Epidemiological studies suggest whole grain consumption is associated with decreased colon cancer risk. However, wheat classes differ in their usage, making the influence of wheat class (red vs. white) and state of refinement (whole vs. refined) difficult to separate, leading to potential confounding. Consequently, we examined the effect of wheat class and refining on colonic precancerous lesions (aberrant crypt foci; ACF) in carcinogen-treated rats. Diets contained wheat flour as whole soft white, refined soft white, whole hard red, or refined hard red. Feeding was begun 10 days prior to carcinogen treatment and continued for 9 wk. Hard red-fed groups had significantly fewer ACF than soft white-fed groups (25-32% reduction, P = 0.013). However, state of refinement had no significant effect on ACF number. Cecal contents supernatant oxygen radical absorbance capacity and fecal bile acid concentration were significantly greater in whole wheat-fed groups than refined wheat-fed groups and did not correlate with ACF number (increase of 21-22%, P < 0.001 and 55-56%, P < 0.001, respectively). Consequently, wheat class, not state of refinement, appears to influence colon cancer risk, with hard red wheat protective relative to soft white wheat. Thus, epidemiological associations of reduced colon cancer risk with whole grain consumption may actually reflect different wheat classes.