The glycaemic index (GI) is an in vivo measurement based on the glycaemicresponse to carbohydrate-containing foods, and allows foods to be ranked on the basis of the rate of digestion and absorption of the carbohydrates that they contain. GI values are normalizedto a reference amount of available carbohydrate and do not reflect the amounts of carbohydrate normally present in foods; for example, a food with a low content of carbohydrates will have a high GI value if that carbohydrate is digested and absorbed rapidly in the human small intestine. This is potentially confusing for a person wishing to control his or her blood glucoselevels by the choice of foods. The rate and extent of starch digestion in vitro has been measured using a technique that classifies starch into three major fractions: rapidly digestible starch (RDS), slowly digestible starch (SDS) and resistant starch (RS). In addition, thistechnique gives a value for rapidly available glucose (RAG), which includes RDS, free glucose and the glucose moiety of sucrose. When the values for thirty-nine foods were expressed on the basis ofthe available carbohydrate content of these foods, highly significant (P<0·001) positive correlations were observed between GI and both RDS and RAG. The measurement of RAGin vitro provides values for direct calculation of the amount of glucose likely to be rapidly absorbed in the human small intestine and,thus, to influence blood glucose and insulin levels. These values can be used to compare foods, as eaten,on an equal-weight basis. Food-table RAG values would allow simple calculation of the total amount of RAG provided by single foods, by whole meals and by whole diets. Studies are planned in which RAG and the glycaemic response in man will be measured for identical food products.