Few studies have compared the gastric emptying of digestible and non-digestible solids in man. In 9 normal subjects and 16 duodenal ulcer patients, we measured simultaneously and during 3 h, the gastric emptying of: (a) radioopaque pellets (n = 100; volume = 30 mm3) ingested with an ordinary meal (450 kcal, 500 ml) and counted on X-ray series; and (b) digestible solids of the meal (99mtechnetium tagged egg white) detected by an isotopic method. Gastric emptying of liquids (labelled with 113mindium DTPA) was also determined. In normal subjects, emptying rates of the pellets was 5 +/- 2 p. 100 per hour while that of digestible solids was 29 +/- 1 p. 100 per hour (P less than 0.01). No correlation was evidenced between the emptying rates of either kind of solids. Emptying of liquids was faster than that of both digestible solids and pellets. Similar results were obtained in duodenal ulcer patients; no significant difference was found between patients and normal subjects for the three meal phases tested. These results are consistent with the hypothesis that digestible solids are retained in stomach until they are ground, and show that inert particles do not always reflect the gastric emptying of natural solids of the meal.