Why Animals See Colors You Can't: The Evolution and Science of Color Vision

Here is the single most useful thing you can learn about color, and almost no one teaches it: color is not in the world. It's in you. A red apple is not red the way it is round or heavy. Its "redness" is a sensation your nervous system constructs — a signal manufactured in a few millimeters of retina and cortex — in response to a narrow slice of electromagnetic radiation the apple happens to reflect. Change the nervous system, and you change the color, without changing the apple at all.

This sounds like philosophy, but it is hard, measurable biology with enormous practical consequences. A dog looks at the same red apple and does not see red. A honeybee, landing on a flower you call yellow, is responding to an ultraviolet bullseye you will never once see in your life. A bird's mate is performing courtship displays in colors that, to you, simply do not exist. And the animal once celebrated as having the greatest color vision on Earth — the mantis shrimp, with twelve kinds of photoreceptor — was shown in a landmark 2014 experiment to distinguish colors worse than you do.

This article is a tour of how different eyes build different worlds. It covers the biology of cones and opsins, why most mammals are "colorblind" by your standards, the fruit hypothesis for why primates — alone among mammals — recovered rich color vision, how birds and insects see into the ultraviolet, the debunking of the mantis shrimp myth, the murky truth about "tetrachromatic" humans, and what all of it means for anyone who designs with color.

1. The Hardware: How an Eye Sees Color

All color vision, from a human to a mantis shrimp, runs on the same basic piece of machinery: the photoreceptor. These are specialized cells in the retina whose job is to convert light into a neural signal. There are two great families. Rods are the low-light specialists — exquisitely sensitive, able to fire on a single photon, but utterly blind to color. Cones are the color detectors — less sensitive, needing more light, but each tuned to a particular band of wavelengths.

The "tuning" comes from a light-absorbing protein called an opsin. Each cone expresses a particular opsin that is maximally sensitive to a particular wavelength — its peak spectral sensitivity — and falls off to either side. A cone doesn't tell the brain "this light is 530 nanometers." It just says "this light excites me a lot," "a little," or "not at all." Color vision, then, is not a matter of measuring wavelengths directly. It's a matter of comparing the outputs of cones with different tunings. The brain does arithmetic on the differences, and the result of that arithmetic is what you experience as color.

This is the crucial insight, and it explains almost everything that follows: color vision is comparison. The number of colors an animal can distinguish is not set by how many cone types it has, but by how many independent dimensions of comparison those cones support. Two cone types give one axis of comparison. Three give two axes (plus brightness). Four give three axes. Each added cone type adds a whole new dimension to the animal's color space — which is why a bird, with four cone types, lives in a color space a full dimension richer than yours, and why the raw count of photoreceptors turns out, as we'll see, to be a misleading way to measure vision.

There's a second layer to the machinery, too. In vertebrates, the cone signals are not just added up — they're fed into opponent channels: red-versus-green, blue-versus-yellow, and black-versus-white. This opponent processing, first proposed by the physiologist Ewald Hering in the late 1800s, is why certain color experiences are physically impossible for you. You can imagine a reddish-yellow (orange) or a greenish-blue (teal), but you cannot imagine a reddish-green or a bluish-yellow. Those pairings are wired as opposites; the brain can be pushed toward one or the other, never both at once. Color vision is not a rainbow you read — it's a set of opposing dials.

2. Human Vision: Three Cones and an Opponent Brain

You are a trichromat: three cone types, three opsin genes, three spectral peaks. The cones are conventionally named by their peak sensitivity:

• S cones ("short" wavelength) — peak around 420 nm, the blue-violet end of the spectrum.
• M cones ("medium" wavelength) — peak around 530 nm, the green.
• L cones ("long" wavelength) — peak around 560 nm, the yellow-green-to-red.

Notice something odd immediately: your L and M cones are close together, peaking only about 30 nm apart in the green-yellow region, while the S cone sits far off in the blue. This lopsided arrangement is not an accident of design — it's a fossil of your evolutionary history, and it has real consequences. It means your color discrimination is finest in the yellow-green region and comparatively coarse in the deep reds and violets. It's also the reason the classic "visible spectrum" is not spread evenly: most of the wavelengths you can see are crammed into the green-yellow band where your two main cones overlap.

From three cones and opponent processing, the human visual system can distinguish roughly one million distinct colors — a number often cited, and roughly right, though it depends heavily on conditions. That sounds enormous until you remember that the "visible spectrum" itself is only a sliver of the electromagnetic spectrum: from about 380 nm (violet) to about 750 nm (deep red). Beyond violet lies the ultraviolet. Beyond red lies the infrared. Both are real radiation, both carry information, and both are completely invisible to you — not because they're "invisible" in any absolute sense, but because your three opsins simply don't respond to them. Other animals' opsins do.

3. Why Most Mammals Are "Colorblind" by Your Standards

Here is a fact that surprises most people: the vast majority of mammals are dichromats — they have only two cone types, and therefore see the world roughly the way a red-green colorblind human does. Dogs, cats, horses, cows, deer, rats, and most other mammals live in a two-dimensional color space of blues and yellows. A bright red ball on green grass is, to a dog, a muted, low-contrast smudge of yellowish-brown on yellowish-green. The dog is not "blind" in any meaningful sense — its world is perfectly coherent, and its rod-rich retina gives it excellent night vision — but by human standards it is missing an entire axis of color.

Why? The leading explanation is the nocturnal bottleneck hypothesis. The story goes like this: the earliest mammals were small, nocturnal, shrew-like creatures that spent the age of the dinosaurs hiding in the dark. In that niche, color vision was useless — there's no color in near-darkness — but sensitivity was everything. So natural selection favored a retina stuffed with rods, and the two extra cone types our fish and reptile ancestors once had were simply lost. When the dinosaurs fell and mammals emerged into the daylight, most of them kept the pared-down, two-cone retina they had inherited. Color vision, once discarded, is not easily re-evolved — the lost opsin genes are gone.

This is why your dog sees a sunset you will never see in the same way, and why a red toy is one of the worst possible choices for playing fetch in green grass. It is also the backdrop against which the primates' story becomes so extraordinary — because primates did something almost no other mammal did. They rebuilt the missing dimension.

4. Primates and the Fruit Hypothesis

Among mammals, the great exception is the primates — and specifically the Old World primates: apes, Old World monkeys, and humans. These animals are routine trichromats, and their three-cone system arose in a way that turns out to be both elegant and revealing.

The genetic origin is now well understood. The two long-wavelength opsin genes — the M and L genes that tune your green and red cones — sit side by side on the X chromosome, and they are clearly the product of a gene duplication. Somewhere in the primate lineage, a single ancestral long-wavelength opsin gene was copied, and the two copies then drifted apart in spectral tuning — one shifting toward green, one toward red — until a single animal carried two slightly different "long" opsins plus the existing "short" (blue) one. Voilà: three cone types, a new dimension of color, and the ability to distinguish red from green.

That X-chromosome placement has a consequence that echoes all the way into modern design: because males have only one X chromosome while females have two, red-green colorblindness is far more common in men — a topic we'll return to. But the placement also created an elegant natural experiment in the New World monkeys. In many New World species, the population is polymorphic: some individuals carry two versions of the long opsin gene and are trichromatic, while others carry one and are dichromatic. In several species, only the females — who can inherit two different X chromosomes — get the full trichromatic upgrade, a phenomenon called heterozygote advantage.

Which raises the question that has occupied evolutionary biologists for decades: why did primates re-evolve trichromacy at all? The most famous answer is the fruit hypothesis — the idea, reviewed in a 2003 paper by Surridge, Osorio, and Mundy in Trends in Ecology & Evolution, that a new red-green axis evolved to help primates spot ripe fruit and tender young leaves against a background of green foliage. Ripe fruit is often red, orange, or yellow — colors that pop against green only if you can distinguish red from green. A dichromatic primate sees the fruit disappear into the leaves; a trichromatic one sees a vivid beacon. The hypothesis is not without rivals — some argue the target was young, nutrient-rich, reddish leaves rather than fruit — but the core logic is the same: trichromacy evolved as a food-finding adaptation, a way of extracting a signal (ripe or tender, versus mature and green) from a noisy green background.

That, in a sentence, is why you can see red. Your color vision is, quite literally, a fruit detector — a biological adaptation built to solve a foraging problem millions of years ago, now repurposed to judge a brand logo or a traffic light.

5. Birds: Tetrachromats in Ultraviolet

If humans are trichromats and most mammals are dichromats, birds are something richer: the majority of birds are tetrachromats, with four cone types. Their fourth cone reaches into a region your eye cannot touch at all — the ultraviolet.

Birds don't just have more cones; they have a refinement your retina lacks entirely. Many bird cones contain tiny colored oil droplets — microscopic spheres of pigment that sit in front of each cone and act as filters, narrowing and sharpening the wavelengths the cone responds to. Where your cones have broad, overlapping sensitivity curves that the brain must untangle, a bird's cones come pre-filtered, which lets the bird make finer spectral distinctions with less neural effort. The result is a color system that is not only higher-dimensional than yours but, in some ways, cleaner.

The consequences are everywhere in bird life, once you know to look. Many species that look plain to human eyes — a Eurasian blue tit, for example, or a male budgerigar — turn out to be covered in ultraviolet patterns invisible to you, patterns that matter in mate choice. Researchers have shown, for instance, that female birds in many species prefer males whose UV reflectance is brighter and more saturated — a sexual signal that unfolds entirely in a channel you cannot perceive. Birds use UV for foraging too: some raptors, like kestrels, can spot the UV-reflecting urine trails of rodents from the air. The "visible world" of a bird simply contains a color axis you have never seen and never will.

There's a beautiful design implication buried here. A bird lives in a four-dimensional color space — one dimension richer than yours. When a designer works in an RGB color space, they are working inside the constraints of the human eye, a three-dimensional box. The bird's world reminds us that every color space is a choice about whose eyes you're designing for — and that "color" as you know it is only one coordinate system among many.

6. Bees, Butterflies, and the Flowers That Aren't For You

Flowers are among the most vivid things in the natural world — but here's the twist: flowers are not colored for you. They are colored for their pollinators, and most pollinators see a spectrum shifted from yours.

Take the honeybee. The bee is technically a trichromat, like you, but its three photoreceptors are tuned to ultraviolet, blue, and green — not blue, green, and red. The bee's long-wavelength receptor stops well short of your red cone; as a result, a bee cannot see red at all. A red poppy, to a bee, is essentially black. (Which is exactly why truly red flowers tend to be pollinated by birds, which can see red, rather than by bees.) In exchange for losing red, the bee gains something you lack: an entire channel in the ultraviolet.

And flowers exploit this relentlessly. Many flowers that look uniformly yellow or white to you carry, in the ultraviolet, dark patterns called nectar guides — bulls-eyes, streaks, and landing strips that point the bee toward the nectar and pollen, like runway markings. The flower is broadcasting a message in a frequency band you cannot receive. You and the bee are looking at the same petal and seeing fundamentally different images.

Insects push the hardware even further. Some butterflies carry five or more distinct photoreceptor types. The common bluebottle butterfly (Graphium sarpedon) is famous for having at least 15 classes of photoreceptor — one of the richest known in any animal — though, as with the mantis shrimp, the relationship between photoreceptor count and actual color perception is far from simple. The lesson from the insect world is the same one the next section drives home: more photoreceptors does not automatically mean better color vision.

7. The Mantis Shrimp Myth, Debunked

For years, the mantis shrimp — a fierce, hammer-clawed crustacean of tropical reefs — enjoyed a reputation as the animal with the greatest color vision on Earth. The reason was simple: its compound eyes contain 12 (and in some species up to 16) distinct types of photoreceptor, each sampling a narrow band of the spectrum from deep ultraviolet to far red (roughly 300 to 720 nm). By the naive logic — more receptors equals more colors — the mantis shrimp ought to see a Technicolor universe that would make your million colors look drab. Popular science ran with it.

Then, in 2014, Hanne Thoen and colleagues at the University of Queensland put that assumption to the test — and broke it. In a paper titled A Different Form of Color Vision in Mantis Shrimp, published in Science, they trained mantis shrimp to discriminate colors in a two-choice task and measured how finely the animals could actually tell one wavelength from another. The result was a shock: the mantis shrimp's color discrimination was poor — worse than a human's. Animals that could have sampled the spectrum in a dozen narrow slices turned out to be fairly bad at telling nearby colors apart.

What was going on? The answer reveals how profoundly the mantis shrimp's strategy differs from yours. Your brain takes three broad, overlapping cone signals and blends them, doing elaborate opponent arithmetic to extract fine color distinctions. The mantis shrimp appears to do almost no blending at all. Instead, each of its narrow-band photoreceptors acts as a kind of preset detector that simply flags "this wavelength is present," and the shrimp compares them almost one-by-one — a strategy Nature's news team summed up as relying on "a simple, efficient and previously unknown mechanism that operates at the level of individual photoreceptors." The shrimp doesn't synthesize a rich color image from its many channels; it scans with a dozen preset filters. It's a fundamentally different — and in the narrow task of color discrimination, less powerful — kind of vision.

The mantis shrimp's eyes are still extraordinary, just not in the way the myth claimed. Its true superpowers lie elsewhere: it has some of the only eyes known to detect circularly polarized light, and its many photoreceptors may serve speed, target detection, and a kind of "color coding by recognition" rather than fine discrimination. The deeper lesson for anyone who works with color is humbling and useful: raw channel count is not color vision. The number of sensors matters less than what the system does with them. A three-cone eye that blends and compares can outperform a twelve-cone eye that merely sorts.

8. Tetrachromatic Humans: Facts and Myths

Given that birds are tetrachromats, a natural question arises: could some humans be tetrachromats too? The answer is a fascinating "maybe" — and a cautionary tale about how science gets distorted as it travels.

The biology is real. Because the L and M opsin genes sit on the X chromosome, and because those genes are polymorphic — they come in several slightly different versions with slightly different spectral peaks — it is possible for a woman to carry two different versions of the L gene, one on each X chromosome. If both are expressed, she would have four distinct cone types: one S, one M, and two slightly different L cones. That is the definition of tetrachromacy. Estimates suggest something like 12% of women may carry the genetic potential for it — a figure often quoted, though it should be treated as an upper bound.

But — and this is where the myths creep in — carrying four opsin genes is not the same as seeing in four dimensions. The definitive review, by Gabriele Jordan and John Mollon in a 2019 paper titled Tetrachromacy: the mysterious case of extra-ordinary color vision, is careful on exactly this point. Rigorous functional studies of human tetrachromacy, they note, are "sparse," and most of what the public believes — women who see "a hundred million colors," or who can perceive colors the rest of us cannot — vastly outruns the evidence. In fact, it has proven surprisingly difficult to demonstrate that a four-cone woman can actually make color discriminations a trichromat cannot.

Why so hard? Partly because the two L cones in question are tuned so close together (often only a few nanometers apart) that the extra dimension they would add is extremely narrow — far less dramatic than the difference between, say, a bird's fourth cone and its other three. Partly because the brain, having been wired for three channels, may not spontaneously build a fourth axis of comparison just because a fourth cone shows up. And partly because our everyday world — screens, paints, pigments — is engineered for trichromats, so a tetrachromat would have almost nothing to exercise the extra channel on even if she had one.

The honest summary: human tetrachromacy is genetically possible, plausibly present in a minority of women, and still not clearly demonstrated to produce a functional fourth dimension of color experience. It's a real phenomenon wrapped in a thick layer of speculation — a useful reminder that in color science, as in design, "possible" and "proven" are not the same thing.

9. What the Colorblind Actually See

The same X-chromosome genetics that makes tetrachromacy possible also produces the most common difference in human color vision: red-green color blindness, formally called deuteranomaly and protanomaly (and the rarer, more severe deuteranopia and protanopia).

The numbers are striking and often underestimated. Roughly 8% of men — about 1 in 12 — have some form of red-green color vision deficiency, versus only about 0.5% of women. The asymmetry is a direct consequence of genetics: the L and M opsin genes live on the X chromosome, so a man, with his single X, needs only one faulty copy to be affected, while a woman, with two X chromosomes, needs both. The most common form, deuteranomaly, doesn't remove the green cone entirely — it just shifts its sensitivity closer to the red cone, which compresses the red-green axis and makes those colors harder to tell apart.

It's important to understand what colorblindness is not. It is not seeing the world in black and white — that condition, achromatopsia, is rare. Most colorblind people see a perfectly rich world; they simply have a narrower red-green axis. A red-green colorblind person can often see a red and a green side by side and describe them as "different," but the difference is subtler, and specific pairs — red on green, red on brown, green on brown, pink on gray — collapse into near-invisibility. For a designer, this is the concrete failure mode: not "everything looks gray," but specific hue pairs that carry your meaning become indistinguishable.

This is not a rare edge case. In a stadium of 12 men, one of them cannot reliably read your red-versus-green status indicator. Which is exactly why the next section matters.

10. What This Means for Designers

Strip away the biology, and the science of comparative color vision hands you a set of concrete, defensible design principles. Here they are.

11. Conclusion: The World Is Rendered, Not Given

Every animal on Earth — from the tetrachromatic bird to the dichromatic dog to the twelve-sensor mantis shrimp — lives inside a rendered world. Light is the raw input; the nervous system is the renderer; and color is the final image. There is no single, correct version of the scene. There are as many versions as there are eyes to build them, and none of them is any more "real" than the others.

For a designer, this is not a philosophical curiosity. It is a practical fact with consequences in every file you touch. It means colorblindness is not an edge case but a predictable, measurable feature of a share of your audience. It means the "visible spectrum" is a spec, not a truth. It means the color spaces you work in are built on a model of one particular animal's eyes — yours — and that model is both a gift and a blind spot. It means, above all, that color is not something you apply to a design. It is something a viewer constructs, inside their own head, using hardware you can only partially predict.

So the next time you reach for a red and a green to mean "stop" and "go," pause and remember the genetics: one man in twelve is looking at your screen through a narrower red-green axis than you are. The next time you're told a display has "a billion colors," remember the mantis shrimp, who has twelve kinds of sensor and a worse sense of color than your dog's owner. And the next time you admire a flower, remember that its most important patterns are painted in a color you will never see — drawn not for you, but for the bee.

Color is not in the world. It's in the eye, and the brain, and the evolution that built them. Design accordingly — for the eyes that will actually do the seeing, not the ones you assume.

"The eye is not a camera; the world is not a photograph. Each species paints its own picture, in its own colors, on the same light."
— adapted from the comparative vision literature