Observations of fluctuations in the intensity and temperature of the O I (557.7 nm) airglow taken at Arecibo in 1989 are reported and interpreted on the assumption that they are caused by gravity waves propagating through the emission layer. The data give the magnitude of Krassovsky's ratio as 3.5 ± 2.2, at periods between about 5 and 10 hours. Comparison with theory shows that the gravity waves responsible for the measured airglow variations must have long wavelengths of several thousand kilometers. The observed phases of Krassovsky's ratio are in good agreement with theoretically predicted values at the long wavelengths and large periods for about half the cases. In the other cases, observed phases are near −180°, suggesting that the waves responsible for the airglow fluctuations have experienced strong reflections in the emission layer. The observations emphasize the importance of knowing the full altitude profiles of temperature and winds for extraction of wave information from the airglow fluctuations.