Recent developments of computer hardware have brought opportunities for programing more complicated teaching materials in CAI. New AV equipment makes it possible to give instruction using sounds and pictures with computers. Use of visual materials is popular in foreign language teaching for the purpose of setting foreign-like situations, inhibiting the use of native language, retaining long-lasting memory, and making the foreign culture interesting. These effects of visual materials, however, are hardly verified scientifically although they have played an important role in the classroom since early days. In order to improve CAI courseware, it is necessary to study the effects of visual materials used in the foreign language classroom. This study attempts to clarify whether pictures have a greater effect on learning than words only, to compare the effects using two kinds of pictures (drawings and photograghs), and to examine the interaction between learners' traits and three different types of stimuli (words only, drawings, photographs). Subjects were 127 male high school students. Teaching content was 21 English prepositions, selected by a test beforehand and ranked from one to five on their level of difficulty to learn. Drawings and photographs corresponding to these prepositions were prepared as teaching materials and made into programed booklets. In these booklets the prepositions were presented in each of the following modes : in a single sentence with or without pictures ; with only a drawing ; or with only a photogragh. Each frame required subjects to fill in a blank, choose a right answer, or compose a single English sentence using a preposition. Items were counterbalanced across presentation modes. A posttest was administered five days later. The results showed that records during the step by step learning process indicate significant effectiveness of pictures and drawings. However, the scores of the posttest proved that there were no significant differences in learning effects among the three different modes. Nor was any interaction found between learners' traits and the different modes of presentation. The results of this study suggest two implications. First, if a learner already possesses one concept, it makes no difference whether or not pictures are used in teaching that concept in a foreign language. The native language may be more precise and efficient than pictures in defining the meaning. Second, photographs may fail to represent the intended information because of too many distracting details. Effectiveness of pictures during the learning process can be interpreted as due to the novelty effect of pictures. Further study is needed to scrutinize the effects and use of pictures in the foreign language classroom.