Observations of winds in the upper mesosphere and lower thermosphere (~80--110 km) measured at the geographically conjugate locations of Adelaide (35 ¿S, 138 ¿E) and Kyoto (35 ¿N, 136 ¿E) in the years 1983--1985 are used to study the hemispheric differences in the behavior of the solar diurnal and semidiurnal tides. Comparisons of the tidal amplitude and phase parameters show that there are distinct hemispheric asymmetries. The diurnal tide is usually much larger in amplitude at Adelaide, and it exhibits a height structure which indicates a relatively larger contribution from the propagating (1, 1) mode at this location. Comparisons with recent theoretical work suggest that while some of these differences may be due to hemispheric differences in forcing and/or to interactions between the tide and the background winds in the middle atmosphere, it is also possible that they indicate weaker turbulent induced damping of the (1, 1) mode in the southern hemisphere. The characteristics of the semidiurnal tide agree well at the two locations.

It is found that the zonal wind components are about 6 hours out of phase and the meridonal components are in phase in all seasons. This behavior, taken together with the inferred vertical wavelengths, which are greater than 100 km in summer and ~45--80 km in winter, indicates that there are strong contributions from the antisymmetric (2, 3) mode, probably forced by wind-induced mode coupling from the symmetric (2, 2) mode. The phases near 90 km exhibit a ''bimodal'' or ''two-state'' behavior with the phases being almost constant in one state near the June solstice and then making a rapid transition of between 4 and 6 hours to another phase-stable state near the December solstice. The duration of the state near the June solstice (~6--8 months) is longer than the duration of the state at the December solstice. Although the agreement between recent models and our observations is relatively good, better agreement may be expected when more realistic wind distributions, especially for the southern hemisphere, are used in the models. ¿ American Geophysical Union 1988