A brief introduction to colour vision and deficiencies

Normal colour vision

The first handbook on the Cambridge Colour Test (CCT) was written by Mollon and Regan in 2000. Below we provide a modified version of their a section of their original manuscript, which has been adapted to describe the features of the CCT for the Metropsis Visual Function Assessment system.


Normal colour perception depends on the absorption of light by the three classes of cone photoreceptor in our retina. The peak sensitivities of the three cones lie in the violet, the green and the yellow-green parts of the spectrum – at wavelengths of approximately 420, 530 and 560 nanometres. It is convenient to refer to the three types of cone as short-wave, middle-wave and long-wave respectively. Their spectral sensitivities are shown schematically below.


Any individual cone obeys the Principle of Univariance: although the stimulus for the cone can vary in two dimensions (radiance and wavelength), the output can vary in only one dimension – the degree of hyperpolarization of the cell. As wavelength varies, there is a variation in the probability of a given photon being absorbed, but once a photon has been absorbed, all information about its wavelength or frequency is lost. So the individual cone, or class of cones, is colour blind. Lights of two widely different wavelengths will produce the same output providing that the two radiances are suitably adjusted. So to analyse colour, the visual system must compare the rates at which quanta are being caught in different classes of cone; this comparison is initiated at a retinal level, in that ganglion cells of the ‘midget’ and the ‘small bistratified’ types draw inputs of different sign from different classes of cone.

Since the normal retina contains just three classes of cone and since each class of cone obeys the Principle of Univariance, all lights can be represented in a three-dimensional space, the ordinates of which correspond to the rates of quantum catch in the short-wave, middle-wave and long-wave cones, as illustrated below.


Any given light will plot as a point in such a space. As the light’s radiance is increased or decreased, the plotted point will move along a vector that passes through the origin. The angle of the vector – its direction relative to the three axes – corresponds to the chromaticity of the light. Lights that plot at the same point in the space, though they have different spectral compositions, will appear alike. However, the chromaticity of the light – the relative quantum catches it produces in the three cones – is not sufficient to tell us the colour that it will appear: for the appearance of a light depends not only on its own chromaticity, but also on the chromaticities of other stimuli in the field and on many other factors, such as the recent history of stimulation.



Owing to the simple nature of this space, it is possible to choose a pair of physical stimuli that yield the same quantum catches in two of the three classes of cone but differ with regard to the remaining class. By alternating between such stimuli, we can probe the integrity of one class of cones in isolation. This principle is exploited in the present colour test, as it has been in many earlier tests. Consider, for example, a point that lies in the plane defined by the quantum catches of the middle- and long-wave cones – the plane of the paper in the present diagram. Now imagine a line passing through this point and orthogonal to the plane of the paper. Lights along the line vary only in the quantum catch they yield in the shortwave cones. The line is called a ‘tritanopic confusion line’, because someone who lacks the short-wave cones – a ‘tritanope’ – will confuse all chromaticities that lie along the line. By alternating between analogous pairs of lights, it is similarly possible to detect those who lack the middle-wave cones or the long-wave cones and who are called ‘deuteranopes’ and ‘protanopes’ respectively. About 2% of the male population are congenital dichromats of the deuteranopic or protanopic kinds.


Congenital tritanopia

Unlike deuteranopia and protanopia, is not sex-linked, and it is much rarer. However, diseases and drugs that affect the receptor layer of the retina appear disproportionately to affect the short-wave cones. The acquired colour deficiency is often of a tritan type, in that thresholds for discrimination are elevated along a tritan confusion line.


Anomalous trichromacy

About 6% of men exhibit anomalous trichromacy, a congenital form of colour deficiency that is milder than dichromacy. Unlike dichromats, they require three variables in colour-matching experiments, but they make different matches from the normal observer and, in most cases, their discrimination of colours is poorer than normal. Currently, the predominant view is that the anomalous trichromat lacks either the normal long-wave or the normal middle-wave pigment, but achieves his residual discrimination in the red-green range by a neural comparison of the quantum catches in two slightly different versions of the middle-wave pigment or two slightly different versions of the long-wave pigment. Anomalous trichromats who behave as if they lack the long-wave pigment are called ‘protanomalous’ and those who behave as if they lack the middle-wave pigment are called ‘deuteranomalous’. Characteristically, the two types of anomalous trichromat show reduced discrimination between chromaticities that are confused by the corresponding type of dichromat: their thresholds, when plotted in a normal colour space, typically form an ellipse oriented along a protan or a deutan confusion line. However, the term ‘anomalous trichromacy’ covers a large variety of phenotypes: some anomals may exhibit discrimination that is nearly as poor as that of the corresponding dichromat, a few enjoy colour discrimination within normal limits, while the majority have discrimination ellipses that lie somewhere between these extremes.

The basis of this phenotypic variation is not yet well understood, but some part of is likely to depend on the spectral separation of the two long-wave or two middle-wave pigments on which the anomal is thought to depend for his residual discrimination. Indeed, the distinction between dichromacy and anomalous trichromacy is nowadays less clear-cut than was traditionally held. For it is possible that some of those classified as dichromats by the Nagel anomaloscope in fact express two genes, encoding pigments with very similar spectral sensitivities, but differing very slightly in peak wavelength or optical density: under some circumstances these residual differences may sustain some limited discrimination in the redgreen range.


Post-receptoral channels

The trichromatic colour vision found in Man and in the Old World primates did not evolve in a single step, and it is increasingly clear that our colour vision depends on two subsystems, which emerged at different times. These two subsystems remain morphologically distinct at early stages of the human visual system.

A phylogenetically ancient subsystem compares the signal of the sparse short-wave cones with some combination of the signals of the long- and middle-wave cones. Forming the anatomical substrate for this pathway are the ‘blue cone bipolar cells’, the small bistratified type of ganglion cell, and cells of the koniocellular laminae 3 and 4 of the lateral geniculate nucleus.

A duplication of a gene on the X-chromosome is thought to have led to the newer subsystem of colour vision, sometime after the divergence of the Old and New World monkeys. This second subsystem compares the quantum catches of the long- and middle-wave pigments, which diverged from a single ancestral pigment after the duplication. Its signals are thought to be carried by the midget bipolars, the midget ganglion cells, and cells of the parvocellular laminae of the lateral geniculate nucleus.

The cells of these two pathways are distinct, in morphology, in size and in immunoreactivity. So we might expect them to be differentially affected by particular diseases, toxins or drugs. If the phylogenetically older pathway is affected, we can expect thresholds to be elevated along a tritanopic confusion line. If the phylogenetically younger pathway is affected, we can expect a discrimination ellipse that is elongated in a direction intermediate between the protan and deutan confusion lines.