This paper reports a follow-up study of 5-, 7-, and 9-year-old subjects who had participated in an investigation of the nature of children's and adults' ability to graphically represent expressive qualities (i.e., happy, sad, angry, loud, quiet, hard). In the original study, the use of literal representation (such as a smiling face on a tree) and non-literal representation (such as a drooping tree for sadness) was scored. The use of non-literal or expressive representation in drawing was found to progress from an initial ability to produce expressive forms based on single cross-modal similarities (such as big for loud) toward the use of more complex multi-dimensional types of expressive forms and the integration of literal and non-literal repesentation in the form of visual metaphors (such as a broken swing for sad). Ninety percent of the sample of the original study was seen at follow-up one year later. Subjects were asked to perform the same drawing tasks. For example, subjects were asked to draw a happy, a sad, and an angry tree so "someone else could tell them apart." In addition, one week later, a comprehension task was given in which subjects were asked to identify the expressive attributes in the drawings they had produced the previous week and previous year. Children were also asked to identify the expressive content of adults' drawings. At this time, a new sample of 4-year-old children performed both the drawing and comprehension tasks. Results are discussed. (Author/RH)