An algorithm for retrieving the first two moments of the photon path length probability density function for both the oxygen A-band and the 0.820 ¿m water vapor band from measurements of the second generation Rotating Shadowband Spectrometer (RSS) is developed and applied to data from the Atmospheric Radiation Measurements (ARM) Program Southern Great Plains (SGP) site. In the algorithm, solar direct-beam measurements are used to characterize the instrument response function in pixel (i.e., wavelength) space. By using nonlinear least squares regression with a two-parameter gamma function constraint, the mean and variance of the photon path length probability density function for cloudy skies are subsequently retrieved from spectral measurements in both bands. Two case studies illustrate that the variance of the photon path length probability density function is more sensitive than the mean of the probability density function to vertical cloud structure. Interestingly, the first two moments of the photon path length probability density function appear to exhibit sufficient sensitivity to detect cirrus that the ARM SGP millimeter-wave cloud radar failed to detect. Photon path length probability density functions from both the oxygen A-band and 0.820 ¿m water vapor band provide additional insights into radiative transfer through a variety of cloudy conditions, improving our understanding of water vapor absorption of solar radiation in these conditions.