Between 1 September 1980 and 31 August 1983, 352 patients with inactive or mildly active Crohn's disease but not taking drug treatment apart from sulphasalazine were entered from 40 hospitals into a prospective trial to assess the effects of two different diets on disease activity over two years. One hundred and sixty two patients were randomly allocated to take a diet unrestricted in sugar and low in fibre and 190 to a diet with little or no sugar and high in unrefined carbohydrate. No clear difference in clinical course was detected among patients who accepted the two different types of dietary advice.