Three experiments are reported which investigated the effects of hyperbaric air on long-term memory. In the first, word lists were learned at 1 and 8.6 ATA using a variable input-free recall paradigm. It was found that learning was affected but not clustered memory organization, and it was concluded that disorganization of memory is not a factor contributing to the learning deficit found with hyperbaric air. In the second and third experiments it was found that the recall of words, which had been learned when non-narcotic, was disrupted at 10 ATA by hyperbaric air and that this disruption was not overcome by providing memory-cues at the time of recall. Two possible explanations for these results are discussed. A hypothesis is put forward to reconcile the results of various hyperbaric memory experiments by pointing out that a relationship between stress and learning found with nitrous oxide could be applicable to these studies also.