Previous studies have reported contradictory findings regarding the effects of item repetition on the subsequent encoding of contextual details associated with items (i.e., source memory). Whereas some studies reported repetition-induced enhancement in source memory, other studies observed repetition-induced impairment. To resolve these conflicting results, we examined the modulatory role of preexperimental stimulus familiarity in the relationship between item repetition and new source memory formation by orthogonally manipulating preexperimental stimulus familiarity and intraexperimental item repetition. In a series of experiments consisting of three phases (item repetition, item-source association, and source memory test), we found that item repetition impaired source memory for preexperimentally familiar items (famous faces or words), whereas the same manipulation improved source memory for preexperimentally novel items (nonfamous faces or pseudowords). Crucially, item repetition impaired, rather than improved, source memory for preexperimentally novel items when these items had been preexposed to participants before the three-phase procedure. Collectively, these findings provide strong evidence that preexperimental stimulus familiarity determines the relative costs and benefits of intraexperimental item repetition on the encoding of new item-source associations. By demonstrating the interaction between different types of stimulus familiarity, the present findings advance our understanding of how prior experience affects the formation of new episodic memories.