Perception of colour in unilateral tritanopia.

The unilateral tritanope described in the previous paper (Alpern, Kitahara & Krantz, 1983) was able to match every narrow-band light presented to his tritanopic eye with lights from a tristimulus colorimeter viewed in the adjacent field by the normal eye. In two regions of the spectrum (called isochromes) physically identical lights appeared identical to the observer's two eyes. One isochrome was close to 'blue' for the normal eye, the other was in the long-wave spectral region seen by the normal eye predominantly as 'red'. Between these isochromes the normal eye required less than spectral purity to match, dropping to near zero purity at 560-570 nm. A mixture of the two isochromes that appeared purple to the normal eye appeared neutral to the tritanopic eye. Hence dichoptic matches grossly violate Grassmann's additivity law. For the normal eye colour naming conformed to typical normal results. For the tritanopic eye the results were coherent with those found by dichoptic matching: the spectrum was divided into two regions by the achromatic neutral band. To the short-wave side, only the colour names 'blue' and 'white' were ever used. To the long-wave side the predominant colour names were 'red' and 'white' with some 'yellow'. Spectral lights appeared neither 'red-blue' nor greenish. Surrounding the test with an annulus either 430 nm, 650 nm, or a mixture of these, fails to induce any greenish appearance, although the achromatic band shifted in the expected directions. It is concluded that there must be exactly three functionally independent, essentially non-linear central codes for colour perception, and that these codes are different from those suggested in existing theories of colour perception.