potential for further reductions is great. In the United States, cancer death rates for all cancers combined are decreasing, if lung cancer (90% of which is due to smoking), is excluded from the analysis. We review the research on causes of cancer and show why much cancer is preventable. The idea that traces of synthetic chemicals, such as DDT, are major contributors to human cancer is not supported by the evidence, yet public concern and resource allocation for reduction of chemical pollution are very high, in part because standard risk assessment uses linear extrapolation from limited data in high-dose animal cancer tests. These tests are done at the maximum tolerated dose (MTD) and are typically misinterpreted to mean that low doses of synthetic chemicals and industrial pollutants are relevant to human cancer. About half the chemicals tested, whether synthetic or natural, are carcinogenic to rodents at such high doses. Almost all chemicals in the human diet are natural. For example, 99.99% of the pesticides we eat are naturally present in plants to ward off insects and other predators. Half of the natural pesticides that have been tested at the MTD are rodent carcinogens. Cooking food produces large numbers of natural dietary chemicals. Roasted coffee, for example, contains more than 1000 chemicals: of 27 tested, 19 are rodent carcinogens. Increasing evidence supports the idea that the high frequency of positive results in rodent bioassays is due to testing at the MTD, which frequently can cause chronic cell killing and consequent cell replacement-a risk factor for cancer that can be limited to high doses. Because default risk assessments use linear extrapolation, which ignores effects of the high dose itself, low-dose risks are often exaggerated.