The purpose of the investigation was to determine the meanings attached by students to the different kinds of molecular structure representations used in chemistry teaching. The students (n = 124) were from primary (aged 13-14 years) and secondary (aged 17-18 years) schools and a university (aged 21-25 years). A computerised "Chemical Visualisation Test" was developed and applied. The research indicates that students' appreciation of three-dimensional molecular structures differs according to the kind of representation used. The best results were achieved with the use of concrete, and pseudo-concrete types of representations (e.g. three-dimensional models, their photographs, computer-generated models). However, the use of more abstract types (e.g. schematic representations, stereochemical formula) was less effective. A correlation between students' results on the Chemical Visualisation Test and their educational level, spatial visualisation, and spatial relations skills was shown statistically, but no statistically significant gender differences were observed. (Contains 2 tables, 8 figures and 2 notes.)