Records obtained with the Sheffield meteor wind radar have been subjected to harmonic analysis resolving a longest period of 130 h. The spectra of oscillation generally show three distinct peaks: periods near 12, 24, and, surprisingly, between some 35 and 72 h. The 24 and 12 h peaks can be readily accounted for as the effect of the solar diurnal and semidiurnal tides but regular oscillations with periods in excess of one day appear more difficult to explain. Although upward propagation of long period waves is severely restricted there seems to be some evidence that the primary source of these oscillations lies in the lower atmosphere as shown by the correlation between meteor winds and the variation of atmospheric pressure as well as tropospheric winds near the principal meteor collecting areas.