[1] The mesosphere-stratosphere-troposphere (MST) radar at Jicamarca, Peru (12°S, 77°W), made extended (15 day or longer) observations of the horizontal and vertical winds that were used to infer the diurnal and semidiurnal tides. The measurements were made during several months from mid-1997 through mid-1998 and using a higher-power transmitter and finer range resolution during 10 days of August 1998. The three-component winds are used to estimate amplitudes, phases, and momentum fluxes associated with the tides. Thermal forcing of the diurnal tide is also examined with diurnal water vapor heating rates calculated using data from the NASA Water Vapor Project (NVAP). For the region near Jicamarca the calculations from NVAP showed the temporal variability of the diurnal heating to be dominated by an annual cycle with maximum around the summer solstice. When projected into tidal modes, about 25% of the total water vapor heating rate amplitude near Jicamarca is found to be nonmigrating. The meridional amplitude of the semidiurnal tide was found to be generally greater than the zonal amplitude, although tidal theory predicts that the zonal amplitude should be much greater at the latitude of Jicamarca (assuming the tide was migrating). The phase of the semidiurnal tide lagged (by about 3 hours) the phase expected from surface pressure climatologies. According to tidal theory the migrating semidiurnal tide should transport little meridional momentum flux. However, substantial southward fluxes (hv 0 w 0 i $ À1 Â 10 À3 m 2 s À2 ) were observed at Jicamarca, and the meridional component of momentum flux was typically larger in magnitude than the zonal component was. The diurnal tide was somewhat weaker, was less coherent, and transported less momentum. The semidiurnal tide had a very long vertical wavelength throughout the troposphere and into the lower stratosphere, while the diurnal tide was only observed to propagate at heights above the tropopause with a much shorter ($10 km) vertical wavelength. Below the tropopause the dominant diurnal motions were not traveling waves, but rather convective motions that exhibited little phase progression with altitude. These motions were broadly peaked in frequency around 24 hours and were presumably standing oscillations with no horizontal propagation and probably with small horizontal scale. Despite the lack of coherence of these quasi-diurnal motions, the associated vertical wind amplitudes were sizable ($0.02 m s À1 ), and thus the fluctuations can presumably transport significant horizontal momentum.