To examine whether the form of dietary carbohydrate influences glucose and insulin responses, we studied the glucose and insulin responses to five meals--each containing a different form of carbohydrate but all with nearly identical amounts of total carbohydrate, protein, and fat--in 10 healthy subjects, 12 patients with Type I diabetes, and 10 patients with Type II diabetes. The test carbohydrates were glucose, fructose, sucrose, potato starch, and wheat starch. In all three groups, the meal containing sucrose as the test carbohydrate did not produce significantly greater peak increments in the plasma concentration of glucose or greater increments in the area under the plasma glucose-response curves than did meals containing potato, wheat, or glucose as test carbohydrates. Urinary excretion of glucose in patients with diabetes was not significantly greater after the sucrose meal. The meal containing fructose as the test carbohydrate produced the smallest increments in plasma glucose levels, but the differences were not always statistically significant. In healthy subjects and patients with Type II diabetes, peak serum concentrations of insulin were not significantly different in response to the five test carbohydrates. Our data do not support the view that dietary sucrose, when consumed as part of a meal, aggravates postprandial hyperglycemia.