Energy transfer from N2(A³Σu+) to O has been extensively studied as a mechanism for the production of 5577‐Å emission in the aurora, but it has yet to be fully accepted as the dominant process. This reluctance arises from the three requirements that must be met. First, the correct absolute emission rate must be produced by this mechanism. Second, the lifetime of N2(A) at auroral altitudes must be compatible with the observed time delays between the N2+ and 5577‐Å emissions. Finally, the generally accepted high degree of covariance between the N2+ and 5577‐Å emissions must be explained. This last aspect is the subject of this paper, in which we investigate this covariance in terms of spectral ratios. Measurements made from the Intercosmos Bulgaria 1300 satellite include the emission rate of the O I 5577‐Å and N2+ 4278‐Å emissions, and the flux of precipitating auroral electrons in the energy range 200 eV to 15 keV. Using the measured fluxes and atmospheric composition models, the emission rate ratio I(5577)/I(4278) for the N2(A) energy transfer mechanism is calculated and compared with the observed values. The observed ratios vary mostly from 2 to 5, with a few values outside this range. We show that this variation arises from variations in the atomic oxygen concentration, partly as an altitude effect associated with changes in primary electron energy, but more dominantly as day‐to‐day variations that are larger than those indicated by the mass spectrometer/incoherent scatter 1983 (MSIS‐83) model. Independent evidence that atomic oxygen is the source of the ratio variations is provided by correlated variations in the I(7320)/I(4278) ratio. The aspects of absolute emission rate and time delay are also reviewed, and it is concluded that the evidence is conclusively in favor of N2(A) energy transfer as the dominant auroral mechanism.