IT HAS LONG BEEN BEOOGNIZED that certain pictures appear to change from one form to a second and back again with continued observation. The Necker Cube is a familiar example of such a "reversible" figure. Warren and Gregory (1958) found reversing changes of apparently the same character in the perception of continuously repeated verbal patterns, but Warren (1961a, 1961b) later claimed that these "verbal transformations" showed several features that distinguished them from the visual reversing figures. He considered the verbal transformation effect to be more akin to the continuous distortions observed by Marks (1949) and studied by Honisett and Oldfield (1961). On the other hand, Sakurabayashi (1954) found that the perceived organization of simple line drawings almost always changed abruptly with continued inspection, that the perceived forms differed from in- dividual to individual, that distortions of the stimulus occurred, and that the fluctuations were among many forms. These were the characteristic features of the verbal transformation effect that led Warren to make the distinction between it and the visual reversing figure effect. Axelrod and Thompson (1962) showed that the time course of the rate of change followed the same pattern for the verbal transformation effect as for both the Necker Cube and the rotating spike pattern used by Brown (1955). This observation also suggests that the verbal transforma- tion effect is due to the same mechanism as the reversing figure effect, although Axelrod and Thompson found low correlations among the numbers of changes reported by the same subjects on different tasks. A further possible method of comparing the verbal transformation effect with the visual reversing figures depends on the demonstration (Taylor & Henning, 1963) of a close relationship between the number of different forms and the number of transitions reported in the verbal transformation effect. At any sufficiently late stage in the observation, the number of transitions that had been given was directly proportional to the number of possible different transitions among the reported forms. The same relation seemed to hold for data that Warren published