Gastric emptying of liquid was studied in 10 normal volunteers and in 27 patients previously treated with truncal vagotomy and drainage. Thirteen of the twenty-seven patients complained of persistent postvagotomy diarrhoea. For each study 300 ml 15 per cent dextrose, labelled with 99mTc-diethylene triamine penta-acetic acid (DTPA), was ingested at a standard rate by subjects who sat facing a gamma camera. Imaging proceeded for 30 min. Gastric area activity curves were corrected for emptying of the test meal during ingestion, and for movement using a new image alignment technique. Gastric emptying at 15 min was 10 +/- 2.6 per cent (mean +/- s.e.m.) in healthy volunteers, 48 +/- 7.3 per cent in patients without diarrhoea, and 84 +/- 2.3 per cent in those with diarrhoea (P less than 0.001, ANOVA). Gastric emptying from 15 min onwards was slower than normal in both patient groups (P less than 0.001). These results show that initial gastric emptying is rapid following truncal vagotomy and drainage, and this change is greater in patients with postvagotomy diarrhoea. No patient with diarrhoea had normal initial gastric emptying.