An assay was developed to evaluate the bioavailability of dietary carbohydrate by slope-ratio analysis of weight gain and plasma ketones of rats fed a carbohydrate-free diet supplemented with glucose as a standard and selected food items, pure carbohydrates and polyols. The diet was based on 35% food grade oleic acid, 12% casein protein supplemented with methionine (0.3%) and arginine (0.2%), 7.5% soybean oil, 37% cellulose, minerals and vitamins. Glucose, as the reference standard, or test materials were incorporated in the diet by replacing equi-energetic amounts of oleic acid and adjusting cellulose to equalize total weight. Optimal levels of protein (casein) and triglyceride (soybean oil) in the basal diet were defined as those levels above which additional protein or fat would serve only to provide endogenous glucose from their glucogenic amino acids and glycerol constituents. A standard dose-response curve was obtained by feeding diets containing 0%, 1% and 2% glucose. Similar dose response curves were obtained by feeding test materials. Under these experimental conditions, additional glucose or additional protein were growth stimulating and casein had approximately 50% of the value of an equal weight of glucose, which was consistent with its content of glucogenic amino acids. The specific carbohydrate value of a food was estimated in the assay by subtracting the calculated glucogenic value of its digestible protein from the total response. The apparent availability of the carbohydrates (i.e., nitrogen-free extract) in soybean meal, lima bean (Phaseolus lumatus), and chick pea (Cicer arietinum) were 35, 70, and 80%, respectively, as available as glucose. Galactose, a 1:1 mixture of galactose + glucose, fractose and starch were as available as glucose. Lactose, ribose, sorbitol, and xylitol were 50 to 65% as available as glucose while mannitol and inulin were not available to a significant extent. Possible metabolic bases for these differences are discussed.