BackgroundnnLow concentrations of albumin in serum and long gastric emptying times have been returned to normal in dogs by salt and water restriction, or a high protein intake. We aimed to determine the effect of salt and water balance on recovery of gastrointestinal function after elective colonic resection in human beings.nMethodsnnWe randomly allocated ten patients to receive postoperative intravenous fluids in accordance present hospital practice (greater-or-equal, slanted3 L water and 154 mmol sodium per day) and ten to receive a restricted intake (less-than-or-equals, slant2 L water and 77 mmol sodium per day). All patients had no disease other than colonic cancer. The primary endpoint was solid and liquid-phase gastric emptying time, measured by dual isotope radionuclide scintigraphy on the fourth postoperative day. Secondary endpoints included time to first bowel movement and length of postoperative hospital stay. Analysis was by intention to treat.nFindingsnnMedian solid and liquid phase gastric emptying times (T50) on the fourth postoperative day were significantly longer in the standard group than in the restricted group (175 vs 72·5 min, difference 56 [95% CI 12–132], p=0·028; and vs 73·5 min, 52 [9–95], p=0·017, respectively). Median passage of flatus was 1 day later (4 vs 3 days, 2 [1–2], p=0·001); median passage of stool 2·5 days later (6·5 vs 4 days, 3 [2–4], p=0·001); and median postoperative hospital stay 3 days longer (9 vs 6 days, 3 [1–8], p=0·001) in the standard group than in the restricted group. One patient in the restricted group developed hypokalaemia, whereas seven patients in the standard group had side-effects or complications (p=0·01).