We evaluated the copying abilities of ten subjects with Williams syndrome (WS; age 6-14 years) and ten normally developing children (age 3-6 years) matched for mental age using the matrices component of the Kaufman Brief Intelligence Test (mKBIT). Each subject copied six figures, including line drawings of closed and open geometrical shapes (alone and in combination), crossed lines, and geometrical shapes made of distinct small, filled circles. Qualitatively, subjects of both groups made comparable copies, although several subjects with WS drew a continuous line when copying figures composed of distinct circles. Quantitatively, the goodness of the copies was assessed by three human observers who rated on an analog scale the similarity of each copy to its visual template. Ratings were converted to a scale from zero (completely different) to 100 (the same) for statistical analyses. We found the following. First, the overall goodness of copies of the templates was very similar between the WS and control groups (WS: mean=46.7, range=0.89-95.4; control: mean=54.5, range=0.89-98.2). Second, there were systematic differences in the goodness of copies between the two groups, depending on the features of the figures. Specifically, the goodness of copies of control subjects was almost the same as that of WS subjects for simple line figures, but was consistently better for composite line figures, and even better for figures in which the shape was made of small, filled circles. Third, there was a significant relation between the goodness of copies (dependent variable) and mental age (mKBIT, independent variable) in both groups, although it was stronger and more highly statistically significant in the control than the WS group. These findings indicate that the principles guiding copying are similar in the two groups and suggest that WS is a case of developmental rather than deviance disorder.