Background: The role of fermentation compared with the source or type of the fermentable material in colon tumorigenesis remains an issue in refining the definition of dietary fiber (DF).Objective: The aim of this study was to investigate the fermentation and source-specific effects of various carbohydrates in a medium-term colon tumorigenesis model.Methods: Six-week-old male Fischer 344 rats were randomly allocated into 6 groups (n = 36/group) to receive either AIN-93G (control) or diets containing fructooligosaccharides, wheat bran (WB), oat bran (OB), polydextrose, or high-amylose maize starch (HAMS), each adjusted to contain a total DF concentration of 7% (wt:wt) and have a fermentability of 3% (wt:wt). After 2 wk, 24 rats/group received 2 subcutaneous doses of azoxymethane (at 15 mg/kg body weight) 1 wk apart while 12 rats/group were injected with a saline vehicle; all rats were maintained on the assigned diets for 24 wk postinjection and then killed. Colon tumor outcomes and pathology together with cecal short-chain fatty acid composition were assessed.Results: No tumors were found in saline-injected rats, and all subsequent analyses were restricted to azoxymethane-injected rats. Colon tumor incidence was significantly lower in the polydextrose (21%) and WB (13%) groups than in the control group (63%; P < 0.05) but not different from the fructooligosaccharide (58%), HAMS (46%), and OB (33%) groups. In comparison to the control group (8 proximal/31 total tumors), fermentable materials reduced the number of tumors (P < 0.05) originating in the proximal colon: HAMS (5/15), polydextrose (2/7), OB (2/9), fructooligosaccharides (1/21), and WB (1/3). The mean ± SEM number of tumors/tumor-bearing rats was significantly lower in the WB (1.00 ± 0.00), OB (1.13 ± 0.13), and HAMS (1.36 ± 0.15) groups than in the control group (2.07 ± 0.27; P < 0.02); other groups did not differ. The mean ± SEM tumor burden/diet group was lower in the WB (1.2 ± 0.7 mm2), polydextrose (6.7 ± 3.2 mm2), and OB (7.0 ± 3.0 mm2) groups than in the control (21.4 ± 5.9 mm2) and fructooligosaccharide (22.1 ± 7.1 mm2; P < 0.05) groups but not significantly different from the HAMS group (15.1 ± 6.1 mm2). Total cecal SCFA concentrations did not differ among diet groups (overall mean ± SEM: 81 ± 4 μmol/g wet weight).Conclusion: The rate and extent of fermentation of the carbohydrate material as well as the characteristics of the material in the lumen of the lower gastrointestinal tract all appear to have an important role in tumor outcomes in the azoxymethane-induced rat colon tumorigenesis assay.