A population-based case-control study was conducted from 1985 to 1989 in western Washington State to assess the relation between nutrients and the incidence of colon cancer in men and women aged 30-62. A food frequency questionnaire was used to document the usual diet 7 years before diagnosis for 424 cases and at a similar time for 414 controls. Alcohol consumption was strongly related to the risk of colon cancer in both men and women, with age-adjusted odds ratios (ORs) of colon cancer = 1.0, 1.9, 1.7, and 2.6 for 0, < 10, 10-29, and > or = 30 g/day intake for men and adjusted odds ratios = 1.0, 1.3, 1.8, and 2.5 for the same categories for women. The trend odds ratio associated with a one-category increment in the four-level alcohol consumption variable was 1.3 (95% confidence interval (CI) 1.0-1.5) in men and 1.4 (95% CI 1.0-1.7) in women. For both sexes, higher dietary fiber intakes were associated with lower relative risks for colon cancer, with age-, energy-, and alcohol-adjusted odds ratios = 1.0, 0.9, 0.8, and 0.6 across quartiles of consumption for men (trend OR for a one-quartile increment = 0.8, 95% CI 0.7-1.1) and adjusted odds ratios = 1.0, 0.9, 0.5, and 0.5 for women (trend OR = 0.8, 95% CI 0.6-1.0). In men, this was mostly attributable to intake of cereal fiber (trend OR = 0.8, 95% CI 0.6-1.0) while, in women, this association mostly reflected the effect of fruit fiber (trend OR = 0.8, 95% CI 0.6-0.9) and vegetable fiber (trend OR = 0.8, 95% CI 0.7-1.0). Calcium was associated with a decreased risk of colon cancer among women only (adjusted OR across quartiles = 1.0, 0.5, 0.6, 0.3; trend OR = 0.7, 95% CI 0.6-1.0). There was no indication of an association between colon cancer and fat or protein consumption or dietary vitamins.