From the experimental point of view, theories derived to account for forgetting phenomena have traditionally relied on interferences from other tasks as a critical factor. The interference is said to stem from the learning of other tasks following the learning of the task to be recalled (retroactive inhibition) and from tasks acquired prior to the learning of the task to be recalled (proactive inhibition). There has been a shift in the relative importance assigned to these two sources of interference. In earlier years, retroactive inhibition was assumed to be the major source of interference in producing forgetting. Recently, however, the proactive inhibition produced by the spontaneous recovery of prior learning has been considered more important. The purpose of this study was to provide some experimental evidence in testing the validity of the spontaneous recovery theory of forgetting. 　　Eighty college students were randomly assigned to 10 groups with 8 subjects of each, and these groups were randomly assigned to two experiments, with 5 groups of each. Each group learned two lists of 15 paired-associate two-Chinese-character abstract terms. The two lists were in A-B, A-C pattern (i.e. identical stimuli, unrelated responses). In experiment I, the learning criterion on each list was one errorless trial. How, ever, five consecutive errorless trials was needed for the first list in experiment H. The retention intervals for 5 groups were: 1 min., 20 min., 1 hour, 24 hours, and 7 days. The MMFR (method of modified free recall) was used for measuring retention. 　　The findings in the experiment I were: (1) the number of trials required to reach the criterion for the second list was significantly greater than that for the first list; and (2) after 24 hours of the conclusion of the learning of the second list, no significant differences between the retentions of the two lists were found. In experiment H, since overlearning was introduced to the first list, the diference between the number of trials for the two lists was not significant. The retentions for the two lists were similar to those in experiment 1. These results partially support the theory.