A 3 x 3 factorial experiment was conducted to examine how dietary fiber (wheat bran) and fat (lard) interactively affect the genesis of N-methyl-N'-nitro-N-nitrosoguanidine-induced colon cancer in rats. Groups of 30 male 4-week-old Wistar rats were fed ad libitum one of nine experimental diets containing either 15 (low), 27.5 (medium), or 40% (high) energy as fat in combination with 0.7 (low), 2.2 (medium), or 3.8 g (high) fiber/100 kcal for a period of 37 weeks. After 4 weeks, each rat received a total of five weekly intrarectal instillations of 6 mg N-methyl-N'-nitro-N-nitrosoguanidine/kg. The highest colon carcinoma incidence and the highest total number of carcinomas of the colon were observed in the animals fed the medium-fat/medium-fiber diet. The highest number of polyps and a relatively high polyp incidence occurred in the animals on the high-fat/low-fiber diet. An enhancing effect of fat on both the tumor incidence and tumor multiplicity was clearly present for the low-fiber diets, whereas fat had no effect when the fiber content of the diet was high. In general, the results showed a nonlinear dose-response relationship for fiber and fat. These results indicate that both dietary fiber and fat affect colon carcinogenesis in a complex, interactive manner.