We have studied the effects of glucose, sucrose, and various starches on postprandial plasma glucose and insulin responses in 19 subjects. All carbohydrate loads were calculated to contain 50 gm. of glucose, and the response to each carbohydrate was tested twice: when given alone in a drink or when given in combination with other nutrients as a meal. The data demonstrate: (1) Glucose and sucrose elicited similar plasma glucose response curves, but sucrose elicited a somewhat greater (20 per cent) plasma insulin response. (2) Raw starch ingestion resulted in a 44 per cent lower glucose response and a 35-65 per cent lower insulin response than did either glucose or sucrose ingestion. (3) When carbohydrate was given as a meal the plasma glucose responses were 40-60 per cent lower than when the same carbohydrate was given as a drink, while the insulin responses were generally similar, and (4) when different cooked starches were compared, the plasma glucose and insulin responses to rice were significantly lower (50 per cent) than to potato. In conclusion, the size of the carbohydrate molecule appears to influence the postprandial glucose and insulin responses such that more complex carbohydrates (starches) elicit lower responses. This effect may be related to differences in digestion rather than to differences in absorption.