Comparisons of gastric emptying between different centres are difficult because of wide variations in methods. Reproducibility of a method is very important before it is used to compare different subjects or to assess the effect of treatment. The aim of this study was to measure the reproducibility of gastric emptying of a solid and liquid meal in normal subjects. Ten males were studied on two occasions. After an overnight fast, the subjects ate a radiolabelled solid and liquid meal. There were no significant differences in T50 on the 2 days (136.6 +/- 23.2 vs 121.3 +/- 26.7 min for solid and 30.7 +/- 12.6 vs 32.6 +/- 18.7 min for liquid; mean +/- SD). Intra-subject variability was between 7 and 21% for the solid component and 1.5 and 63% for the liquid component. The mean difference in T50 between the 2 days was 15.3 +/- 21.9 min for the solid component and -5.1 +/- 19.7 min for the liquid component. Only one difference between the T50 results was not in the 95% confidence interval for the liquid component. Thus despite some inter- and intra-subject variability, the method showed good reproducibility.