Colour-matching data and the spectral absorption curves of visual pigments

Abstract

The failure of the hypothesis of three visual pigments with spectral absorption curves of rhodopsin type as an explanation of the colour-matching data of the normal trichromat, is examined in a detailed calculation. Surprisingly, the hypothesis predicts for the chromaticity coordinates of the spectral colours (W.D.W. system), curves that are in satisfactory agreement with the empirical values. However, the necessary choice of the peak wavelengths of the pigments to obtain this result entails (a) a peak wavelength for the “red” sensitive pigment at 617 nm which other evidence shows to be much too far in the red, and (b) a completely unacceptable transmission curve representing light losses in the eye prior to visual absorption. The principal responsibility of the long-wave visual pigment for the failure of the rhodopsin-type hypothesis is emphasized in a second calculation that shows that no visual pigment consistent with the empirical colour-matching data can have a peak wavelength greater than 591 nm, and that if the absorption curve is of rhodopsin type its peak wavelength must be less than 556 nm. For the pigment associated with the “green” sensitive cone process and absorbing maximally at about 540 nm a rhodopsin shape is not inconsistent with the colour-matching data.