The rapid development of computer graphics technology has made possible an easy integration of dynamic visualizations into computer-based learning environments. This study examines the relative effectiveness of dynamic visualizations, compared either to sequentially or simultaneously presented static visualizations. Moreover, the degree of realism in the visualizations was manipulated experimentally. One hundred-and-twenty university students were randomly assigned to one of six conditions (3 x 2; between-subjects; presentation format x realism). Learners' visuo-spatial abilities were considered as a continuous moderator for the presentation format. Learning outcomes were measured by a pictorial locomotion pattern classification test. Dynamic conditions outperformed static-sequential ones, but not static-simultaneous conditions, in classification performance. Realism had no main effect and did not interact with the presentation format as expected. Learners' visuo-spatial abilities had a positive effect on learning outcomes, but did not moderate the effects of the presentation format. Implications of the results for the design of instructional materials are discussed. (Contains 3 tables and 3 figures.)