To test the effects of pictorial presentation and a pictorial review task on science learning, all-verbal or verbal-pictorial presentation conditions were followed by either pictorial practice or a control task. The 96 undergraduate participants were randomly assigned to one of the four conditions presented on a computer-assisted instruction terminal. The groups differed significantly on a pictorial recognition test but not on a verbal posttest. The results favored the pictorial presentation with review condition. Experiment Two tested the effectiveness of the pictorial review task compared to a read-twice control. The same all-verbal presentation condition was given to 50 undergraduates who were then randomly assigned to the pictorial review task or to the re-read condition. The pictorial review condition facilitated performance on the pictorial posttest relative to the re-read condition and produced more favorable attitudes as well.