Owing to their sampling properties in local solar time, both the TIMED and UARS spacecrafts provide the opportunity to quantitatively study the mesospheric thermal tides. The information though is limited because it takes the TIMED instruments 2 months to cover the full range of 24 hours, almost twice as long as UARS. We present here the results of an analysis, in which the seasonal and global-scale variations of amplitude and phase for the migrating diurnal tide are retrieved from atmospheric temperature and wind measurements on the TIMED (SABER, TIDI) and UARS (MLS, HRDI) satellites. Comparisons are made with results from the Numerical Spectral Model (NSM), which is characterized in part by the parameterization of small-scale gravity waves. At 60 km and above, the amplitude and phase variations of the diurnal tide from the NSM agree generally well with those derived from the SABER temperatures. In the SABER temperature amplitudes at around 60 km there are evident asymmetries in latitude with respect to the equator, which are not apparent in the MLS data near 55 km or in the NSM. (The asymmetry in the SABER temperatures disappears above 60 km.) For the meridional winds of the migrating diurnal tide derived from TIDI, the phase and relative amplitude variation with latitude and season are in qualitative agreement with the HRDI measurements and the NSM. The amplitudes from TIDI winds are systematically smaller, but this may in part be due to interannual variations. However, the results for the zonal winds from TIDI more often do not agree well with those from HRDI or the NSM.