Extended exposure to study material can markedly improve subsequent recognition memory performance in amnesic patients, even the densely amnesic patient H.M. To understand this phenomenon, the severely amnesic patient E.P., 3 other amnesic patients, and controls studied pictorial material and then were given either a yes-no (Experiment 1) or a 2-alternative, forced-choice (Experiment 2) recognition test. The amnesic patients and controls benefited substantially from extended exposure, but patient E.P. consistently performed at chance. Furthermore, confidence ratings corresponded to recognition accuracy. The results do not support the idea that the benefit of extended study time is due to some kind of familiarity process made available through nondeclarative memory. It is likely that amnesic patients benefit from extended study time to the extent that they have residual capacity for declarative memory.