Although anorexia nervosa may vary widely in its severity and outcome, it is usually regarded as an illness in its own right, as patients do not display major changes in the form of the illness over time (Russell, 1970). It readily lends itself to being accorded clear-cut diagnostic criteria (Russell, 1977) and this is reflected in ICD–10 (World Health Organization, 1987). However, the nosological independence of anorexia nervosa has undergone vigorous assault since its classic description by William Gull (1874). For example, the psychological nature of anorexia nervosa was obscured for 30 years by Simmond's (1914) description of anterior pituitary lesions and cachexia, and Kay & Leigh's (1954) influential study of anorexia nervosa concluded with their doubts about its status as a ‘psychiatric entity’.