Five groups of 5 patients with maturity-onset diabetes were given an oral glucose tolerance test with 75 g glucose, then 2 days later were given 75 g carbohydrate as rajmah (Phaesolus vulgaris), channa (Cicer arietinum), green gram (Vigna radiata), rice or wheat. There was a highly significant decrease in blood glucose concentrations when carbohydrate was given in the form of legumes (rajmah, green gram and channa) than when it was given as cereals or glucose. The decrease in blood glucose concentrations was directly related to the dietary fibre content of the food given; it was greatest for rajmah and least with rice.