Groups of housewives (n=8) were allocated to the cells of a 2 x 2 factorial design that contrasted the effects of physical reinstatement (testing in the same or different room) with mental reinstatement of context (spontaneous recall or guided rehearsal of the original incident) on memory for the appearance of an unexpected intruder. Irrespective of room used for testing, memory guidance increased the quality of facial composites produced by witnesses. Physical reinstatement of context also improved composite quality, though to a lesser extent. No significant effects for either guidance or context were found on either recall of the physical details of the intruder or a simulated mugbook identification task. (Publisher abstract)