Two experiments examined decay and additivity of semantic priming produced by DRM false memory lists on a naming task. Subjects were presented with study lists containing 14 DRM items that were either all 14 related, the first 7 related, the second 7 related, or all 14 unrelated to the non-presented critical item. Priming was measured on a naming task that presented the critical item in test position 1, 3, or 9. Priming occurred for the 14 related list in the first position only. However, there was also evidence for long term semantic priming in other conditions that was not due to relatedness-checking. Across experiments, an underadditive pattern of activation was obtained. The results are interpreted within the context of Activation Monitoring Theory and suggest that the laws of rapid decay and additivity established for single prime target pairs may not apply to mixed DRM lists or delayed testing. (Contains 1 table.)