A remote sensing method has been developed for the determination of the [O>/[O-MSIS> ratio in aurora, using ratios of the O I (557.7 nm) and N+2 (391.4 nm) emissions. It is shown that the method can be used for the analysis of measurements integrated along the line of sight, provided data only above the emission rate peak are used. The method is applied to the case of horizontal viewing from a vertically oriented rocket so that a large volume of space was sampled around the rocket. The method can potentially be applied to satellite limb images, provided some independent information about the location of the aurora is available, as it was for the rocket observations. Photometric measurements of the N+2 (391.4 nm) and O I (557.7 nm) emissions obtained during the Energy Budget Campaign 1980 on flight E-2 with the instrument EF11 and its reflight in 1981 were used in the analysis presented. During the first flight the rocket horizontally viewed two distinct aurorae, a nearby diffuse patch, and a more distant pulsating aurora. Results obtained by the same EF11 instrument on a second flight through an auroral arc in 1981 are also presented. Two types of atomic oxygen variability were found in both of the flights. In the first type, [O> is increased above [O-MSIS> by a factor of 1.5 at 180 km, is equal to the MSIS model at 160 km, and is less than MSIS below that; that is, the scale height of [O> was increased. The experimental I(557.7)/I(391.4) ratio was constant with altitude. In the second type, the [O> was depleted by about a factor of 2 over the altitude range of 120--180 km, while the I(557.7)/I(391.4) ratio decreased with altitude. The inferred atomic oxygen concentrations of 0.5 to 2 with respect to MSIS suggested different vertical flows on the two cases. Independent evidence is provided by atmospheric composition measurements made during the same campaign. ¿ American Geophysical Union 1995