The relationship between various types of fiber and colorectal cancer risk was investigated using data from a case-control study conducted between January 1992 and June 1996 in Italy. The study included 1953 cases of incident, histologically confirmed colorectal cancers (1225 colon cancers and 728 rectal cancers) admitted to the major teaching and general hospitals in the study areas and 4154 controls with no history of cancer admitted to hospitals in the same catchment areas for acute nonneoplastic diseases. Dietary habits were investigated using a validated food frequency questionnaire. Odds ratios (ORs) were computed after allowance for age, sex, and other potential confounding factors, including physical activity and protein, fat, and carbohydrate intake. Fiber was analyzed both as a continuous variable and in quintiles. For most types of fiber, the OR of colon and rectal cancers was significantly below 1, and no appreciable differences emerged between the two. When the unit was set at the difference between the upper cutpoints of the fourth and first quintile, i.e., the 80th and 20th percentiles, the ORs for colorectal cancer were 0.68 for total fiber (determined by the Englyst method as nonstarch polysaccharides), 0.67 for soluble noncellulose polysaccharides (NCPs), 0.71 for total insoluble fiber, 0.67 for cellulose, 0.82 for insoluble NCPs, and 0.88 for lignin. When fiber was classified according to the source, the OR was 0.75 for vegetable fiber, 0.85 for fruit fiber, and 1.09 for cereal fiber. The ORs were similar for the two sexes and the strata of age, education, physical activity, family history of colorectal cancer, and energy intake. Likewise, no appreciable differences emerged when subsites of the colon and rectum were investigated separately. This study provides additional support for a protective and independent effect of fiber on colorectal cancer, particularly for cellulose and soluble NCPs, and of fiber of vegetable or fruit origin.