The most important factor contributing toward certainty of reproduction is frequent repetition, of course with attention, for without attention no memorizing is possible. The experimental investigation of the influence of repetition has yielded, among minor ones, two particularly interesting results. One of them justifies an educational practice which had already been adopted by teachers because it seemed to be advisable. In order to memorize any material we should not try to force the desired end by accumulated repetition without pause. It is much more economical to devote a short time to learning, long enough for a few repetitions, to do this again after a pause of some hours or days and again after the same interval, until the desired effect is obtained. The total time required for obtaining this effect would be much greater if the total process of memorizing were to occur at one time without intermission. Another result of experimental investigation is contrary to the tradition of educational practice. It has been proved that, in order to learn a long poem, monologue, or piece of prose, this should not be divided into smaller parts. It is uneconomical to learn each stanza or sentence separately. The whole should always be read from the beginning to the end, without introducing points of division which are not desired at the time of reproduction. The method of involuntary reproduction has recently been applied to a problem of much practical significance. The attempt has been made to reveal thus associations of ideas which have been firmly established, but which the subject has strong reasons for keeping secret, for instance, the ideas forming the memory of a crime which he has committed. He is asked to tell or write as quickly as possible a word suggested by each of a great number of words presented to him in succession. Among these latter words are given some which have a special relation to the knowledge which the subject is suspected of possessing. If the suspicion is correct, it is likely to be shown in either of two ways in the answers to these test words. Either the expected (for instance incriminating) answers are actually given and reveal thus the subject’s knowledge; or if these answers are inhibited and voluntarily replaced by others of a more innocent appearance, the time of answering, the reaction time, is considerably increased. It may also happen that the subject, under these conditions, becomes confused and gives absolutely meaningless answers.