The relationship between various types of fibre and colorectal cancer risk was investigated using data from a case-control study conducted in the Swiss Canton of Vaud between January 1992 and December 2000. The study included 286 cases of incident, histologically-confirmed colorectal cancers (149 colon and 137 rectal cancers) admitted to the University Hospital of Lausanne, and 550 controls whose admission diagnosis was of acute, non-neoplastic diseases. Dietary habits were investigated using a validated food frequency questionnaire (FFQ). Odds ratios (ORs) were computed after allowance for age, sex, education, physical activity and energy intake. Fibre was analysed both as a continuous variable and in tertiles. There was a significant inverse relationship of total fibre intake (determined by the Englyst method as non-starch polysaccharides) and of its components with the risk of colorectal cancer. ORs for a difference in intake of one standard deviation from the mean fibre intake of the control distribution was 0.57 for total fibres, 0.55 for soluble non-cellulose polysaccharides (NCPs), 0.58 for total insoluble fibres, 0.57 for cellulose, 0.62 for insoluble NCP and 0.62 for lignin. When fibre was classified according to its source, the OR was 0.60 for vegetables, 0.78 for fruit and 0.74 for grain fibre. The ORs were similar for colon and rectal cancer and consistent across the strata of the major covariates and of several types of fibres.