The plasma glucose and serum insulin responses were determined in untreated Type 2 (non-insulin-dependent) diabetic patients following the ingestion of foods containing sucrose, glucose, fructose or lactose in portions that contained 50 g of carbohydrate. The results were compared to those obtained following the ingestion of pure fructose, sucrose, glucose + fructose and lactose. The objectives were to determine 1) if the glucose response to naturally occurring foods could be explained by the known carbohydrate content, and 2) whether the insulin response could be explained by the glucose response. The glucose response was essentially the same whether the carbohydrate was given as a pure substance, or in the form of a naturally occurring food. The glucose response to each type of carbohydrate was that expected from the known metabolism of the constituent monosaccharides. The glucose areas following the ingestion of the foods were: Study 1: glucose 11.7, orange juice 7.3, sucrose 5.2, glucose + fructose 6.3, and fructose 0.7 mmol X h/l; Study 2: glucose 14.6, orange juice 7.3, apples 5.5, and apple juice 4.7 mmol X h/l; Study 3: glucose 12.6, ice cream 8.1, milk 3.7, and lactose 4.1 mmol X h/l. The insulin response was greater than could be explained by the glucose response for all meals except apples. Milk was a particularly potent insulin secretagogue; the observed insulin response was approximately 5-fold greater than would be anticipated from the glucose response. In summary, the plasma glucose response to ingestion of fruits and milk products can be predicted from the constituent carbohydrate present. The serum insulin response cannot.