How Your Eyes See Color: Cones & Trichromacy

Color does not arrive in your eye. It is built there, from just three signals.

Short answer: color vision starts with cone cells in the retina. Most people have three types: short (S), medium (M) and long (L) wavelength cones. Incoming light stimulates each type by a different amount, and your brain reads that pattern of relative signals as a color. This three-receptor design is called trichromatic vision.

Three kinds of cone

The retina holds two families of light-sensing cell. Rods handle dim light and motion, while cones handle color and fine detail in brighter conditions. A typical human retina carries several million cones, packed most densely in the fovea, the tiny central pit you use whenever you look straight at something. Cones come in three types, each tuned to a different part of the spectrum. They take their names from the wavelengths they respond to most strongly: S (short), M (medium) and L (long). An important point sits underneath all of this: no single cone sees a color on its own. Color emerges only when the brain compares the outputs of all three.

Where each cone peaks

Each cone type has a peak sensitivity, the wavelength it responds to most strongly. Direct measurements of the cone pigments, notably by Bowmaker and Dartnall in 1980 and later refined in the widely used Stockman and Sharpe estimates, place the peaks roughly as shown below.

Cone	Common name	Peak sensitivity	Responds most to
S	Short / “blue”	≈ 420 nm	Violet-blue light
M	Medium / “green”	≈ 534 nm	Green light
L	Long / “red”	≈ 564 nm	Yellow-green through red light

Two things stand out. First, none of the cones actually peaks at pure red. Even the L cone tops out in the yellow-green and then keeps responding, more weakly, into the reds, which is why the popular nicknames “blue”, “green” and “red” are convenient but a little misleading. Second, the M and L peaks sit remarkably close together, around 534 nm and 564 nm, only about thirty nanometers apart. That near overlap is not a design flaw. It is exactly what lets the eye separate fine differences across the yellow, orange and red range, where a great deal of everyday color judgment happens.

Color as a pattern of three signals

The central idea of trichromacy is that a color is not a property carried by any one cone. It is the relative strength of the three cone signals taken together. Feed the eye a single wavelength of pure yellow light and it produces a particular balance of S, M and L responses. Now feed it a suitable mixture of red and green light. If that mixture happens to produce the same balance of responses, the two look identical, even though the physical light reaching the eye is completely different. Pairs of physically different lights that look the same are called metamers, and they are the reason a screen built from only red, green and blue subpixels can convincingly reproduce a yellow banana or an orange sunset. Your brain has only those three numbers to work with, and every color you perceive is its reading of their balance.

Rods, cones, and why night looks gray

Because cones need a reasonable amount of light to respond, color vision fades as the light drops. In dim conditions the cones fall quiet and the far more sensitive rods take over. Rods come in a single type and carry no color information, so a moonlit garden reads as shades of gray even though the flowers are as colorful as ever. The same handoff explains a familiar stargazing trick: a faint star is often easier to see slightly off to the side than dead ahead, because the rod-rich outer retina catches weak light that the cone-packed center misses.

A common misunderstanding

It is tempting to picture each cone as a little meter that reports “this much red”, “this much green”, “this much blue”. In reality every cone answers to a broad band of wavelengths, and the bands overlap heavily. A mid-green light stimulates the M and L cones strongly and the S cone barely at all, yet no cone is ever truly silent and none acts alone. Color is a comparison, not three separate readings. This is also why there is no single “color-detecting” wavelength and no pixel that is really red in itself. The redness lives in the pattern of response, not in the light.

A theory confirmed by measurement

The three-receptor idea is surprisingly old. Thomas Young suggested in 1802 that the eye needs only a few receptor types to explain how colored lights mix, and Hermann von Helmholtz developed the proposal into what is now called the Young–Helmholtz trichromatic theory. For a long time it stood as an inference from color-mixing experiments rather than something anyone had observed directly. Then, well over a century later, researchers measured the light absorption of individual cone pigments and found exactly three, with peaks close to the values in the table above. A prediction drawn from mixing colored lamps turned out to match the biology of the retina.

Put your three cones to work

Everything you see is your brain reading three signals, so it makes sense to train the reading. In Hueddle, you narrow in on a hidden target color from gentle feedback, sharpening how precisely you interpret those cone responses. It is a direct, playful way to exercise the trichromatic system this page describes.

Play Hueddle →

Frequently asked questions

How do eyes see color?

Color vision starts with cone cells in the retina. Most people have three kinds of cone, each most sensitive to a different band of wavelengths. When light arrives, the three cone types respond by different amounts, and the brain reads that pattern of relative signals as a particular color.

What are cone cells?

Cones are light-sensing cells in the retina responsible for color vision and for fine detail in bright light. There are three types, named short (S), medium (M) and long (L) for the part of the spectrum each responds to most strongly. They are packed most densely in the fovea, the small central pit you use for sharp, direct looking.

What are the peak wavelengths of the three cones?

Measurements of the cone pigments place the peaks at roughly 420 nm for the S cone, 534 nm for the M cone and 564 nm for the L cone. Notably, even the L (“red”) cone peaks in the yellow-green rather than in pure red, and the M and L peaks sit close together, only about 30 nm apart.

What is trichromatic vision?

Trichromatic vision means color is built from three cone signals. Because there are three receptor types, any color a typical person sees can be described as a particular combination of the S, M and L responses. This is the basis of the Young–Helmholtz trichromatic theory, which direct measurement of the cone pigments later confirmed.

Why does color fade in dim light?

Cones need a fair amount of light to respond. In dim conditions the cones fall quiet and vision passes to the rods, which are far more sensitive but come in a single type and carry no color information. That is why a moonlit scene reads as shades of gray even though the objects in it are as colorful as ever.