Chromatic adaptation refers to the ability of the human visual system to adjust to the color of the illumination, or other prevailing stimuli, such that perceived object colors vary far less with changes in illumination than would be expected from simple radiometry or colorimetry. Models of chromatic adaptation are generally formulated as extensions of the von Kries hypothesis of some sort of independent gain control mechanisms operating on the three types of cone signals. This paper introduces a new way to model the phenomenon with no requirement for the first stage chromatic processing. This model is referred to as a spectral adaptation model since it acts upon spectra rather than chromatic signals such as tristimulus values. The spectral adaptation model was compared with other models of adaptation both computationally and through limited psychophysical data. It is shown to have reasonable, and flexible, performance and could be of practical value in applications such as spectral image reproduction. A limiting case of the spectral model, a model of perfect color constancy, is also described and compared with traditional chromatic adaptation models.