The rate with which perceptual information becomes available was manipulated in 2 word naming experiments. Word priming effects, in terms of reduced naming latencies for repeated items, and recognition memory measures were obtained with matched groups of amnesic patients and control participants. In both experiments, the amnesic patients evidenced significantly reduced priming effects compared to control participants under difficult task conditions. Under easy task conditions the baseline naming latencies of the amnesics were significantly longer than those of controls, but the difference in priming effects failed to reach significance. The findings are consistent with the Information Availability model of priming positing that both priming and explicit memory are mediated by episodic information from a study or information processing episode. It is argued that word priming does not represent a memory function that is spared in amnesia.