Excitation processes for [Formula: see text] in aurorae, in the nightglow, and in laboratory sources are discussed. It is shown that the observed vibrational distribution in aurorae is consistent with the [Formula: see text] + NO charge-transfer mechanism. Arguments are presented for the case that quenching of O2(b) in vibrational levels above ν′ = 1 is rapid, and that therefore the auroral source is much larger than previously supposed. It is suggested that oxygen atoms are an efficient quencher for O2(b) levels above ν′ = 1.