While differences in glucose and insulin responses to specific carbohydrate foods have been reported, few data are available for mixed meals incorporating such foods. This study compared the plasma glucose (PG), serum insulin (SI), and C-peptide (CP) responses tothree different isocaloric test breakfasts given in random order to eight insulin-treated non-insulin-dependent diabetes mellitus (NIDDM) patients. The test meals were selected from a hospital food exchange list and contained similar quantities of carbohydrate, protein, fat, and dietary fiber. The postprandial PG, SI, and CP responses to two of the test breakfasts (meal A: eggs, toasted wholemeal bread, orange juice, margarine, and milk; meal B: wheatflake biscuits, toasted wholemeal bread, milk, and margarine) were similar (meal A: 104.3 ±, 23.0 mg · h · dl−1, 5996 ± 1108 μU · min · ml−1, and 89.8 ± 25.4 pmol · min · ml−1, respectively; meal B: 104.9 ± 21.6 mg · h · dl∼l, 6268 ± 1161 μU · min · mi−1, and 99.8 ± 26.4 pmol · min · ml−1, respectively). Meal C, consisting of toasted muesli and skim milk, produced smaller glycemic and insulin responses (46.8 ± 8.8 mg · h · dl−1; P < .02, and 4369 ± 700 μU · min · ml−1; P < .05, respectively) than meals A and B and less endogenous insulin secretion (CP response 62.8 ± 19.9 pmol · min · ml−1; P < .05 compared with meal A, NS compared with meal B). The lower glycemic response after meal C could be explained by differences in method of food processing resulting in a decreased availability of starch to amylolytic enzymes, the higher content in meal C of sucrose, lactose, and fructose, which are associated with a low glycemic index, and by quantitative and qualitative differences in fiber. While food exchange lists are generally useful in planning diets for diabetic persons, some modification to current lists may be necessary to take into account the processing method and nature of the carboyhydrates in the food when considering the equivalence of individual food items.