2 8 groups of 20 university undergraduates each, learned 1 of 4 transfer paradigms to a List I criterion of either at least 1/10 or at least 5/10, and a List II criterion of 1 perfect trial. Paradigms used were A-B, C-A ; A-B, A-Br; A-B, A-D; and the A-B, C-D control. In all cases List II stimuli were midrange CVCs and List II responses were highfrequency-of-occurrence real words. This was also true for List I with the exception of the A-B, C-A paradigm where the stimuli of List I were the same as the responses of List II, and the responses of List I were a new set of midrange CVCs. List II trials to criterion data indicated essentially the same pattern of results at both levels of List I learning. A-B, A-B r produced the greatest negative transfer and A-B, A-D the least. A-B, C-A was not significantly different from the A-B, C-D control group, although reaching errorless performance in fewer trials. 1 This article is based on a dissertation submitted to the graduate school of the University of Rochester in partial fulfillment of the requirements for the PhD degree. The author wishes to acknowledge the considerable assistance of B. G. Andreas and R. C. Radtke. An abridged version of this paper was presented to the Spring 1966 meeting of the Eastern Psychological Association in New York City. 2 Now at HumRRO, Division No. 2, Fort Knox, Kentucky.