Body image disturbances in anorexia nervosa patients have been assessed in an experimental study which dealt with the following issues: (1) Three different procedures for the assessment of body image disturbances were compared in a discriminant function analysis: (a) a Video Monitor Procedure, (b) the Movable Caliper Procedure and (c) the Image Marking Procedure. The Image Marking Procedure showed the best discrimination between groups. (2) Anorexia nervosa patients showed a statistically significant over-estimation as compared to the control group in the variables ‘waist’, ‘upper thigh’ and the compound measure ‘soft body parts', while they did not overestimate control variables (Kruskal-Wallis-Rank-Analysis of Variance). (3) A standard liquid meal of 240 Kal. had no statistically significant influence on the estimation of body width, irrespective of the amount of calories consumed and instruction given. Possibilities and limitations of various measurement procedures are discussed.