A case-control study was carried out in the populations of two Belgian provinces that differed in food consumption habits, particularly fat intake. There were 453 colonic and 365 rectal cancer cases and 2,851 population controls. There were no significant differences in the average intake of the major nutrients, with the exception of carbohydrates; patients had a larger intake, limited to oligosaccharides. In both provinces, in males and females having cancer of the colon or the rectum, the intake of linoleic acid was lower than among controls; for dietary fibers, a smaller intake was observed among patients in one province. The relative risks were computed for four levels of daily intake of each nutrient. A positive trend was found for oligosaccharides in all subgroups, and a negative trend was found for polysaccharides, the latter for colon cancer patients only. There was a constant and significant negative trend for linoleic acid, with a similar negative trend for dietary fiber. None of these trends were affected by further adjustment for total calorie intake. For several vitamins and minerals, less marked, less constant effects were observed. They tended to be negative for vitamin B1, vitamin B6, and iron; they were positive for retinol and vitamin B2.