Bugelski and Sharlock (1952) credit McGeoch with saying that although the concept of mediation was an old one, it had generated more discussion than experimentation. Bugelski and Sharlock in commenting on McGeoch's statements had this to say, "The concept of mediation is of great potential value for the psychological analysis of learning, thinking, and insight." (Bugelski and Sharlock, 1952, p. 334) The views of Bugelski and Sharlock represent the current thinking in verbal learning circles relevant to the importance of mediation in symbolic forms of behavior. The experimental emphasis, at the present time, is upon the conditions underlying the process. With regard to the nature of mediation Jenkins has stated, The second task, I believe, is to press on in our experimental attack on the conditions of mediation: that is, we should attempt to discover how these implicit processes are acquired, how they are actuated, how they are inhibited, and in general, how they are employed by the subjects. (Jenkins, 1963, p. 212) One method by which the conditions of mediation may be discovered is to determine the relationship between this process and many other forms of intervening variables. In commenting on this latter point Mowrer has stated, "But no one, it seems, has addressed himself systematically to the question of the relation between intervening variables and mediators." (Mowrer, 1960, p. 68) These views of Jenkins and Mowrer on the direction experimentation should take in mediation suggest the need for the current study. The purpose of this study was to investigate the function of visual imagery, type of mediator and associative frequency in mediate association. The role of these factors was examined within an induced mediation paradigm of the form A- B, B-C, A-C.