Summary The role of rehearsal in verbal learning and the learning effects of the interval between the initial presentation of an item and its subsequent rehearsal were investigated in a factorial experiment. Three types of rehearsal were compared: stimulus anticipation, response anticipation, and simultaneous presentation. The Ss practiced either exclusively with one rehearsal mode, with a mixture of two, or with all modes. Rehearsal was either immediately after the initial presentation of an item (zerointerval) or a randomly determined number of other items were interposed between initial presentation and rehearsal. The experimental materials consisted of 8 sentences, each intended to establish a relationship of equivalence between two nouns. The sentences were presented within the context of an informative passage about the superstitions of a fictional primitive tribe. Acquisition was self-paced. Recall, as measured on tests administered about 5 min. after completion of practice, was the chief dependent variable. Clerical personnel (N = 544) served as Ss. The data supported the following main conclusions: (1) Zero repetition interval was appreciably less effective than randomly determined interval. (2) Rehearsal through active anticipation with zero repetition interval was the worst practice condition. It depressed recall performance to the level of a no-rehearsal control group. (3) When Ss operate under simultaneous presentation throughout the experiment, this method is as good or better than active anticipation. (4) Active anticipation is a more effective form of rehearsal than simultaneous presentation when the two conditions are compared for Ss who used both rehearsal modes in practice. (5) Response anticipation is more effective than stimulus anticipation when Ss operate under only one rehearsal mode. The superiority of response anticipation is diminished for Ss who practice with both stimulus and response anticipation, but the results are not conclusive in this respect.