1. Eight deuteranopes matched a mixture of a monochromatic light on the long wave side of the neutral point and a violet (450 nm) primary to a fixed white as well as a monochromatic light on the short wave side of the neutral point mixed with a red (650 nm) primary, to the same white. For lambda greater than 530 nm, the former set of matches defined the action spectrum of the long wave sensitive foveal cones, and for lambda less than 480 nm, the latter that of the short wave sensitive cones. 2. Individual differences in the former matches were approximately correlated with the respective ratio of the sensitivities of the wave-length of the anomaloscope primaries, in a way that individual differences of the latter were not. 3. Assuming that eye media differences alone account for the differences in long wave sensitive foveal action spectra, the spectral reflectivity of the foveal fundus was predicted for these deuteranopes. The prediction is inconsistent with measurement. 4. Thirteen deuteranopes matched monochromatic spectral lights with a green (535 nm) and a blue (460 nm) primary. The result were analysed by von Kries' method in which differences in matching due to differences in eye media absorption are obviated. The matches of five differed significantly from one another when so analysed. It was concluded that at least one of two action spectra of the foveal cones of every one of these five differed from that of all of the others. 5. The canon that deuteranopes accept normal colour matches was evaluated by confronting a single normal with five deuternopes in the analytical anomaloscope of Baker & Rushton, set in the mode of each of the five in turn. Obvious differences existed between this normal's matches and those of four of five deuteranopes. 6. Explanations for differences in the spectrum of erythrolabe in different deuteranopes are evaluated. The possibilities that all have the identical visual pigment but (a) in cones with different optical funnelling properties or (b) in different optical densities are considered. Preliminary results are not in agreement with the expectations of either of these ideas. 7. It is suggested that the visual pigment in the foveal long wave sensitive cones of different deuternopes (and of different normals) may have different extinction spectra. The idea is consistent with micro-spectrophotometric measurements of rhodopsin in individual rods from different frogs (Bowmaker, Loew & Leibman, 1975).