Who has not enjoyed the splash of colour in a market: gorgeous red peppers, the green of basil and what on earth are these purple vegetables over there? All thanks to trichromatic vision, another story of convergence.

Although vertebrates are primitively tetrachromatic (many fish, reptiles and birds have four different types of cone cells, each sensitive to distinct light spectra), the mammals are generally dichromatic (having only two cone types in the eye). A popular notion is that this reflects a transition to a nocturnal way of life. In the primates, however, trichromatic vision has re-evolved – there are three types of cone cell, differentiating colour in the red, green and blue regions of the spectrum. Trichromacy is linked not only to largely diurnal activity, but also to the adaptive advantages of being able to detect coloured fruit and leaves. An important corollary of this argument is that with the shift to acute vision the role of olfaction has become less important. There is probably a lot of truth in this idea, and certainly relative to other mammals (notably the rodents) human olfactory capacity is seriously impaired with many of the relevant genes turned off by being transformed into pseudogenes. However, it is not an absolute rule and the howler monkeys of the New World, which have independently evolved full trichromatic vision, still retain important pheromonal capacities. The story of mammalian colour vision has other complications…

The molecular basis of colour vision

The vertebrate retina contains two classes of photoreceptor cells – rods that are responsible for low-light vision and cones that function best in bright light and provide the basis for colour vision. Seeing colour means being able to discriminate lights that differ in their spectral composition, so at least two types of photoreceptor must be present that are sensitive to light of different wavelengths. The spectral tuning of a photoreceptor is due to a visual pigment called opsin – only a few amino acid changes in this protein shift the receptor’s spectral sensitivity. The ancestral vertebrate set of cone opsins, which has been retained in the tetrachromatic fish, reptiles and birds, comprises two shortwave-sensitive classes (SWS1 and SWS2), one middlewave-sensitive class (RH2) and one longwave-sensitive class (LWS) with spectral sensitivities ranging from ultraviolet to red. Most mammals have lost SWS2 and RH2 and are thus dichromatic with long- to middlewave-sensitive L-cones and shortwave-sensitive S-cones. The absorption spectra of the cones can differ markedly between species, varying from green to yellow to orange for L-opsin and from blue to violet to near-UV for S-opsin, and it is assumed that these variations represent adaptations to the species’ visual ecology. In contrast, the spectral tuning of photopigments is remarkably conservative among the trichromatic primates – they possess red (L), green (M) and blue (S) cones.

These denotations are somewhat misleading, as the receptors themselves are colour-blind – it is the comparison of their signals in the brain that generates a colour signal. Therefore, not only different receptor types are necessary for the perception of colour, but also the neural mechanisms for comparing their responses. A greater number of photopigments allows for more comparisons to be made and thus enhances the potential for colour vision. The colour vision of dichromats probably resembles that of red-green colour-blind humans, whereas the comparison of L- and M-cone signals adds a red-green mechanism in trichromats.

Evolution of trichromacy in primates

In all dichromatic primates, the S-opsin gene is located on an autosome, while the L-opsin gene is found on the X-chromosome. Becoming trichromatic meant acquiring an additional opsin gene that was expressed such that it produced a novel chromatic signal and the neural circuitry to process this colour information. This has happened in different ways and several times independently in primates…

Routine trichromacy

In the Catarrhini (Old World monkeys and apes), a duplication of the L-opsin gene has occurred, resulting in functionally divergent M- and L-opsin genes that are both located on the X-chromosome. This is referred to as routine trichromacy. Interestingly, this gene duplication seems to have happened independently and more recently in the New World howler monkeys (genus Alouatta). It is generally assumed that the L- and M-opsin genes of the Catarrhini diverged after allele duplication, while it might have been the other way round in howler monkeys – different opsin alleles arose and then became fixed as different loci by duplication.

Allelic trichromacy

In other New World monkeys (Platyrrhini), the evolution of trichromacy is a rather different story. They also have an autosomal S-opsin gene but only one X-linked opsin gene. This gene, however, is polymorphic, meaning that it encodes different M- to L-opsins. The interesting consequence is that heterozygous females are trichromatic (as they possess two different variants of the gene), whereas males and homozygous females are dichromatic. This has generated a great deal of discussion and there is still a prevailing sense of puzzlement. What is generally agreed is that this allelic (or polymorphic) trichromacy is adaptive, not least because it evidently evolved millions of years ago and is clearly stable. For example, both dichromacy and trichromacy could convey a foraging advantage in different situations.

Allelic trichromacy is, however, not limited to the New World monkeys. DNA sequence analysis has recently revealed that it has evolved independently in some Strepsirrhini, more “primitive” primates, such as lemurs, aye-ayes and bush babies. While many strepsirrhines are nocturnal and mono- or dichromatic, some show a genetic capacity for trichromacy (with its behavioural relevance still being discussed). These include the Madagascan Coquerel’s sifaka (Propithecus coquereli), red ruffed lemur (Varecia rubra) and black and white ruffed lemur (V. variegata). As these three species are diurnal, it has been suggested that strict diurnality was crucial for the evolution of trichromacy in Strepsirrhini. However, not all diurnal lemurs show the potential for allelic trichromacy and a recent study has demonstrated an opsin gene polymorphism in the blue-eyed black lemur (Eulemur macaco flavifrons), a species that is cathemeral (i.e. active during day and night). While this discovery does not disagree with the widely acknowledged hypothesis that trichromacy has been acquired several times independently in the Strepsirrhini, it would also be consistent with a more ancestral origin of trichromacy in this group.

Trichromatic marsupials

Until recently, trichromatic colour vision amongst mammals was thought to be unique to primates, but then three spectrally distinct cone types were detected in some marsupials (and behavioural trichromacy demonstrated by colour mixture experiments in at least one of them). These are the fat-tailed dunnart (Sminthopsis crassicaudata), the honey possum (Tarsipes rostratus), a small wallaby (the quokka Setonix brachyurus) and a bandicoot (the quenda Isoodon obesulus). Trichromacy in the quenda and the cathemeral fat-tailed dunnart differs from that of primates in that their S-cones confer sensitivity to UV light. The L-cones of the insectivorous fat-tailed dunnart are particularly sensitive in the green-yellow region of the spectrum and could, in combination with the M-cones, allow for detection of cryptically coloured green and brown prey. Also in the honey possum, the spectral sensitivities of its photopigments can be linked with its visual ecology. The L-cones are tuned to longer wavelengths than in the other marsupials studied, conferring sensitivity in the yellow-red region of the spectrum, and it has been suggested that this could be related to its unusual diet of nectar and pollen. Indeed, the L-cones seem to be well adapted for detecting food-rich target flowers and distinguishing them from non-target flowers of other species. However, with respect to these tasks even longer wavelengths would be advantageous, which implies a constraint of some sort. This constraint could be molecular or, more likely, another ecological pressure. For foraging to be efficient, the honey possum needs to determine the maturity of flowers (particularly of its major food plant Banksia attenuata), for which its L-cone tuning seems to be close to optimal.

As these four species are phylogenetically distant, it has been suggested that routine trichromacy could be a common feature of marsupial mammals. However, marsupial trichromacy is not especially well documented, and it is also unclear whether marsupials (in contrast to placentals) have retained M-opsin from their ancestors or whether it has re-evolved, analogous to the situation in primates. Some researchers favour the former scenario, arguing that while placentals mainly escaped from predators by becoming nocturnal, thus reducing the need for trichromacy, the Australian marsupials remained diurnal in the absence of large predators and might therefore have benefited from retaining the ancestral photopigments. However, at least some Australian marsupials, such as the tammar wallaby (Macropus eugenii), are probably dichromatic, supporting an independent origin of trichromacy in this group.

What may be a holdover of a more ancestral trichromacy is seen in the platypus (Ornithorhynchus anatinus), a member of the oviparous monotremes, which split from the viviparous Theria (marsupials and placentals) about 200 million years ago. They are unique amongst mammals in that they have retained visual pigments of the SWS2 class, while losing the SWS1 gene.