We use overhead ionospheric incoherent scatter radar data from Millstone Hill to deduce the stormtime variations of the thermospheric temperature, O and N2 densities, and F1 region ion composition during the Equinox Transition Study of September 17--24, 1984. Using the measured neutral temperature profile and O density at 400 km altitude, we extrapolate to determine O density changes in the lower thermosphere. Using these O densities and the measured electron densities, we deduce the variation of the N2 density. Our initial attempt leads to the deduction of large depletions in both O and N2 in the lower thermosphere. On using the MSIS model neutral temperatures for the extrapolation, however, we compute O depletions having magnitudes more comparable with theoretical predictions and N2 variations consistent with expected thermal expansion effects. We speculate that the radar-measured neutral temperatures were contaminated with ion-neutral frictional heating effects during the disturbed periods of the ETS and that this served to invalidate neutral density extrapolations made with these temperatures. Further, we compute the height of 50% O+ ion composition in the F1 region and the height of the peak electron density in the ionosphere, and, upon comparing these, find that the height of the peak on the disturbed days was located in the lower F1 region molecular-ion layer rather than in the F2 region O+ layer. This occurrence has serious consequences for methods of deducing neutral winds from the height of the F layer, for these methods rely on the peak containing primarily O+ ions. ¿ American Geophysical Union 1990