This study investigated how readily achievable changes to the quantity and processing of starchy foods in a typical Western diet: (1) were reflected in levels of resistant starch (RS) and NSP excreted from the small intestine; and (2) more favourable profiles of butyrate, NH3 and phenol production. Two diets, a low-starch diet (LSD) and a high-starch, low-fat diet (HSLFD) were compared. The LSD with 20 % total energy (%E) from starch was based on a ‘typical’ Australian diet, while the HSLFD (40 %E as starch) was the same Australian diet modified by an increased content of legumes, starchy foods and coarsely-ground cereals and by a reduced fat content. Four subjects with iliostomies consumed each diet for 2 d, with ileal effluent collection on the second day. On the HSLFD compared with the LSD, RS in ileal effluent increased from from 0·49 to 1·7 g/MJ per d (P < 0·005) while ileal NSP excretion increased from 2·0 to 3·3 g/MJ per d (P < 0·05). Ileal effluents obtained after each diet were incubated for 24 h in vitro with a human faecal innoculum. After fermentation, ileal effluent from the HSLFD produced more butyrate relative to other short-chain fatty acids (17·5 v. 15·8 molar %, P < 0·005) and less phenol (2·3 v. 5·7 mg/l, P < 0·05) and NH3 (20·3 v. 23·1 mmol/l, P < 0·005) than the LSD diet. The HSLFD also generated a lower pH (6·15 v. 6·27, P < 0·05). On a wt/wt basis, RS was 2·3-fold higher in the HSLFD effluent while NSP did not increase, suggesting that the change in RS largely contributed to the fermentation effects. Changes in in vitro variables when the HSLFD ileal effluent was ground before fermentation indicated the importance of physical structure in determining ileal excretion of RS. We conclude that: (1) readily achievable modifications to the amount and processing of starchy foods in an Australian diet would produce potential benefits for in vitro fermentation variables; and (2) the physical structure of grains and cereals is important in determining access by colonic bacteria to a carbohydrate substrate.