The oxygen spectrum near 60 GHz is studied empirically with emphasis on its application to the passive remote sensing of the tropospheric and lower stratospheric temperature. For this purpose the oxygen absorption coefficient over a certain range of pressure p and temperature T is fitted to the power law form apx Ty, where a, x, and y are frequency-dependent parameters. A root-mean-square relative error of 8% is obtained by best fitting the parameters to the theoretical absorption expressions proposed by Van Vleck and Weisskopf, Gross, and Rosenkranz. This provides a convenient basis for a quantitative comparison of the three theoretical expressions. The power law form also leads to a basic understanding of the remote sensing of the atmospheric temperature profile from airborne spectrometers. Such concepts as height selectivity and resolution can be naturally defined in terms of a, x, y, and the pressure scale height of the atmosphere. As a practical application it is shown how the power law form can help attack the forward and inverse problems in remote sensing. An example is given in treating satellite data from the Nimbus 5 microwave spectrometer.