Gastric partitioning is undertaken for morbid obesity but the mechanism by which weight loss is achieved is unclear. The effect of partitioning on gastric emptying has therefore been studied. In thirty patients with a mean weight of 124.6 +/- 7.3 kg (+/- s.e.m.), a stapling technique was used to create a 50 cm3 upper gastric reservoir which communicated via a 10 mm stoma with the distal stomach. Twenty-four pre-operative and nineteen postoperative gastric emptying studies were performed using separate 50 cm3 solid and liquid 113m In-labelled meals. The time taken to achieve 50 per cent gastric emptying and the delay in isotope reaching the duodenum were measured from gamma camera images. There was no correlation between the pre-operative weight of the patients and their gastric emptying rates. The median 50 per cent gastric emptying time for solids was increased from 21 min (range 8-34) to 37.5 min (range 8-60) (P less than 0.001), unpaired Wilcoxon), but was little changed from 17 min (1.5-30) to 16 (1.5-30) min using liquids. Delay in arrival at the duodenum was increased for solids from 0.5 min (0.5-13) to 2.5 min (0.5-42.5) (P less than 0.05) but again there was no significant effect on liquid emptying. Postoperative 50 per cent emptying time for solids correlated with mean weight loss at 3, 6, 9 and 12 months (P less than 0.05). Following gastric partitioning there is delayed emptying of solids with related weight loss but liquid emptying is unaffected.