Binocular mixtures of equiluminous components of different wavelengths were matched with additive monoptic mixtures of the same components. After a satisfactory match had been achieved, the luminance of each colour in the monoptic mixture was measured photometrically. After presentation of an orthogonal grating superimposed on the colour shown to one eye, the colour matching was repeated. The grating induced a strong dominance of the colour with which it was combined. Yet the uncontoured colour was not entirely suppressed, but contributed to the binocular colour to various degrees. In three subjects with anisometropic amblyopia in one eye the colour presented to the amblyopic eye contributed little or nothing to the haploscopic colour mixture, depending on the degree of amblyopia. This diminished contribution could not be enhanced by a grid. In three cases with strabismic amblyopia and in one case with strabismus alternans no haploscopic colour mixture effects could be demonstrated. The observations are discussed in the context of neurophysiological findings in the visual system of primates, and it is suggested that colour and contour are not transmitted through independent channels from the retina to the cortex.