Plasma glucose and insulin responses to six different meals were determined and compared with values predicted by published glycemic indices of the component foods. The test meals were of different ethnic origins: Indian (lentil curry with rice), Italian (spaghetti bolognaise), Chinese (stir-fried vegetables and chicken with rice), Greek (lentil stew), Western (sirloin chop and vegetables); and Lebanese (sandwich with unleavened bread and hummos). Eight healthy volunteers were given 50 g carbohydrate portions of the above meals after an overnight fast. The glycemic and insulin indices were highest for the Lebanese meal and lowest for the Greek with significant differences among the meals (ANOVA, p less than 0.05). The observed glycemic indices correlated well with the predicted glycemic indices (r = 0.88, p less than 0.01) and insulin responses parallelled the glycemic responses (r = 0.83, p less than 0.05). These results suggest that the glycemic index approach will be useful in planning diets for diabetic people.