Recently, we conducted a study1 in which the responses of lysergic acid diethylamide (LSD) subjects on a word association test (WAT) were compared with those of a control group, as well as with the responses of groups of neurotics, psychotics, and controls tested by Rapaport, Gill, and Schafer.2 In the course of this investigation, a marked "practice" effect was observed when the WAT was administered to subjects a second time. Since the subjects were unfamiliar with the WAT and its scoring, they had no way of consciously knowing what constituted an adequate or deviant response. It was, therefore, hypothesized that any improvement on a retest might be due largely to an unconscious mobilization of psychological defenses, suggesting that the degree to which a subject is able to "correct" deviant responses may be a measure of the integrity of his ego defenses. Accordingly, a separate study of this intriguing practice effect