SEEING COLOR: Stylized image of the eye’s photoreceptors, the rods and cones, which respond to particular of wavelengths of light to give animals their sense of sight© ALFRED PASIEKA/SCIENCE PHOTO LIBRARY/GETTY IMAGES

In a steamy Eocene jungle, a newborn monkey opens its eyes for the first time. The world it sees is unlike any other known to its primate kin. A smear of red blood shines against a green nest of leaves. Unbeknownst to its mother, this baby is special, and its eyes will shape the human experience tens of millions of years in the future. Were it not for this little monkey and the series of genetic events that created it, we might not have the color vision we do: Monet’s palette would be flattened; the ripeness of a raspberry would be hidden among the leaves; traffic lights? They likely would never have been invented.

Like most...

Adding a third opsin gene doesn’t simply introduce 50 percent more colors; its effect is multiplicative. If a single opsin gives an animal the ability to distinguish 100 shades, say, the addition of a second opsin, “amazingly, multiplies that by 100,” says color vision researcher Jay Neitz of the University of Washington in Seattle. “Adding a third photopigment has been the greatest invention of all, because it multiplies color vision by another 100 times.”

Such a profound expansion of our visual experience actually required very minor genetic alteration. In 1991, Neitz, working with his wife Maureen and their postdoc advisor Jerry Jacobs of the University of California, Santa Barbara, demonstrated that just three amino acid substitutions account for the 30 nm difference in peak absorption between the modern-day red and green cones in humans, with each change shifting the photopigment’s color spectrum by 5 nm to 15 nm.1 “It’s absolutely stunning,” says Jacobs. “A single nucleotide change can change your color vision.” (See illustration below.) Yet, despite this simplicity, the evolutionary circumstances that allowed our primate ancestors to adopt trichromacy—the three-cone system that gives humans and some other primates the ability to see the world in full-spectrum color—are remarkably intricate.

To understand how that first trichromatic monkey and its similarly equipped primate descendants responded to their heightened sense of sight remains an ongoing quest. But experiments by the Neitzes and others that provide dichromatic animals, such as mice or squirrel monkeys, with an extra opsin are helping to fill in the story of the evolution of human color vision. The results suggest that the first trichromatic monkey may have been able to respond immediately to its new, more vibrant world—see the ripe fruit among the green buds; the red ants on the leaves. The work may also point the way to a future in which scientists could treat color blindness by replacing malfunctioning opsin genes, and perhaps, one day, even supercharge humans’ color perception to reveal a new rainbow altogether.

A colorful duplication

It’s 1980. Jeremy Nathans, then a graduate student at Stanford University, is driving back to campus after visiting a slaughterhouse in San Jose. Beside him jiggles a bucket of cow eyeballs on ice.

He’s heading to the lab of his advisor, David Hogness, where he plans to use the eyeballs, along with a revolutionary new tool called recombinant DNA, to answer a question that had been posed decades before: What is the molecular basis of color vision? “It seemed clear to me that the way to solve these problems was not to study light-absorbing proteins, which are extremely rare, hard to work with, and intermixed with far more abundant proteins,” says Nathans. “This was a problem that was going to be solved by going directly to the genes.”

At the time, very few human genes had been cloned, and recombinant DNA methods were crude. “It was hard,” recalls Nathans, now a professor at Johns Hopkins University School of Medicine. “We went up quite a few wrong paths.” But four years of labor eventually paid off. In 1983, Nathans and Hogness published the amino acid sequence of bovine rhodopsin, and a year later they published the human rhodopsin sequence.2

Rhodopsin, expressed in rod photoreceptor cells, enables animals to see in dim light. To understand color vision, Nathans and his colleagues had to track down the three opsins embedded in the cell membranes of the three varieties of cones, which absorb short (blue), medium (green), or long (red) wavelengths of light. Fortunately, despite about a billion years of divergence between them, rhodopsin and the cone opsin genes shared enough sequence homology for the known sequence to serve as a probe for the unknown genes.

What makes the tale of primate color vision so special is that it can be told, from beginning to end, in exquisite genetic and molecular detail.

A couple of years later, using his own DNA, Nathans and colleagues cloned the cone opsins. Two of them—the red and the green—reside on the X chromosome and are 96 percent similar in their amino acid sequence. The results provided support for the idea that an ancient X-linked opsin gene underwent a single duplication event and that subsequent mutations in the copy shifted the absorbance spectrum of the photopigment.3

“The work illuminates not only the physiology of color vision, but basic mechanisms of evolution,” geneticist David Botstein, now at Princeton, wrote in a commentary accompanying Nathans’s paper in Science. “It has been thought for some time that a major theme of evolution is duplication followed by divergence.”

But, as it would turn out, duplication and divergence are not the whole story.

Around the same time that Nathans was cloning human opsin genes, the Neitzes were working on squirrel monkeys in Jacobs’s Santa Barbara lab. Previously, Jacobs had found that these New World monkeys do not have the same color vision as Old World primates and humans. In particular, squirrel monkey color vision is highly polymorphic—some females see much larger ranges of shades than males or other females.

Digging into the genetics of this unusual variation, Jay Neitz and his colleagues discovered that while squirrel monkeys have just two opsin genes (one on chromosome 7 and one on the X chromosome), they have several opsin alleles. Three alleles, which resemble the human red and green opsins, are present in the same locus on the X chromosome. With just one X chromosome, all males are dichromats, but because females carry two Xs, they can carry two different alleles for the X-linked opsin gene, granting such heterozygotes trichromatic vision.4

“The first step then to getting to trichromacy [in primates] was to just get diversity, polymorphism, in the one gene [they] have,” says Neitz. Then, somewhere along the line a genetic translocation likely plucked an opsin allele from one X chromosome and plunked it next to a different opsin allele on the other X chromosome, giving that animal two opsin genes adjacent to one another, as humans have today.

“The fact that we have the same few amino-acid substitutions as New World monkeys argues there was a single ancestral variation that gave rise to [the cone opsins of] both Old and New World primates,” Nathans says. “It leads to an interesting twist on the evolutionary dogma of gene duplication.”

X marks the spot

In Living Color

Human color vision is based on the different wavelengths of light absorbed by three cone opsin proteins, which are responsible for the spectral tuning of the cone cells in the retina. The red and green opsins, whose genes reside on the X chromosome, are thought to have evolved from an ancestral cone opsin gene that duplicated itself.


Three amino acid substitutions in the red opsin protein account for the spectral tuning of the green opsin. At position 180, swapping serine for alanine produces a 6 nm shift of the absorption spectrum; tyrosine to phenylalanine at position 277 provides a 9 nm shift; and changing a threonine to an alanine at position 285 confers another 15 nm shift in maximum absorption. Together these three changes produce the 30 nm gap between the maximum absorption of the red and green opsins.


While a gene duplication of the X-linked opsin was necessary to grant all the animals in a group, including the males, trichromatic vision, this part of the color vision story is, once again, not so simple. There needed to be some mechanism to ensure that both genes on the X chromosome were not coexpressed in the same cone cell.

The prevailing model for how the brain discriminates colors is that it assigns cone classes—green, red, or blue—to each cell by comparing how it and its neighbors respond to various wavelengths. For instance, if red light hits the eye and one cone activates while an adjacent cone stays silent, then the brain figures out that those two are in different cone classes. But cone cells are only useful in discriminating colors in this way if each cell expresses only one type of opsin. If an individual cone carried two different opsins and responded to the absorption spectra of each, its firing wouldn’t be very informative. So how do cone cells, which carry an organism’s full genome and thus the genes for all three opsins, limit the expression of two of them?

Researchers are still unsure how the blue opsin gene, positioned on an autosomal chromosome, is not coexpressed with either of the opsins on the X chromosome, but Nathans’s work has yielded clues regarding the mechanism that allows only one of the two X-linked opsins to be expressed in a given cell. Studying people who have only one functional opsin that absorbs short wavelengths of light and can thus discriminate only among blue hues, Nathans discovered that some of these so-called blue-cone monochromats had deletions about 4,000 base pairs upstream of the red and green opsins on the X chromosome.5 “It smelled like an enhancer,” says Nathans, referring to short genetic sequences adjacent to promoters that help initiate transcription. He later showed in transgenic mice that this enhancer sequence is required for the expression of red and green opsins and that it selects which one will be transcribed.6

Importantly, the opsin gene duplication on the X chromosome did not include the enhancer, resulting in a single enhancer being responsible for turning on both genes. But it acts on only one opsin gene in any given cell—likely chosen at random—meaning that the enhancer will lead to the expression of the red opsin in one cone and the green opsin in another. Without this mechanism, cones would likely express a gobbledygook of green and red opsins, and our perception of color would be drastically different.

Brain power

Whether that first trichromatic monkey could actually take advantage of the expression of all three opsin alleles—whether it could see the blood on the leaves as a different color—is not entirely clear. And in fact, a basic quandary in evolutionary biology is how animals process new sensory input. “When she got this new cone, did she say, ‘Great, this is a nicely colored world,’” says Jacobs, “or did she say, ‘Oh, now I suppose I have to redesign my nervous system’?” Around 2000, he and Nathans got the opportunity to answer that question.

The researchers replaced one of the medium-wavelength opsin alleles on the mouse X chromosome with a human long-wavelength opsin gene to create a line of trichromatic mice. They then trained the mice to select the color in a panel of three that differed from the other two. Next, the mice were tested for their ability to perform this task across a range of shades, including those within the absorption spectrum of the new opsin. The newly trichromatic animals excelled at the task. In 2007, the team published data showing that heterozygous female mice, which carried the long-wavelength gene on one X chromosome and the medium-wavelength gene on the other, plus their short-wavelength gene on chromosome 6, were able to discriminate additional colors compared to animals with only medium- and short-wavelength alleles. This suggested that the animals’ brains were making use of the new opsin.7 “I think that’s the cool part of it, that [plasticity] is just an intrinsic property of a sophisticated nervous system like a mammal’s,” says Nathans.

A couple of years later, the Neitzes and their colleagues tried something similar with adult squirrel monkeys, using viral-vector gene therapy to introduce a third cone opsin into full-grown males. In this case, the animals were not immediately able to discriminate colors, but after about four months, the monkeys showed marked improvement, detecting previously indistinguishable colors in blue-green and red-violet hues.8 The delay corresponded to the timing of robust transgene expression, as if having the new visual pigment was all it took. “Marvelously, the monkeys gained what looks like full trichromatic vision,” says Jay Neitz, adding that the effects remained stable for a few years.

Both studies suggest the possibility that the primate brain was primed to accept the new stimulus offered by a third cone opsin—no major rewiring required. Such an ability may reflect our far-distant ancestors’ perception of even more colors than we see today. “If you go back to the vertebrate ancestor, they used to have five different kinds of pigments,” says Shozo Yokoyama, who studies vertebrate opsins at Emory University. Mammalian ancestors presumably lost some of these opsin genes along the evolutionary way, but their brains may have retained the capability to interpret the activity of additional opsins. (See “Animals’ Diverse Palettes” at bottom.)

But others in the field aren’t convinced that the animals were able to process new colors as soon as the retinal hardware was in place. Shortly after Nathans’s mouse study came out, Walter Makous, a vision science researcher at the University of Rochester, commented in Science that the mice might not be discriminating colors with their new opsin. Rather, they could have detected blotchiness in the color presented as a result of the human and mouse cones responding differently to the same stimulus, as if they were detecting different luminosities of the same color.9 It’s possible then that true color vision does not explain the mice’s improved performance. Instead, the brain “could see changes in color as essentially changes in texture,” says David Brainard, a vision scientist at the University of Pennsylvania. “Or it could be the brain is primed to figure this out and gives you color vision. I don’t think we know.”

A recent modeling study by Brainard’s group points to the latter scenario, at least for primates. The research team used computer learning to simulate human color vision and found that the configuration of the human retina—its particular mosaic of cone types, the ratio of long- to medium-wavelength cones, and the differences in their maximum absorption spectra—allows for such learning.10 These variables present the brain with enough information to determine the class (red, green, or blue) of each cone based on the wavelengths that each cell responds to.

“It was possible for the information-processing system to observe the signal of a cone and successfully assign labels for each cone class that are highly accurate,” Brainard says. Whether that was the case when trichromacy first appeared among primates is not certain, but Brainard’s simulation demonstrates that the modern human retina can make use of three opsins. Given the similarities among primate retinas, perhaps the first trichromatic monkey could do the same.

Colors with benefits

At whatever point primates were able to perceive additional color, the advantages would have spurred trichromacy’s quick spread through the population, researchers presume. It’s been thought that the monkeys would have been able to better distinguish between ripe and unripe fruit, for example, and to spot reddish young leaves among less protein-rich older ones, allowing them to forage more efficiently and improve the nutritional quality of their meals. (See photographs here.) But demonstrating the existence of such benefits has proven difficult.

PLOS ONE, doi:10.1371/journal.pone.0084872.g001, 2014. PHOTO CREDITS PRIMATES: F. CAMPUS; FRUITS: A. MELIN; PUMA: N. PARR

Amanda Melin of Washington University in St. Louis has spent years traveling to the forests of Costa Rica to observe capuchin monkeys in their natural habitat. Like squirrel monkeys, these New World primates have dichromatic males and either dichromatic or trichromatic females. Melin spends days at a time watching the animals forage, walking kilometers through the forest as the monkeys move from tree to tree, and collecting DNA from fecal samples, in an effort to determine which colors each animal can see. Much to her surprise, she’s found that fruit feeding rates between dichromats and trichromats are the same, initially suggesting that color vision doesn’t offer an advantage for foraging.11

Digging deeper into the data, however, Melin uncovered a subtler effect. “Where we see the difference is in accuracy,” she says. “Trichromats are making way fewer mistakes, but foraging at a more leisurely pace.” Dichromats, on the other hand, appear more frantic, touching, sniffing, and biting more fruits, including unripe or inedible ones.12 The question Melin is trying to answer now is whether that sloppier foraging behavior has any nutritional impact on the animals.

For people, of course, the importance of color vision is immeasurable, and Neitz is hopeful that the gene therapy he has used in monkeys could translate to a therapy for patients with color blindness. If successful, this would not only bring color acuity into the lives of those whose color experience is limited, it would allow researchers to explore how humans experience the leap from dichromacy to trichromacy.

But why stop there? Neitz wonders if it might also be possible to expand the range of normal human vision. This summer, he used gene therapy to give two monkeys a fourth cone opsin, such as birds have. Its absorption peak sits between that of the short- and medium-wavelength opsins, somewhere between violet and green. As this article goes to press, it’s too soon to know whether the supplemental gene has given the monkeys added sensitivity at that end of the rainbow, but if it works, the monkeys may soon be able to see 100 times more colors. It is enticing to think of how such a therapy could enrich human vision. Would a forest no longer look homogeneously green, but as diverse in hue as it is in individual trees? What beauty might we experience in art if we are able to perceive an order of magnitude more colors?

“Sometimes when I’m driving around I ask myself if the IRB [institutional review board] would let me do that to myself,” says Neitz. Why not try it out, he muses, to see if he can “cure” himself of trichromacy? Indeed, whether it happens by the hand of evolution or by human intervention, perhaps the story of human color vision’s progression is not over, Neitz says. “It’s a brave new world.” 


 ANIMALS' DIVERSE PALETTES | See full infographic: JPG | PDF

Most mammals, such as dogs, express just two types of opsins in the distal ends of their eyes’ cone cells, which are responsible for color vision. Humans and some primates have three. Other animals, including birds, fish, and insects, have even more opsins, although insects don’t have cones, but instead use other types of cells to detect color. Such diversity yields whole new worlds of color, with each opsin adding an order of magnitude more hues. Reconstructing the evolution of opsin genes, Shozo Yokoyama of Emory University and his colleagues have found that substitutions at only a couple dozen amino acid sites in opsin proteins account for this diversity of spectral tuning found among vertebrates.

© MALEXEUMLike most mammals, dogs (Canis familiaris) see in color, just far fewer colors than other animals. From a behavioral study of two Italian greyhounds and a toy poodle, researchers figured out their limited color discrimination is due to dichromatic color vision (Visual Neuroscience, 3:119-25, 1989).



© WIKTOR RZEZUCHOWSKChickens have four types of cone opsins (PNAS, 89:5932-36, 1992), and in some birds, the short-wavelength opsin is shifted to absorb in the ultraviolet. Bird cone cells also have an oil droplet that serves to filter or concentrate particular wavelengths of light.



© LEEKRIThe eye of the American chameleon (Anolis carolinensis) has no rods and uses multiple cone opsins to detect color. The peaks here show the maximum absorption of the photopigments reconstituted in vitro (Vision Research, 38:37-44, 1998).



© LOONGERIn its ocean habitat, the coelacanth (Latimeria chalumnae) receives only blue light. Correspondingly, its rod-enriched eyes absorb light in this range. The peak here represents the absorption maximum of the visual pigment in vitro (PNAS, 96:6279-84, 1999).



© LROCHKA_TLike many fish, goldfish can see in the ultraviolet, thanks to a shift in their short-wavelength opsin. Using their long-wavelength opsin (yellow peak), they can also see red, likely an adaptation to their shallow aquatic environment, in which red light is not filtered out (Genetics, 153:919-32, 1999).



DIDIER DESCIUENS/WIKIMEDIA COMMONSThe small white butterfly (Pieris rapae) expresses four types of opsins but has at least six types of photoreceptors (PLOS ONE, 5:e15015, 2010). Filters in the eye adjust the spectral sensitivity of the photoreceptor cells. In males, the violet receptor is modified into a second blue one (not shown).



© NANCY NEHRINGRepeated florets called ommatidia in the compound eye of the fruit fly (Drosophila melanogaster) are made of a central color-vision cell surrounded by six blue-light absorbing cells. Shown here are the absorption maximums for the opsins expressed in the central cell of the ommatidium (J Neurosci, 19:10716-26, 1999).


  1. M. Neitz et al., “Spectral tuning of pigments underlying red-green color vision,” Science, 252:971-74, 1991.
  2. J. Nathans, D.S. Hogness, “Isolation and nucleotide sequence of the gene encoding human rhodopsin,” PNAS, 81:4851-55, 1984.
  3. J. Nathans et al., “Molecular genetics of human color vision: The genes encoding blue, green, and red pigments,” Science, 232:193-202, 1986.
  4. G.H. Jacobs, J. Neitz, “Inheritance of color vision in a New World monkey (Saimiri sciureus),” PNAS, 84:2545-49, 1987.
  5. J. Nathans et al., “Molecular genetics of human blue cone monochromacy,” Science, 245:831-38, 1989.
  6. Y. Wang et al., “A locus control region adjacent to the human red and green visual pigment genes,” Neuron, 9:429-40, 1992.
  7. G.H. Jacobs et al., “Emergence of novel color vision in mice engineered to express a human cone photopigment,” Science, 315:1723-25, 2007.
  8. K. Mancuso et al., “Gene therapy for red-green colour blindness in adult primates,” Nature, 461:784-87, 2009.
  9. W. Makous, “Comment on ‘Emergence of novel color vision in mice engineered to express a human cone photopigment,’” Science, 318:196b, 2007.
  10. N.C. Benson et al., “Unsupervised learning of cone spectral classes from natural images,” PLOS Computational Biology, 10:e1003652, 2014.
  11. A.D. Melin et al., “Fig foraging by dichromatic and trichromatic Cebus capucinus in a tropical dry forest,” International Journal of Primatology, 30:753-75, 2009.
  12. A.D. Melin et al., “Food search through the eyes of a monkey: a functional substitution approach for assessing the ecology of primate color vision,” Vision Research, 86:87-96, 2013.

Interested in reading more?

Magaizne Cover

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!