Satellite observations of solar EUV irradiance at wavelengths ranging from 140 to 1850 ¿ show unexpectedly great differences between cycle 21 and the previous cycle 20, posing significant problems to the applicability of current aeronomical models based on observational data gathered primarily during cycle 20. Whereas the O2-dissociating fluxes for &lgr;>1250 ¿ show but little change, the irradiance at wavelengths below 1027 ¿, known as the major source of ion production in the thermosphere, differs rather drastically not only with respect to the average absolute values but also with respect to the relative amplitudes of temporal variations of EUV compared with the associated amplitudes of the conventional activity indices, F10.7 or Rz. Both the absolute fluxes and the relative variations were found to be considerably higher for the developing new sunspot cycle compared with previously observed characteristics of cycle 20. No simple interpretation seems possible because concurrently obtained observations of atmospheric ootical depths at various EUV wavelengths suggest that the actual atmosphere around the 300-km level does not show any comparable increases in density or temperature (optical depths being essentially proportional to the product of both quantities). This seems to imply that solar-cycle-dependent hermospheric heat sources other than the solar EUV irradiance may play an important role not adequately include in the present theoretical basis of atmospheric models.