Vegard's identifications of a number of lines appearing in the auroral spectrum are discussed. The conclusion is reached that a good many of these lines may be attributed to the oxygen atom in several stages of ionization. It is shown how measured line intensities and theoretical line strengths are combined to give the excitation temperature of the auroral region. Preliminary results indicate that this temperature is in the range 3000° to 6000° K. The meaning of this result is discussed briefly.