When we test people’s working memory, we usually ask them to remember some rapidly presented information for a few seconds. This kind of task probably engages not only working memory but also long-term memory: Although long-term memory can hold information for long durations, there is no reason why it should not also contribute to a memory test after a few seconds. However, assuming that all information we ask people to remember in a working-memory test is also maintained in long-term memory raises the question: Why does performance in working-memory task show such a severe capacity limit, given that long-term memory has no known capacity limit? The fact that people cannot remember more than a handful of stimuli in a working-memory test implies that the contribution of long-term memory is fairly limited. In this project we investigate three questions: (1) How does maintaining information in working memory contribute to the formation of a long-term memory trace for that information? (2) Under which conditions can episodic long-term memory contribute to performance in a working-memory task, and how much? (3) How can we distinguish the contributions of working and long-term memory to performance in a task?