Absorption of solar radiation by collision pairs of oxygen is a small but significant part of the total budget of incoming shortwave radiation. It is shown that the 1.26-μm band of O4 and O2⋅N2 is likely to contribute 0.64--1.55 W m-2 to the total atmospheric absorption for overhead Sun, clear sky conditions, bringing the total estimated absorption by all collision pairs of oxygen over all wavelengths from 300- to 1300-nm to 2.2--3.11 W m-2 for overhead Sun under clear sky. The globally averaged all-sky absorption by collision complexes of oxygen is likely to be about 0.9--1.3 W m-2, roughly twice the value of previous estimates neglecting the 1.26-μm band and the contribution of O2⋅N2. Significant uncertainties are shown to remain in the spectroscopy of O4 and O2⋅N2 in the near-infrared band at 1.26-μm (and possibly also at 1.06-μm), requiring modern laboratory measurements to better quantify these processes. Atmospheric measurements are also presented in this paper to verify the spectral shape of the absorption by O4 and other trace gases (O2, H2O, NO2, and O3) in the 610- to 680-nm and 400- to 450-nm regions. These observations rule out the existence of significant unknown anomalous or unexplained absorbers displaying structured features under either clear or cloudy conditions at these wavelengths. However, the possibility of anomalous broad continuum or quasi-continuum absorption cannot be evaluated using our measurement technique. The observed wavelength dependence of water vapor absorption in the bands near 442- and 650-nm are shown to agree well with their expected spectral structure under both clear and cloudy sky conditions for the instrumental resolution employed. Further, both the shape and magnitude of the measured atmospheric O4 absorption near 630-nm agree with expectations based on recent laboratory measurements of this band to better than 10% over a broad range of zenith angles. ¿ 1998 American Geophysical Union