The purpose of this study was to investigate the effect of elaborations on memory . Two types of elaborations (self-generated elaboration and experimenterprovided elaboration) were examined. The experiment consisted of three phases (incidental learning phase, immediate test phase, and delayed test phase). In the incidental learning phase, subjects were asked to make plausibility judgments about 28 fictitious episodes. Half of these were about well-known individuals and the other half were about unknown individuals. Each name (either well-known or unknown) was presented with either two supportive facts or without the supportive facts. During the immediate test phase, subjects were given unexpected memory tests. One week later, unexpected delayed memory tests were administered. Results from both immediate and delayed tests indicated that self-generated elaborations based on prior knowledge subjects had about well-known individuals enhanced the retention of target information, whereas experimenter-provided elaborations involving the presence of supportive facts only benefited memory performance when the subjects had prior knowledge about the individuals. Experimenter-provided elaborations were also effective to the extent that the encoding context was reinstated at testing.