A case-control dietary study of 198 patients with cancer of the colon and two matched control groups demonstrated a significantly lower fiber consumption frequency among the cancer patients. This difference was not confined to a few items. Of the 73 items on the fiber list, 61 were eaten less often by the cancer patient than by a neighborhood control, and 57 were consumed less frequently than by a surgical control. These findings support the hypothesis that low-residue foods play an etiologic role in colon carcinogenesis. A mechanism related to the possible potential carcinogenic properties of degraded biliary compounds may be implicated.