Four experiments are reported which examined the effect of element concreteness on discovery performance by university students during the mental synthesis task. In this task, subjects attempt to combine given elements into a new figure using mental imagery. Discovery performance was not significantly different whether the subjects synthesised novel figures from the names of simple geometric shapes and alphanumerical characters, the names of concrete objects, visual presentation of the objects, or when physically manipulating the objects. A decrement in discovery performance when subjects attempted to synthesise novel figures from the names of elements which have variable shapes suggested that characteristics of the elements can have functional consequences for mental synthesis performance. The scope of the mental synthesis task ranges widely over stimulus materials and subject abilities. The results are discussed in terms of automatic and controlled processes in figure construction and interpretation.