Plant foods have been associated inversely with colon cancer. Since a major focus of this study was to identify components of plant foods which may account for their association with colon cancer, nutrients which are commonly found in plant foods also were evaluated. A population-based case-control study was conducted in Northern California, Utah, and the 'Twin Cities' area of Minnesota (United States). Complete data were available from interviewer-administered questionnaires on 1,993 cases and 2,410 controls. Higher intakes of vegetables (for highest relative to lowest quintile of intake) were associated inversely with colon cancer risk: the odds ratio (OR) was 0.7 for both men (95 percent [CI] confidence interval = 0.5-0.9) and women (CI = 0.5-1.0). Associations were stronger among those with proximal tumors. Total fruit intake was not associated with colon cancer risk although, among men, higher levels of whole grain intake were associated with a decreased risk (OR = 0.6, CI = 0.4-0.9 for older men); high intakes of refined grains were associated with an increased risk (OR = 1.5, CI = 1.1-2.1). Dietary fiber intake was associated with a decreased risk of colon cancer: OR = 0.5 (CI = 0.3-0.9) for older men; OR = 0.7 (CI = 0.4-1.2) for older women; OR = 0.6 (CI = 0.4-1.0) for men with proximal tumors; OR = 0.5 (CI = 0.3-0.9) for women with proximal tumors. Other nutrients, for which plant foods were the major contributor--such as vitamin B6, thiamin, and niacin (women only)--also were associated inversely with colon cancer. Neither beta-carotene nor vitamin C was protective for colon cancer. Adjustment of plant foods for nutrients found in plant foods or for supplement use did not appreciably alter the observed associations between plant foods and colon cancer.