One version of deficient-processing theory attributes the lag effects observed in free recall to a diversion of attention or processing activity from repeated items when the spacing between repetitions is small (cf. Hintzman, 1976). This hypothesis indicates that spacing effects should be sensitive to many instructional variables that alter the rules for allocating processing capacity to the to-be-remembered items. In an effort to manipulate such allocation rules, two free-recall experiments were carried out. The first manipulated overt rehearsal instructions so that subjects were either left free to rehearse in a manner of their own choosing or were restricted to the item currently being presented. The second was similar except that overt rehearsal was not required and a third condition (imageability-rating task) was added. The results indicate that spacing effects are sensitive to instructional variables and suggest that displaced rehearsal plays an important role in the spacing effect.