30 patients with active Crohn's disease, mean Crohn's Disease Activity Index 301 (SE 32), who would otherwise have been treated with steroids, were randomised to receive for 4 weeks either an elemental diet ('Vivonex') (n = 16) or a polymeric diet ('Fortison') (n = 14). Assessment on days 10 and 28 showed that clinical remission occurred in 5 (36%) of the 14 patients on fortison compared with 12 (75%) of the 16 patients assigned to vivonex. The difference in remission rate was significant (p less than 0.03). Dietary treatment resulted in little change in the nutritional state and various laboratory indices of activity over a 4 week period despite clinical improvement. Polymeric diets do not seem to offer an effective therapeutic alternative to elemental diets in patients with acute exacerbations of Crohn's disease.