Visual search, that is the search for a specific target object amongst distractor objects, is an essential human behaviour. A question which arose about a decade ago concerns the involvement of memory during visual search. In an influential paper, Horowitz and Wolfe (1998) proposed that memory does not support visual search. However, the conclusion of memory-less search was highly debated ever since (e. g., Gilchrist & Harvey, 2000; Horowitz, 2006; Horowitz & Wolfe, 2001; Körner & Gilchrist, 2008; Kunar, Flusberg, & Wolfe, 2008; Peterson, Kramer, Wang, Irwin, & McCarley, 2001). Although most researchers now agree that memory supports visual search the properties of this memory are still unclear (e.g., Beck, Peterson, & Vomela, 2006). For instance, Körner and Gilchrist (2007) suggested that short-term memory is involved in repeated visual search. It is possible to remember recently inspected objects from a previous search and find them faster in a subsequent search. One aim of the experiments presented here was to answer the question whether items are not only found faster in a subsequent search when they have been inspected during a previous search but also when they have not been inspected before. Furthermore, the involved memory processes were investigated. To this end, five experiments were conducted. Results showed that there is not only a benefit for recently inspected items but also for non-inspected items. When the underlying memory processes were investigated, results showed that previously inspected items were inhibited within search but not across the two searches. Such an inhibition was expected to be found when the benefit for non-inspected items is rather due an automatic memory process. Thus, it was suggested that the same memory mechanism which causes the benefit for recently inspected items in repeated visual search also causes the benefit for non-inspected items. Such a memory process is assumed to guide search rather actively and flexibly.