Molecular oxygen data during spring and fall are deduced from Lyman &agr; ion chamber measurements. The data show the existence of wave type structure. Random errors in the technique are estimated from simultaneous independent measurements and are found to be low. A major uncertainty in the densities results from lack of information on the appropriate absorption cross sections of molecular oxygen for the temperatures encountered in the atmosphere. Potentially large overestimates in the densities could result from absorption by nitric oxide above 105 km and to a lesser extent by nitric oxide at lower altitutdes and water vapor.