SUMMARY1. Results were collected from thirty-three published and unpublished studies of gastric emptying. The volumes of the meals ranged from 50 to 1250 ml., and composition varied from pure carbohydrates to ordinary food.2. From the published composition of the meals, their nutritive density, as kcal/ml. (4.18 KJ/ml.) was computed: it ranged from zero to 2-3 kcal/ml.3. The volume of each meal, or test meal, delivered to the duodenum in 30 min was determined, assuming that gastric emptying was exponential.4. The greater the nutritive density of a meal, the less was the volume transferred to the duodenum in 30 min. The original volume of meal given was not a determinant of the rate of emptying (ml./min).5. The slowing of gastric emptying with a meal of high nutritive density was not sufficient to prevent an increased rate of delivery of energy to the duodenum (nutritive density x volume delivered in unit time) with a meal of high nutritive density.6. Assuming an appropriate relationship for the interaction of a stimulus (kcal/ml.) and duodenal receptors, it was possible to predict a rate of gastric emptying for each meal, given its nutritive density. Knowing the initial volume of the meal, it was possible to predict the mean half time for its emptying. 7. There were eight sets of anomalous results: in four the volumes of meal given were less than 200 ml.; explanations of the anomalies in the other four results could not be provided.8. The results are consistent with equal slowing of gastric emptying by the duodenal action of the products of digestion of isocaloric amounts of fat, protein and carbohydrate, for example, 4 g fat or 9 g carbohydrate, both 36 keal, taking carbohydrate and protein as 4 kcal/g and fat as 9 kcal/g.