Enteral feeding has been shown to be as effective as primary therapy for Crohn's disease, but it requires high patient motivation, may be unpalatable and is expensive. However, in adolescents with growth failure and when corticosteroid therapy is contra-indicated or has failed, it may become the treatment of choice. Furthermore, dietary therapy allows circumvention of the adverse side-effects of repeated courses of steroids. A number of different hypotheses have been proposed to explain the effect of enteral feeds but none has reached universal acceptance. Prospective trials suggest that the exclusion of whole protein is not necessary. Comparison of feeds with differing composition suggests that a low fat content increases efficacy and various explanations have been offered. The reduction of colonic bacterial load may also be important. Because symptoms of Crohn's disease may be provoked by eating, there is a risk of falsely attributing symptoms to specific foodstuffs. However, in many individuals foods can be identified which affect disease activity, and their exclusion leads to prolongation of disease remission. Dietetic supervision during food testing is important to avoid detrimental effects on nutrient and micronutrient intake.