More and more clinicians are facing young girls suffering from anorexia or bulimia, but the question as to whether these diseases are really increasing remains open. This paper reviews the epidemiological literature published over the last 30-40 years. Until 1980, the authors of retrospective studies claim that the incidence of anorexia has increased. From 1990 on more strictly defined retrospective researches conclude that the incidence and prevalence of anorexia remains steady, although atypical anorectical disorders--often self limited--are increasing. Thus, the incidence of anorexia--strictly defined--is evaluated at 1/100,000 inhabitants within the general population and at 20 to 30/100,000 inhabitants among 15 to 45 year-old females. Moreover, from 1980 on, using new diagnostic tools such as the EAT (Eating Attitude Test), authors describe new nutritional disorders such as bulimia, the incidence of which is calculated at 2 to 5% among 15-45 year-old females. The variability of the results--i.e. the rates--is probably explained by the heterogeneity of the epidemiological approaches, as well as a tendency to denial from the subjects suffering from severe diseases. In the future, one will have to remain cautious in the interpretation of such data: self-administered questionnaires can more easily detect minor eating disorders, and the diagnosis of true anorexia or severe bulimia should be assessed through clinical interviews.