Thermospheric and ionospheric behavior over a 6-day period beginning on September 24, 1981 (day 267), is discussed using DE 1 far ultraviolet dayglow data and ground-based ionosonde data. The DE 1 data are processed to produce images of O/N2, the column density of O relative to N2. Ionospheric data are in the form of Nmax, the maximum electron density that is equivalent to NmF2 provided an F2 layer is present. Day-to-day variations in these parameters and their relationship to one another are examined in three geographical regions: Siberia/Japan, Europe, and North America. The 6-day period begins and ends with undisturbed days and includes a 2-day geomagnetic storm period with onset on day 269. Several disturbances in the form of reduced O/N2 are observed, with the most severe occurring during the storm. For those ionosonde sites located within the disturbed regions, negative ionospheric effects are observed similar to those reported by Strickland et al. [2001> for a different storm period. In Siberia, 4 consecutive days of negative effects are observed that coincide with reductions in O/N2. Middle- to low-latitude enhancements in O/N2 are also observed that are more global and less structured than regions of reduced O/N2. Percentage changes in the enhancement from quiet time values are typically less than ~20% in contrast to much larger changes in well-developed regions of reduced O/N2. Several instances of positive ionospheric effects are observed that are in phase with coincident O/N2 enhancements, but the coupling is less convincing than that between negative ionospheric effects and O/N2 reductions. It is, nevertheless, seen often enough in this work to argue that compositional effects in addition to dynamical effects must be considered in addressing the causes of long-duration positive ionospheric storm behavior. ¿ 2001 American Geophysical Union