Sodium superoxide (NaO2) is believed to be a major sink for meteor-ablated Na in the upper atmosphere. However, the rate constant for the reaction between NaO2 and O is not known, and its determination is the subject of this study. NaO2 was produced by the reaction between Na vapor, and excess of O2, and the carrier gas N2 in the upstream section of a fast flow tube reactor. Excess atomic O was then added, so that a steady state developed in which the sodium was partitioned between its atomic form and a variety of oxides. The steady state fraction of atomic Na, monitored by laser-induced fluorescence, was then observed as a function of and 2>, from which it was deduced that k(NaO2+O)=(2.2¿1.0)¿10-11 cm3 molecule-1 s-1 at 300 K (3&sgr; uncertainty). As a prelude to this experiment, the recombination reaction between O2 and Na was studied by the conventional flow tube technique, yielding k(Na+O2+N2)=(3.14¿0.20)¿10-30 (300/T)-(1.52¿0.27) cm6 molecule-2 s-1. These experimental results were then incorporated into a one-dimensional model of sodium in the upper atmosphere between 65 and 110 km. It is shown that formation of NaHCO3 rather than NaO2 is the dominant removal process for atomic Na below 90 km. This new model, whose only important assumption is the rate constant for the reaction NaHCO3+H→Na+H2CO3, predicts a seasonal variation of the atomic Na layer in excellent agreement with recent lidar observations at 40¿ and 69 ¿N and also demonstrates that temperature fluctuations produced by gravity waves should induce significant chemical responses in the Na layer below 88 km. ¿ American Geophysical Union 1993