In order to assess the extent of errors and to provide correction factors for gamma camera gastric emptying studies of liquid meals labelled with radionuclides (99Tcm or 113Inm), a series of phantom studies were performed with different gastric emptying procedures, gamma cameras and data handling systems. To validate the overall accuracy of the method, 24 combined aspiration and gamma camera gastric emptying studies were carried out in three normal volunteers. Gastric meal volume, expressed as a percentage of the original volume of the meal ingested was underestimated due to scattered radiation from the stomach. The underestimation was 7-20% varying with the size of the gastric region of interest (ROI), the energy of the nuclide and the fraction of meal in the stomach. The overestimation, due to scattered radiation from the gut, was negligible (1-3%) for any of the procedures with 10% of the meal situated outside the stomach at a distance of at least 1 cm. Sections of the duodenal loop were always inside the stomach ROI and were likely to contain a constant fraction of the meal (3-10%). The gamma camera technique eliminated much of the error due to variations of stomach geometry and produced accurate quantitative gastric emptying data comparable to those obtained by evacuation (P greater than 0.10), when the entire field maximum 1-min count achieved within the first 20 min of a study was taken as representing the original volume of the meal ingested, and when corrections for area related errors due to scattered radiation from the stomach were performed.