It has long been suspected that the large amounts of reminiscence found by Ballard, Williams, and others were attributable to a flaw in the experimental design used by these investigators. However, it was possible that the differences in the results of these early studies and the later ones by Ward, Hovland, and others could be attributed to other differences in experimental procedure, and, in fact, from the findings of Ballard and Williams, it was possible to predict that only small amounts of reminiscence would be obtained when nonsense material was learned by adult Ss. Accordingly, the present study was designed to attempt to reproduce Ballard''s results under his conditions, and then to investigate the effects of introducing the proper controls. It was possible to reproduce Ballard''s results with considerable precision. However, when the proper controls were introduced, the "reminiscence" effect vanished. There is a strong presumption, therefore, that the results obtained by Ballard (1), Huguenin (2), Williams (6), and G. O. McGeoch (3) are spurious and that reminiscence obtained under this type of design is an artifact of the experimental method employed.