This study aimed to evaluate whether serial ultrasound measurements of antral volumes are reproducible and, if so, whether they are representative of the entire gastric contents and reliably measure gastric emptying. The antral volume was measured in man after oral and intragastric administration of known amounts of 5% glucose solutions, and on two separate occasions in the same subjects, after the ingestion of a standardised solid-liquid meal (1050 kcal). Antral volume measurements were performed in both supine and upright positions; inter- and intraobserver measurement errors were also assessed. After ingestion or intragastric administration of equal amounts of liquid, antral volumes determined by ultrasound showed a wide intersubject variability. In each individual subject, however, a linear trend was found between the antral volume and amount of ingested or administered liquid. Intrasubject variability of antral volume measurements was at its minimum in fasting conditions and 300 minutes after ingestion of the solid-liquid meal. Intra- and interobserver variations and whether the patient was erect or supine did not affect measurements of antral volumes. These results support the use of real time ultrasound in determining the gastric emptying time. Results at fasting and 300 minutes after a solid-liquid meal are stable end points of measurement.