Sixty Croatian children who had experienced directly the recent war in Croatia drew a man, followed on a separate page by either (i) a Croatian soldier, (ii) an enemy soldier or (ii) a second drawing of a man. Half of the sample had lost their father due to the war. There were no significant effects of Trauma group (with/ without father) or Topic of drawing on the drawings' size or their placement relative to the child's self-portrait drawing. The paper discusses the many mechanisms of influence of emotion on drawing but argues that these are idiosyncratic and unpredictable. The use of drawings in clinical assessments of adults and children has had a long tradition. Indeed, diagnostic drawing tests still rank in the top ten of all tests used by clinicians (Lubin, Larsen, Matarazzo & Sever, 1985; Watkins, Campbell, Nieberding & Hallmark, 1995). Furthermore, there are numerous case studies using drawings as a therapeutic tool (e.g., see Dalley, 1984; Winner, 1982). Both forms of practice indicate that it is commonly considered by those working with disturbed clients that drawings can provide a useful insight into the client's problems. Although there is little dispute that even line drawings can be expressive using aesthetic criteria (Arnheim, 1949; Gardner, 1974; Werner & Kaplan, 1963), the evidence for drawings conveying an emotional maladjustment in the drawer is far less convincing. Theoretical links between drawings and the drawer's maladjustment can be broadly categorized under three traditions (see Cox, 1992; Thomas & Silk, 1990; Thomas & Jolley, in press). First, Machover (1949) claimed that drawings can be used as a personality assessment. She devised the "Draw-a-Person" test in which the client is asked to draw a person and then make a second drawing of a person of the opposite gender to that depicted in the first drawing. Interpretations of the drawings are then made by the clinician using psychoanalytic theory. In extensive reviews of the literature, Swenson (1968) and Roback (1968) argue that the Draw-a-Person Test has weak reliability and validity, and that particular aspects of a drawing cannot be related to particular types of maladjusted personality (see also Falk, 1981). This is likely to be due to the tradition's reliance on the body-image assumption. That is, a drawing of an unidentified person conveys the artist's self-concept and possibly their physical image.