Data reported by Butter (1970, Experiment II) indicate that effects of stimulus concreteness-imagery on incidental recall may reverse over a long retention interval. Consistent with previous results, Butter also found a reversal of stimulus-arousal effects over time. In the present attempt to replicate Butter 's findings , no reversal was found for either the concreteness-imagery or arousal variable. A strong concreteness-imagery main effect was found , however, with high stimulus concreteness-imagery facilitating recall on both immediate and delayed tests.A number of studies, beginning with that of Kleinsmith & Kaplan (1963) , have shown that incidental recall of paired associates is affected by arousal value of the stimulus terms. Responses paired with low-arousal stimuli are recalled better than those paired with high-arousal stimuli on an immediate retention test, but the effect reverses after a longer retention interval. Results of a study by Butter (1970, Experiment II) suggest that a similar reversal occurs with respect to the effects of stimulus concreteness (C) and imagery value (I).The specific procedures used by Butter were to present a list of eight paired associates to Ss, with half of the pairs havinghigh and half having low C-I stimuli. For Ss tested 2 min after list presentation, recall of responses paired with high C-I stimuli was better than recall of responses paired with low C-I stimuli. For Ss tested 2 days after list presentation, however, the effect was reversed . Since Butter had obtained GSR data from her Ss during stimulus presentation, she was also able to determine the effects of arousal on retention . Analysis of the retention data, with respect to arousal effects, showed the pattern of recall to be in agreement with that of Kleinsmith & Kaplan (1963) and just the reverse of that associated with the C-I variable. Consistent with the retention results, Butter reported additional data indicating that low C-I items tended to be classified as high arousal, while high C·I items were low in arousal.Yuille (1971) attempted to replicate Butter's findings with respect to the reversal of C-I effects, but was unable to do so. Yuille suggested two possible reasons for this failure to replicate. One possibility is that Butter's monitoring of GSRs, a procedure not followed by Yuille, may have modified the learning situation in some way. The second possibility suggested by Yuille is that Butter had made an error in her analysis of C-I effects. On the basis of recall percentages reported by Butter, 'This research was supported in part by Grant HD06420 from the National Institute of Child Health and Human Development to the second author.Bull. Psychon. Soc., 1973, Vol. 2 (1) Yuille concluded that only 12 Ss per group were included in the analysis rather than the 16 Ss per group Butter had reported testing. While Yuille's observation that Butter had omitted data from analysis appears to be correct, the suggestion that Ss had been eliminated is contraindicated by the values Butter reports as degree...