Fabry-Perot Interferometer measurements of neutral winds and European Incoherent SCATter radar measurements of plasma velocities have shown a significant equinoctial asymmetry in the average behavior of the thermosphere and ionosphere above northern Scandinavia. Existing standard models of the upper atmosphere use forcing functions that are symmetric about the solstices, therefore these observations are unexpected. It is suggested that the asymmetry arises from the diurnal variation in the cross polar cap potential difference (CPCPD) because there is a 12 hour phase difference between the variations at the March and September equinoxes. The variation in the CPCPD is caused by an annual and diurnal variation in the orientation of the magnetosphere with respect to the interplanetary magnetic field. This is known as the Russell-McPherron (R-M) effect. The plausibility of this explanation of the equinoctial asymmetry in thermospheric winds is supported by investigation of the effect of their geomagnetic history, i.e., the repercussions on the winds of the activity levels in the few hours prior to the observation. The consequences of the R-M effect have been simulated in the University College London/Sheffield/Space Environment Laboratory coupled thermosphere-ionosphere model by imposing a diurnally varying high-latitude electric field pattern. The results are used to test the predictions, given in an earlier paper, of the average behavior expected at other high-latitude sites. A corollary to the study is that the evidence presented here implies that the auroral oval may be smaller at solar minimum, which is also unexpected. ¿ 1997 American Geophysical Union