Monkeys With Superpower Eyes Could Help Cure Color Blindness

In the video, a preposterously cute grey squirrel monkey named Dalton bonks his head against a computer screen in front of him. Wide-eyed and muttonchopped, Dalton has quite the set up—the screen, wide in squirrel-monkey terms, displays dots of varying sizes and colors. Below that is a monkey-sized basin, like a sink in a dollhouse kitchen remodeled with stainless steel fixtures.

Dalton is doing science. Male squirrel monkeys don’t see color well; they have a kind of red-green colorblindness. Dalton's eyes really only see medium and short wavelengths of light—blues and greens, and their overlap color, yellow. He's what vision scientists call a protanope. With no receptor for reddish hues, he sees reds as dark yellows and yellow-browns, and greens as mostly yellow—to the extent that human color words mean anything to a monkey.

He isn’t really bonking his head; Dalton is trained to indicate when he can see a color on the screen. “He’s actually fairly carefully touching his tongue to the screen,” says Jay Neitz, a color vision researcher at the University of Washington. Dalton sticks his tongue out, Jay says, because he knows that when he recognizes a color, a drop of grape juice will appear in the basin. Dalton really likes grape juice. And a little click will sound in the background, another bit of reinforcement. When he sees a color, he gives it a little kiss.

When Dalton can’t find a color, or kisses the wrong part of the screen, a less pleasant buzz replaces the click. Also: no grape juice. When that happens, Dalton sometimes takes a random guess. Or he just looks around the room, seemingly a little frantic.

“Is he angry?” I ask.

“It’s more like, ‘what the hell?’” Neitz says. “Sometimes they’ll grab ahold of the tray. It’s a frustration thing.” In clip after clip, shades of red scattered amid greys go unseen, unlicked. The buzzer buzzes, the grape juice does not appear. Dalton’s body assumes the posture of a primate who would very much like to speak to a manager, please.

Then there’s a discontinuity, a weeks-later time jump. Off screen—this was in 2009—Dalton undergoes a delicate operation. A surgeon inserts a long needle into Dalton’s eye, all the way to the receptor-dense, light-sensing retina at the back. With this microliter syringe, the surgeon injects a tiny bleb of fluid. “It creates a retinal detachment that looks like a blister,” Neitz says. The surgeon does this in three places, each 120 degrees from the others, in both eyes.

This is where Dalton apparently gets superpowers . In this origin story, Neitz and his wife Maureen, a geneticist, are the scientists who create the super-monkey serum.

In the fluid is a virus, specifically an “adenovirus,” a common variety of pathogen that includes the common cold. This one has been scraped clean of the things that make it germy, repurposed to carry a carefully designed stretch of DNA wrapped inside a ball of protein.

Viruses are good at hijacking a cell’s genetic machinery. Usually they do it to trick cells into making more virus; that’s called infection. Here, in Dalton’s eyeball, the modified adenovirus is carrying instructions to teach the cone-shaped cells in the monkey’s retinas that normally sense medium-wavelength, greenish light to instead (or maybe also) sense longer, reddish wavelengths.


Sign up for the Backchannel newsletter and never miss the best of WIRED.

A lot has to go right. The virus has to stick to the cell and evade the monkey’s immune system. It has to get the new gene into the cell’s nucleus and integrate into the existing DNA. The gene has to actually get turned on and start making proteins. It rarely goes right. “We’ve been working on ways to improve the efficiency,” Neitz says. At the highest viral titer, just 30 percent of cells infected actually turn on the gene. But those that do will go on to express not just one photopigment, but two. A once-middle-wavelength cone will also have a long-wavelength receptor. Nominally, it’ll see red, and Dalton will have a monkey superpower.

Now Neitz cues up a new set of videos. This is Dalton again, post operation. Red dots appear amid green on the screen. Dalton picks them right out. Lick. Click. Juice. Again: different red dots amid different colors. Lick, click, juice. Dalton is unstoppable. He gets almost all of them, one after the other.

A monkey with genetically engineered eyes isn’t even the most striking thing about the Neitzes' work here. It’s what Dalton says about the genetics of color vision. Most mammals are dichromats—they have only two photoreceptors for color. But some primates, including humans, are trichromatic. How and why that ability evolved remains something of a mystery, but being able to induce it in a dichromat like Dalton says a lot about color vision. It also says a lot about how the brain deals with color. Oh, and it might be a cure for red-green colorblindness in people, too.

Back in 1672 the Philosophical Transactions of the Royal Society —the first real scientific journal—published Isaac Newton’s breakthrough work showing that a prism could break the white light of the sun into the colors of what he called the spectrum. Slightly more than a century later, in 1777, that same journal published an account that cemented that discovery, a literal in-sight. Light might be made of color, but not everyone could see those colors. In this case, a Cumberland shoemaker named Harris couldn’t tell when objects were red. Neither could his brothers. And when one of those brothers saw a rainbow, he “could distinguish the different colours; meaning only, that it was composed of different colours, for he could not tell what they were.”

More investigators started coming forward with similar stories about color blindness, a deficit that turned out to be crucial to the scientific understanding of human color perception. In 1798, a decade before introducing the idea of atoms to chemistry, the English chemist (and squirrel monkey namesake) John Dalton revealed in the Memoirs of the Literary and Philosophical Society of Manchester that all those times when he had to ask his fellow amateur botanists what color a flower was, he wasn’t kidding. “Notwithstanding this, I was never convinced of a peculiarity in my vision, till I accidentally observed the colour of the flower of the Geranium zonale by candle-light, in the Autumn of 1792,” Dalton wrote. “The flower was pink, but it appeared to me almost an exact sky-blue by day; in candle-light, however, it was astonishingly changed, not having then any blue in it, but being what I called red.” Dalton’s brother saw the flower that way, too.

Becoming his own research subject, Dalton started experimenting. Most people, he realized, see six colors in the Newtonian spectrum. “I see only two or at most three distinctions,” Dalton wrote. Red, orange, yellow, and green were all “yellow” to him. Everything else was blue. Colors looked different under candle-light than daylight.

Dalton the man, like Dalton the monkey, was a protanope. Even today the condition is sometimes called Daltonism. (It affects about 1 percent of human men and a far smaller fraction of women; the more common form of red-green colorblindness, a mutation called "deuteranopia" in the cone that senses greenish, medium-wavelength light is a little more common—like 6 percent of men.) The implication of all those likewise-envisioned brothers was that Daltonism somehow ran in families—though Darwin’s Origin of Species and the concept of evolution were still six decades in the future.

This wasn’t just about how the eye worked. At the end of the 18th century, the problem of color attracted not just the intense interest of scientists studying the eye and brain but also philosophers concerned with the very nature of consciousness. The questions color raised exposed fractures in art and technology on the cusp of the industrial revolution.

Was color something that surfaces absorbed or reflected? Was light made of teeny-tiny particles (as Newton had said) or waves? And if it was waves, what medium did they travel through? As the art historian Melissa Katz has written , over the next 100 years, the palate of available industrial pigments would double—a long journey from paleolithic reds and yellows and classical blues and greens to the colorized, modern world, to chrome yellow, iodine scarlet, zinc white, cadmium yellow, manganese violet. And yet, even as a wild array of new colors proliferated, there was no single, agreed-upon theoretical construct explaining how people’s eyes drank them all in. In fact there were two.

In 1801 the genius polymath Thomas Young determined that not only was light composed of waves (traveling in a “luminiferous ether” that wasn’t actually a thing, but stick with him). Using calculations from Newton himself and a pretty good estimate for the speed of light, Young was able to quantify the wavelength of different colors. Red, he said, was “482 millions of millions of undulations in a second.” Green was 584 millions of millions. As the vision researcher John Mollon notes in Normal and Defective Colour Vision , if you convert these measurements to nanometers, they’re damn close to modern values.

Young didn’t stop there. Wavelength is in effect a continuous metric with an infinite number of possible colors; estimates for the number of colors humans see have ranged from 2 million to upward of a billion. But “as it is almost impossible to conceive each sensitive point of the retina to contain an infinite number of particles, each capable of vibrating in unison with every possible undulation, it becomes necessary to suppose the number limited, for instance, to the three principal colours, red, yellow, and blue,” Young said .

Whatever was going on in the eye, Young was saying, it was mixing sensations together to produce all those millions of colors from just three. He wasn’t the first person to suggest that the human eye had three receptors for different colors, but he was the one who made the most noise.

In 1852 Hermann Ludwig Ferdinand von Helmholtz published a paper on color vision that made the distinction between the mixing of pigments—stir red and green paint together and you get yellow (or, in real-world mixing, brown, but that's really just dark yellow anyway)—and the mixing of light. Pick the right colored lights, complements of one another, and mix them together, and you’d get white. Hemholtz was originally skeptical of Young’s approach, in part because when he mixed other colored lights together, the results were desaturated, like pastels.

By coming up with a way to quantify people’s perceptions of colors, James Clerk Maxwell—who’d go on to write the equations that still define electromagnetism—brought Helmholtz around. Maxwell realized that every perception of color affected, to a greater or lesser extent, all three of the sensations that Young had postulated. Nobody ever saw a real, hypersaturated spectral hue. But to prove that there could be colors outside a person’s fundamental ability to see them, Maxwell realized he’d need to quantify the color observations of people who didn’t have one of Young’s sensations. In other words, he needed people who were colorblind.

Maxwell developed a theoretical colorspace, a triangle with red, green, and blue at the vertices. He’d learned to generate an equation for any possible color in that space using spinning disks and colored paper; working with colorblind people he found that for any color, all the other colors they confused with it were on a line from that color to the red corner of the triangle. Hence, the person was missing the red receptor, or as Maxwell described him, a dichromat. The fundamental sensation of red was outside that person’s triangle.

Helmholtz bought it—hard. By the 1850s, everyone started calling the idea of three receptors for color in the eye the Young-Helmholtz Theory. Less than a decade later, scientists had largely agreed that the cone-shaped cells in the retina were where these perceptions resided.

I promised you two theories, though. An adherent of a more spiritualist, mystical German scientific tradition, Ewald Hering, a scholarship kid from Saxony, had been fighting with Helmholtz since Hering got his MD from the University of Leipzig in 1860. (I'm particularly indebted to R. Steven Turner's book In the Eye's Mind: Vision and the Helmholtz-Hering Controversy for much of this history.) Hering wondered: Why can a person imagine a green that is bluish or yellowish, but not reddish? Or a blue that is reddish or greenish, but not yellowish? If you decide that red, green, blue, and yellow are the four fundamental colors—urfarben, Hering called them, though today scientists say “unique hues”—then why can’t people see, or even conceive of, red-green and blue-yellow? The colors were opponent, or antagonistic. German has a word for it! They’re gegenfarben .

And the idea that just three sensors can account for all the colors? “One is then forced to describe yellow, for instance, as a red green or a green red, blue as a violet green or green violet,” Hering wrote in Outlines of a Theory of the Light Sense . No big deal in and of itself, but “such a way of assigning names to colors does not at all express in what way and to what extent the colors appear to be interrelated.” In other words, the three-colored triangular colorspace just isn’t how people see.

Hering re-drew the colorspace according to his opponency theory, a circle with red opposite green and blue opposite yellow. Where red overlaps blue, you get various purples; those oppose the chartreuses where yellow overlaps green. Red-green colorblindness was, to Hering, a lack of ability to perceive the red-green opponency axis. Trying to figure out how the more phenomenological, observational experience of color meshed with what people knew about how the eye worked (or sometimes didn’t work) consumed the second half of the 20th century’s vision research.

The Neitz lab takes up nearly two full floors of a building in Seattle’s South Lake Union neighborhood, which used to be known for houseboats and nightclubs with a casual attitude toward the age of their patrons. Today it’s home to a dizzying number of cleanly modern near-cubes—University of Washington research labs, Google’s Seattle HQ, more Amazon offices, and businesses with headscratchingly sci-fi sounding names like “NanoString.” A new Facebook HQ under construction next door to the Allen Institute. It feels like a lot of stuff in this part of town uses blockchains for something.

In small rooms off the lab’s various hallways, the advanced instruments of vision research spread out like stand-up games at a really difficult arcade. Neitz, a lean, smiling man in a Pacific Northwest standard-issue fleece pullover and nice track shoes, shows me one of the newest gadgets: a scanning electron microscope with an automated microtome inside, the science version of a deli meat slicer.

It takes an image and then slices off 50 nanometers of whatever it’s looking at, then takes another picture. A computer assembles the slices together into a whole structure, inside and out. And the whatever-it’s-looking-at right now is the retina of a monkey.

Onscreen, it resembles a mosaic by Joan Miró, all curving forms leaning into each other. Where the interfaces are darkest, Neitz says, are where neurons are talking to each other—literally where the neurotransmitter glutamate is flowing from one neuron across a synapse to another. “We can see what nobody else on Earth can see,” Neitz says, not without a little pride. “The places where the cells are communicating are very, very distinctive.”

Their vision in the lab is even better than that. “We’re actually able to distinguish the blue cones from the red and green,” Neitz says. (And the blue make up only about 6 percent of the total.) “We can trace where information about short-wavelength light is going throughout the retina.”

The Neitzes got married in 1981 and got their PhDs in 1986. Back when they were grad students, Maureen was working on genetics, and Jay was doing neuroscience. He realized that color and colorblindness were a kind of natural experiment on consciousness; Maureen thought molecular genetics might help figure it out. When they got their PhDs, they decided to work on it together. His office is downstairs and hers is upstairs. “Our ability to work together has—” he begins.

“—improved,” she finishes.

The primate eye has four light-sensing photopigments in the retina, at the back. There's rhodopsin, found not in the "cones" that perceive color but other cells called rods, used in low-light conditions. In the cones are three other types—one sensitive to longer wavelengths for reds, one to the middle-wavelength one for greens, and one to the short-wavelength for blues. It's the Young-Helmholtz theory made flesh.

But where the theory is willing, that flesh is weak. So-called Old World primates—macaques, baboons, people—generally have trichromatic vision. New World primates like squirrel monkeys, though, are weird. Some squirrel monkeys are trichromats and some are dichromats—but not all the trichromats see colors the same. Some are “anomalous trichromats,” a common form of colorblindness.

It's the photopigments' fault, somehow. Early on, Neitz's mentor Gerald Jacobs learned that squirrel monkeys have five . There's short, medium, and long, but also one halfway between red and green. Some have only the red and the halfway pigments. They're deuteranomalous. Some have the green and the halfway pigments. They're protanomalous. And some protanopes have only the green. The males are all dichromats. Some of the females are trichromats.

So, like, what the heck, right? Chromosomally typical mammals have two so-called sex chromosomes. Males have an X and a Y; females have an X and an X. So the inheritance pattern is tricky. Male offspring get their father’s Y and one of their mother’s Xes. Female offspring get their father’s X and, again, just one of their mother’s Xes. In males, both those chromosomes make proteins, but in females, one of the Xes in each cell gets silenced—it’s called “X inactivation.”

A tricky bit: The gene for the bluish, short-wavelength photopigment isn’t on a sex chromosome. But the other two are.

In the monkeys, the photopigments come from just one gene on the X chromosome—except that gene has three possible forms, what geneticists call alleles. So all the males were actually three kinds of dichromats.

The females, then, had six different varieties—three on each X. “If you’re a female and you get a red on one X and a green on the other X, then in approximately half the cones you’ll get the X with the red and the other half will get the X with the green. Voila, the female will have those two cones,” in addition to their blue ones, Neitz says. But the males only have their blue cones, “and either just red, just green, or just in between.”

The Neitzes went to work on the basic mechanism for how those photoreceptors functions, and it’s a doozy. A photopigment has two main parts—the “opsin” is a relatively giant protein that threads through the membrane of the retinal cell like a pile of overcooked rotini pasta. At the heart of that complex, no matter what color it senses, is a wee little molecule called a chromophore. This particular one, 11- cis -retinal, is a chain of carbons with a dogleg bend at one double-bond along its backbone.

As a class, this system is called a G-coupled protein receptor—a sensor on the outside of cell connected to machinery on the inside. Some trigger hits the sensor and starts a cascade of action that activates a “g protein” in the cell, setting off some rubegoldbergian chain of biological activity like the release of a neurotransmitter.

In the rods and cones, that trigger is light—a subatomic particle called a photon. Remember, wavelength is only one way to think about light; “color” also roughly corresponds to the amount of energy in a photon. If I tell you about light with a wavelength of, say, 540 nm—that’s a slightly yellowish green—I’m also saying that it is a photon with an energy of 222 kilojoules per mole. Same thing.

When a photon hits the chromophore’s dogleg double bond, one of those bonds breaks. Half of the molecule rotates. The chromophore literally changes shape, straightening out. “And that straightening-out pushes on the opsin it’s sitting inside,” says Greg Horwitz, a neuroscientist at the University of Washington. “When it straightens out, it pushes against the opsin protein a little bit, and the protein changes shape, so that now it can interact with the g-protein.” The rubegoldberg mechanism starts to crank; there’s a change in voltage across the cell’s membrane, and that’s the first step of vision.

Damned if it’s not even crazier than that. The chromophore can’t just straighten back out on its own. It gets transported out of the cone to another cell, bent back into shape like a piece of rebar that a backhoe ran over, and shuttled back into the cones. “It’s a detail, but it’s just so weird,” Horwitz says.

The point is, deep inside this clump of amino acid chains, light converts to mechanical motion, which converts to a neuroelectrical signal, which in the brain becomes a conscious apprehension of color. The specific amino acids in the opsin determine what wavelength of light the chromophore will respond to.

Higher wavelength colors, toward red, have lower energy. “The protein puts pressure on that double bond, twisting it,” says Neitz. “If it’s twisting in the right direction, it takes less energy from a photon to break it and remake it.” More mechanical energy from the protein means it takes less energy from the light to trigger the response—a longer wavelength. And vice versa.

As soon as the cells at the back of the eye transduce the photons that came streaming in through the iris, there’s no “color” anymore. Once the chromophore has absorbed its photon and the opsin around it has determined whether it’ll sproing straight or not, the actual wavelength doesn’t matter anymore.

Technically, that’s called “univariance,” and it means that different things can set off the same cone. The wavelength of light might be closer to the peak sensitivity, or there might be more light overall, or both. So the simple pinging of a cone doesn’t signify a color. How could it? If the system was that direct, you’d only be able to see red, green, blue, or overall-bright— but not yellow . There’s no single receptor for that.

The overall complex only understands how many times that happens—how many photons the system has absorbed. The out-there world of wavelengths and photons, of pigments that reflect and absorb light? It all turns to fiction as soon as it reaches the back of your eye. It’s a story. But inside the brain, stories are all we have.

Here’s where I take a beat. “It’s not how I would design it,” I finally say.

“I’m used to that now,” Neitz answers. “Nothing is like we would design it.”

But that’s not exactly true, because the Neitzes did take a crack at designing their own opsin-chromophore complex.

Maureen grew up in Sunnyvale, just up the road from Palo Alto and the Stanford lab where Jeremy Nathans worked. They’d met in grad school; Nathans had told them he was working on sequencing the genes responsible for making those opsins. So on a visit back to see Maureen’s family, the Neitzes stopped in to see Nathans. And he’d done it. The team had the genes for the L, M, and S opsins—Nathans’ genes, actually. They’d sequenced his.

The Neitzes asked Nathans if they could borrow a gene.

Scientists can be as contentious and competitive as any other human. But Nathans went over to a freezer and took out a sample, pipetted a bit into a tube, capped it, and gave it to Maureen. She put it in her purse.

Maureen was able to tweak a single nucleotide, one of the A-C-G-T code letters that make up the verbiage of DNA, to alter the peak absorbance of the photopigment it coded for. Cones don’t sense specific colors, anyway—their sensitivities are actually bell-shaped curves. The short-wavelength S-opsin tops out at 420 nm, a deep violet, but overlaps with the other two a bit in the high 400s. M- and L-opsin peak at 530 nm and 560 nm, respectively, and overlap almost completely—no surprise, really, since L-opsin is a recent mutation. The genes for L and M are in fact 98 percent identical.

“Little by little, we began to work out the genetics of colorblindness in humans,” Jay Neitz says. “The most common thing that happens is that humans lose one gene on the X chromosome. We say it’s basically a ‘back mutation.’ Humans turn back into a squirrel monkey.”

It was full circle. The Neitzes had a molecular explanation for phenomena that vision scientists had been arguing about since the 1700s. “In ophthalmology, especially in the early days,” Neitz says, “we always said that once we figure out what causes something, we’ll be able to cure it.”

So with all that in mind: Could what the Neitzes did to Dalton work in a human eye. Could you cure a red-green colorblind human male?

It’s not an easy question. Gene therapy is tricky; most things that go right and wrong in people’s bodies and minds don’t come down to just one gene, and even when they do, it’s not always obvious how to tweak that gene to fix the problem, or what the knock-on effects of that tweak will be.

Also, every cell has every gene, but not every gene gets turned on in every cell. The cells that make muscle don’t do that in the brain; the cells that make bone aren’t supposed to do that in muscle. You have to target the right cells and convince them to change—transduction, the process of getting cells to express the new protein you want, is rare. Here the Neitzes had an advantage. “To cure colorblindness, you only want to transduce a random subset of cones,” Neitz says. “You have a billion copies of the virus back where the photoreceptors are. So even though transduction is poor, you can get enough cells transduced.”

For now, the Neitzes are working on a modification of the treatment. Instead of having to cause a retinal detachment, they'd like to be able to get the right amount of transduction with an injection into the vitreous, the clear jelly that fills the eyeball. So they've made the right modifications to the vector and the procedure, and they've injected monkeys. "We don't know the results yet," Neitz says.

They're also working with macaques, Old World trichromatic primates like you and me. They're trying to give those monkeys a fourth photopigment, one derived from an opsin in some gerbils that's tuned to a peak sensitivity between our bluish and greenish. It'd give them more even coverage of the visible spectrum, and tetrachromacy. Another superpower. "We always test things in mice first," Neitz says, "and last weekend I injected the latest version into mice eyes. It takes a couple months."

Cones are necessary to color vision, but not sufficient. The neural wiring between them, and from the eye through various weigh stations to the visual cortex, all play a role in creating what more romantic scientists used to call “ the color sense .”

So, sure, Dalton the monkey is behaviorally trichromatic. He acts like he sees colors the way you (if you’re color-normal) and I (I’m color-normal) do. But is Dalton actually trichromatic? For full-blown trichromacy, you’d have to have the right wiring between the cells of the retina and in the brain. Neitz argues that his monkeys do, that it was already laid down and waiting to receive the new input from his engineering. After all, some of female squirrel monkeys are naturally trichromatic.

Other researchers aren’t so sure. “Showing trichromacy is not a mystery. James Clerk Maxwell figured out how to do that 150 years ago. You have to have the animal make trichromatic matches,” says Qasim Zaidi, a neuroscientist at SUNY College of Optometry. “They’ve had 10 years to do this, and they haven’t done it.”

Other neural architectures could give the same results as Neitz has seen. The retina is a complex layering of ganglion cells, bipolar cells, amacrine cells, and the horizontal cells that wire the rods and cones together; clusters of cells make "receptive fields" that feed information to neurons up the line, toward the brain. The perceptions of light and color are actually combinations of responses and signals from all those things, an arithmetic of negative response from some and positive response from others. It's way more labyrinthine than just red light pinging a red photoreceptor so you see red. “You can get the whole population of responses just by nonselectively firing a ganglion cell with every photoreceptor in its receptive field,” Zaidi says.

But Neitz says their behavioral tests assure him it's more than that. The monkeys couldn't just be seeing a difference in brightness—the color version of loudness, if you will—because "we very carefully varied the intensity of the red or green relative to grey, to make sure no matter what level of brightness used, they could tell it apart," Neitz says. "Dichromats can't."

People argue about the evolutionary value of color overall, but it must have one, or we wouldn’t be able to see it. So how’d we animals get it? To evolve trichromatic vision, you need a brain that can process input from a retinal mosaic of three photoreceptors. But to get any selective advantage from having that brain, you need three photoreceptors and all their neural wiring.

The monkeys have all three photoreceptors. Their cones respond to red light. That’s a threshold, yes. But, I say, trying for delicacy, did the Neitzes every try just dissecting one of their monkeys' eyes and looking at the cones?

“We try never to kill monkeys. We’re kind of opposed to doing that,” Neitz says. “I developed an electrophysiological technique where we could anesthetize the animal, put an electrode on its eye, shine different lights, and figure out what photopigments they have.” Dalton, it turns out, died a few years back—adult-onset diabetes—and Neitz says his illness meant a necropsy wouldn't have told them anything. (That aversion to violence, Neitz says, is also what has led them to turn down invitations from the military to think about potential applications of retinal gene therapy beyond red-green colorblindness—infrared night vision, the ability to see an otherwise invisible difference in the uniforms of allies and enemies, something. Neitz won’t be specific.)

But it’s possible that Dalton could see some kind of difference, but not one that we humans would describe as a color. Or rather, they could now see that something had a different color when, prior to gene therapy, they wouldn’t have been able to. “The Neitz research shows (pretty clearly in my view) that inducing the expression of a second M/L opsin gene in such dichromatic male monkeys allows them to gain discrimination abilities that appear trichromatic,” Jay’s old colleague Gerald Jacobs says in an email.

That success, he says, “suggests pretty strongly to me that one or another of the means for adding a viable photopigment to the retinal array would very likely produce the same change in a human. How their perceptions of color, as opposed to their discrimination abilities, might change is another interesting question.”

Take the mantis shrimp. In addition to a killer right cross, mantis shrimp have 12 photoreceptors with narrow peak sensitivities and minimal overlap, covering a range from ultraviolet to borderline infrared. But no one thinks they’re dodecachromatic. “They don’t compare across photoreceptors, so they have no color discrimination,” Zaidi says. “They have enormous speed. Because they’re not doing any kind of combination, it’s a direct line to behavior. But that tells you having 12 photopigments doesn’t give you 12 kinds of color vision.” And therefore simply adding a third to spider monkeys doesn’t give them three. (In fact, Zaidi and Conway have argued that cells in the inferior temporal cortex of the macaque monkey’s brain have very similar colors as the mantis shrimp’s eyes sense.)

It’s enough to suggest that the science has more questions to answer. So, too, might the ethics. Four or five phase I and phase II trials of gene therapy for human color vision deficiencies seem to be underway right now, but they’re all for a much more severe form, achromatopsia—the total lack of functioning cone receptors. People with the disorder don’t see color, and also see less detail and are painfully sensitive to light. Researchers have also had more success curing it in other mammals, including dogs and sheep. So the disease is more debilitating than red-green colorblindness, and the research track record is better.

Conway says the Neitzes work on the molecular genetics of the photoreceptors is good, and that Jay is a “wonderful, out-of-the-box thinker,” but molecular genetics has limitations. “It’s an ethical grey zone with their experiments to try to fix colorblindness,” he says. “Their gene therapy work is really compelling and interesting from a basic science point of view, but when you say ‘I’m going to fix colorblindness,’ I think you kind of need to know what you’re doing.”

The next step, experimentally, is getting the treatment to work, and show it doesn't have side effects. The approach to the Food and Drug Administration might involve treating another kind of colorblindness, blue-cone monochromats missing both the red and green cones. It's a more serious impairment, of both color and acuity.

Neitz has continued trials, even though he and Maureen haven’t published any new data on animal subjects. “We’ve tried it five times before and it didn’t work. One of these times, it’s going to be it,” he says. “We are trying to get something that can go forward to humans. That’s what being a scientist is—all the times it doesn’t work.”

Now the Neitzes are getting ready to try a further experiment, inserting a “true-blue” photopigment gene from a Mongolian gerbil to give his monkeys a fourth photopigment. It’d cover a troublesome gap in mammalian trichromacy, a dip in sensitivity over the blue-greens between S-opsin and M-opsin to which evolution turned a blind eye.

Other animals see it, though. “We isolated the gene from the gerbil, but for some reason we couldn’t get the complete sequence. We substituted the exact same segment from a human, so it’s actually a chimera,” Neitz says. “You know that scene in Jurassic Park where the guy says, ‘you substituted frog DNA?’ Well, I guess that’s what we shouldn’t have done, because the structure of the genome sequence from the human is incompatible with the gerbil.”

I suggest that people talking about genetically souped-up animals maybe shouldn’t reference Jurassic Park , and Neitz laughs. “Well, we have trained monkeys waiting to do this experiment. And we’ve developed a television that has four different colors.” Gadget-heads: He’s talking about a monitor that replaces the three RGB pixels with four—RGBV, he says, for violet. Take that, Ultra High Def.

It'd be cool to watch a show on that monitor. After all, if you think the quantitative assessment of whether a monkey has acquired metamonkey vision is hard, imagine trying to ask what colors the monkey actually sees . Scientists barely know how to do that with other humans, and humans can speak English. That's what you learn when you study people's color vision. "It's like, 'oh, you aren't seeing the same things as I am. You're just using the same words,'" Maureen Neitz says.

So I ask, tested abilities aside, do the post-gene therapy monkeys act different? Does seeing colors change how they are ?

One of the subjects, Sam, now more than 20 years old, is still around. “I talk to him quite frequently,” Neitz says. “Sam’s still in pretty good shape.”


“The thing about monkeys is, they’re pretty nonchalant about everything,” he says. “Like, ‘oh, yes, I see color now.’”

There is one thing, Neitz says. Outside his office is an old coin-op gum-ball machine, the kind with a glass sphere and a dispenser with a crank. This one's filled with loose M&Ms; I had noticed it on the way in because it contained only red, blue, and just a few green, and I’d asked if that reflected the distribution of cones in the retina. Alas, it was just that green ones were more expensive on Amazon.

The thing is, a human patient of Neitz’s had told him that the red and green ones were indistinguishable to him. And Neitz, well, he’s not supposed to give sweets to the monkeys—no Coca-Cola, even though they love it, and no candy. But he’s known Sam for a long time, and so he often brings in a treat.

Before the surgery, Sam gobbled up M&Ms indiscriminately. Since then? “Sam now has this huge preference for green M&Ms. And he likes green beans,” Neitz says. “I don’t know what that is, but that’s my one story about that.”

Updated 2-10-19 12:15 PT Pigments mix subtractively and darken; clarifed that yellow's dark correspondent is brown.

  • All the times Facebook moved fast (and broke things)
  • What it takes to pull off the country's first online census
  • How to make your home more energy-efficient
  • The world might actually run out of people
  • Finding Lena, the patron saint of JPEGs
  • 👀 Looking for the latest gadgets? Check out our latest buying guides and best deals all year round
  • 📩 Get even more of our inside scoops with our weekly Backchannel newsletter
Read also:   Meet the Rosehip Cell, a New Kind of Human Neuron