We developed a method which we used in the current study to observe the rat colon endoscopically. Our goal was to evaluate the entire course in the development of experimental large bowel tumors through serial observations in the same rat. We compared the effects of dietary lard and cellulose on rat colon tumorigenesis in a 2 x 3 factorial design. Sprague-Dawley rats (n = 90) were divided into 6 diet groups: rats were fed a diet without cellulose that contained 5, 10, and 15 g/100 g lard, or diets containing 15% cellulose diet (15 g cellulose/100 g diet) and the same concentrations of lard. The development of large bowel tumors induced by the administration of 1,2-dimethylhydrazine (25 mg/kg body weight) for 19 wk was examined endoscopically. Tumor induction rates in the 15% cellulose groups were lower than in the 0% cellulose groups (P = 0.008), independent of the lard concentration. These results suggest that the preventive effect of cellulose against large bowel tumorigenesis is greater than the promotive effect of fat under the current experimental conditions.