A &thgr; aurora previously discussed on the basis of Viking images (northern hemisphere) and DE 1 images (southern hemisphere) is reexamined in light of additional data, primarily the auroral plasma distribution as determined from the Viking, DMSP F6, and DMSP F7 satellites. This event, which occurred before a substorm expansion phase on August 3, 1986, appeared in the images to consist of a single arc along the morning side in the northern hemisphere and along the evening side in the southern hemisphere and was isolated from the auroral oval in both sets of images. On the basis of the auroral plasma distribution inferred from three satellites, the brightest arcs do occur at the locations indicated by the imagers, but the arcs are in fact connected to the main oval with continuous precipitation, and weaker secondary arcs (not observed by the imagers) occur in the opposite hemisphere magnetically conjugate to the bright arcs. These observations support the interpretation of the &thgr; aurora as occurring on closed field lines as a result of the expansion of the morning and evening sector ovals into the polar cap. A careful examination of the characteristics of the observed auroral energy plasma suggests additional conclusions. It appears that the ionospheric manifestation of the recently discovered low-energy electron layer can be identified with a complicated structure of soft precipitation at the poleward edge of the main precipitation region. Finally, unlike recent reports, the ions were not observed to have a cutoff in the polar cap that is any sharper than that of the electrons. ¿ American Geophysical Union 1995