Previous work has documented cognitive deficits at high altitudes (15,000-25,000 ft), but there is controversy for lower altitudes. This study looked at the effects of moderate altitudes--12,500 ft and 15,000 ft--on short-term memory in comparison to 2,000 ft. Seventy-two student pilots and instructors were first administered the Vocabulary, Digit Span, and Digit Symbol subtests from the Wechsler Adult Intelligence Scale-Revised, the Vandenberg Mental Rotation Test, and the near-contrast sensitivity portion of the Vistech VCTS 6000 chart. Participants then spent 1 1/2 hr at their designated altitude for cognitive testing. Participants performed a 30 min vigilance task while listening to an audiotape with instructions to recall radio calls prefaced by their assigned call sign. Half of the radio calls were high memory loads (at least 4 pieces of information), and half were low memory loads (no more than 2 pieces of information). No effects of altitude were found in performance on the Vigilance task. However, for readbacks of high memory load, significant deficits in recall were observed at 12,500 ft and 15,000 ft, whereas no effect of altitude was observed on recall of readbacks with low memory loads. These results indicate that, at altitude, short-term memory was exceeded for the readbacks requiring a larger amount of information to be recalled, and that cognitive deficits are found at lower altitudes than previously observed.