The preferable dietary method for epidemiologic studies of diet and cancer is the diet history, which provides individual estimates of the usual intake of foods and nutrients during a fairly long time period. The method permits respondents to indicate their usual frequencies and amounts of selected items. Comparison of the food and nutrient intakes between patients with a particular cancer and controls who are cancer free has demonstrated associations of dietary fat with colorectal, breast and prostate cancers in the multiethnic population of Hawaii. Similarly, protective relationships of dietary fiber with colorectal cancer and of carotenoids and vegetables with lung cancer have also been shown in this population. Further analysis suggests that the incidence of breast and prostate cancers in Hawaii could be reduced by decreasing the intake of saturated fat, whereas increasing the consumption of fruits and vegetables could have a beneficial effect on lung cancer incidence in Hawaii.