A model is presented for the color vision system of the honeybee, which takes the nonlinear phototransduction process in the photoreceptors into account and assumes linear computations of the excitations of the photoreceptors. The model parameters are derived by a least squares fit of the scale values determined by multidimensional scaling analysis of the results of color choice experiments to the excitation values of two hypothetical spectral antagonistic coding cells. The psychophysical scale values are interpreted physiologically. Furthermore, a color difference formula is presented which is based on the color opponent coding (COC) model. The model explains quantitatively (1) the sensitivity of spectral antagonistic neurons measured by Kien and Menzel (1977; Journal of Comparative Physiology, 113, 17-34, 35-53), (2) the color discrimination function measured by von Helversen (1972; Journal of Comparative Physiology, 80, 439-472). The following predictions are derived from the model: (1) excitation/log (I) curves of the spectral antagonistic neurons; and from the model in conjunction with the color difference formula: (2) intensity dependent color shifts (Bezold-Brücke effect); (3) the intensity dependence of wavelength discrimination.