Low glycaemic index foods produce low blood glucose and insulin responses in normal subjects, and improve blood glucose control in Type 1 and well-controlled Type 2 diabetic patients. We studied the effects of a low glycaemic index diet in 15 Type 2 diabetic patients with a mean fasting blood glucose of 9.5 mmol I−1 using a randomized, crossover design. Patients were given pre-weighed diets (59% energy as carbohydrate, 21% fat, and 24g 1000-kcal−1 dietary fibre) for two 2-week periods, with a diet glycaemic index of 60 during one period and 87 during the other. On the low glycaemic index diet, the blood glucose response after a representative breakfast was 29% less than on the high glycaemic index diet (874 ± 108 (± SE) vs ± 204 ± 112 mmol min I−1; p < 0.001), the percentage reduction being almost identical to the 28% difference predicted from the meal glycaemic index values. After the 2-week low glycaemic index diet, fasting serum fructosamine and cholesterol levels were significantly less than after the high glycaemic index diet (3.17 ± 0.12 vs 3.28 ± 0.16 mmol I−1p < 0.05, and 5.5 ± 0.4 vs 5.9 ± 0.5 mmol I−1, p < 0.02, respectively. Urinary C-peptide excretion, as an index of insulin secretion, was 30% lower on the low than the high glycaemic index diet (2.05 ± 0.30 vs 2.93 ± 0.49 nmol mmol-creatinine−1; p < 0.02), urinary urea was reduced by 19% (347 ± 27 vs 402 ± 39 mmol 24-h−1; p < 0.025), consistent with enhanced colonic fermentation. These results suggest that low glycaemic index starchy foods may be beneficial in the treatment of Type 2 diabetes.