Cones and rods

Short answer

The retina has two types of light-detecting cells: cones, which work in daylight and make color vision possible, and rods, which work in dim light but can't tell colors apart on their own. The three types of cones are sensitive to different regions of the spectrum, and the brain extracts color by comparing the responses of these three types. In the dark, only rods are active, and since they're all the same type, there's nothing left to compare; we see the world in light and dark, without color.

Two kinds of receptors: rods and cones

The retina, which lines the back surface of the eye, is packed with cells that convert light into nerve signals. These cells' shapes even gave them their names: the long, thin ones are called rods, and the shorter ones that taper to a point at the outer end are called cones. How the eye focuses an image like a camera is the subject of BI-08; here we're looking at the two different materials of the "film layer" the image falls onto.

The numbers show a striking imbalance: rods outnumber cones by roughly twenty to one. But their distribution is just as different. Cones are tightly packed into a small pit at the center of the retina, the fovea; this is the region you use when reading a book or looking at a face. At the very center of the fovea there are almost no rods at all. Moving away from the center, cone density drops off quickly as rods become more numerous.

This distribution has an everyday consequence. You see color and detail best exactly where you're looking; toward the edge of your vision, colors fade and motion and changes in light stand out instead.

Three cone types and their sensitivity curves

The human eye has three types of cones, and each type responds most strongly at a different wavelength. Short-wavelength cones, or S cones, respond most to the violet-blue region; medium-wavelength cones, or M cones, to the green region; and long-wavelength cones, or L cones, respond most strongly in the yellowish-green region. The name "blue, green, and red cones" is common but misleading: the L cone's peak isn't in red but in yellowish green, and the three curves overlap widely with one another.

An important principle is hidden here: a single cone doesn't know color. An L cone can't report the wavelength of the light falling on it; it can only report how strongly it was stimulated. A small amount of strong yellowish-green light and a large amount of weak red light can produce the same response in an L cone. What makes color possible is that the same light stimulates the three cone types in different proportions, and the brain compares those proportions. The two major theories behind this comparison, trichromatic theory and opponent process theory, are covered in BI-10 and BI-11.

The three cone types aren't equal in number, either. S cones make up only a small fraction of all cones and are almost entirely absent from the very center of the fovea. Because of this, the eye's sharpest region is surprisingly poor at distinguishing very small blue details. The ratio of L to M cones, meanwhile, varies enormously from person to person; yet all of these people pass normal color vision tests. The visual system adapts to whatever mix of cones it has.

Night vision: rods take over

In bright daylight, rods are practically useless: they're so sensitive that they saturate, and their response stops changing. Cones take over vision. As light fades, the job changes hands. At twilight, the two systems work together; in true darkness there's not enough light left to stimulate the cones, and rods carry vision on their own.

This is the answer to the card's question: we can't see color in the dark because rods come in only one type. With a single receptor type, there's no way to tell whether a light produced a weak response because it was dim or because it fell at a wavelength the rod is less sensitive to. There's no second signal to compare it against. The result is a gray world made up only of light and dark. This is why flower colors disappear in a garden under moonlight.

Dark adaptation takes time. When you walk from a sunlit street into a dark movie theater, you see almost nothing for the first few minutes. Cones adapt a little first; after seven or eight minutes, rods become the most sensitive cells, and sensitivity keeps increasing for twenty to thirty minutes. A bright light undoes this adaptation within seconds, which is why people observing the night sky use a red flashlight — rods are only weakly sensitive to red light.

Because rods aren't found at the center of the fovea, night vision has an odd consequence. If you look straight at a very faint star, it can disappear, and it reappears when you shift your gaze slightly to the side. Astronomers call this technique averted vision: it means letting the light fall on the outer region, where rods are abundant.

The Purkinje effect

In 1819, the Czech physiologist Jan Evangelista Purkyně noticed that flowers that looked bright red in the afternoon looked very dark at dawn, while the leaves around them stayed relatively light. He concluded that the eye has two separate systems, one suited to bright light and the other to dim light. Today this phenomenon is called the Purkinje effect, after him.

The reason is that the two systems' sensitivity curves peak in different places. Daytime vision is most sensitive in the yellowish-green region, while the rods' curve is shifted toward blue-green, around 500 nanometers. As light fades and rods take over vision, long-wavelength reds drop into darkness quickly, while blue-greens stay relatively bright.

Example — faces in a garden at dusk

Picture yourself talking with someone in a garden on a summer evening. As the sun sets, the red roses darken first, then the leaves turn gray, and the blue-violet flowers vanish last. The faces of the people in front of you go through the same shift: the pink in their cheeks and the red of their lips are the first tones to disappear, and the warmth and undertone that distinguish different skin tones go with them. What's left is only the difference between light and dark. Once the light has dropped enough, every face, light-skinned or dark-skinned, reads in the same gray language; recognizing a face turns into reading shape and shadow instead of color.

Common misconception

Summary

1. The retina has two types of receptors: about 92 million rods work in dim light, and about 4.6 million cones work in daytime and color vision; cones are concentrated in the fovea.
2. The three cone types are most sensitive in the violet-blue, green, and yellowish-green regions; a single cone doesn't know color — color arises from comparing all three responses.
3. In the dark, color can't be seen because only the single type of rod is active; the shift of sensitivity toward blue-green in dim light is the Purkinje effect.

Test yourself

1. (Easy) Which cells in the retina enable vision in dim light?
2. (Easy) What is the region of the retina where cones are most densely packed called?
3. (Medium) Why can't we see color in the dark?
4. (Medium) Why do red flowers fall into darkness before leaves do in the evening?
5. (Hard) How does the principle that a single L cone can't know color explain why three cone types are necessary?
6. (Hard) A faint star disappearing when you look straight at it and reappearing when you look to the side results from which distribution feature?

Answers

1. Rods; they're highly sensitive to light but can't distinguish color because they come in only one type ("Two kinds of receptors").
2. The fovea; cones are at their densest in this pit at the center of the retina, and there are almost no rods at its very center ("Two kinds of receptors").
3. In the dark there's not enough light left to stimulate the cones, and rods alone carry vision. Since rods come in only one type, there's no second signal to compare the effects of different wavelengths against; the brain only receives light-dark information ("Night vision").
4. In dim light, rods take over vision, and the region rods are most sensitive to is blue-green; since their sensitivity to red is low, reds darken quickly. This is the Purkinje effect ("The Purkinje effect").
5. A cone only reports how much it was stimulated; different combinations of wavelength and intensity can produce the same response. When the responses of types with different sensitivity curves are compared, this ambiguity is resolved and color can be extracted ("Three cone types and their sensitivity curves").
6. There are next to no rods at the center of the fovea; looking to the side lets the faint light fall on the outer region, where rods are abundant ("Night vision").

Sources

• Curcio, C. A., Sloan, K. R., Kalina, R. E., and Hendrickson, A. E., "Human photoreceptor topography," Journal of Comparative Neurology 292, 1990.
• Bowmaker, J. K. and Dartnall, H. J. A., "Visual pigments of rods and cones in a human retina," The Journal of Physiology, 1980.
• Hofer, H., Carroll, J., Neitz, J., Neitz, M., and Williams, D. R., "Organization of the Human Trichromatic Cone Mosaic," Journal of Neuroscience 25(42): 9669–9679, 2005.
• York University, "Spectral Sensitivity curves of rods and cones"; "Dark adaptation."
• Hanover College, "Purkinje Shift" interactive demonstration.
• EyeWiki, "Color Vision."

Next step

BI-10 takes up the question this lesson leaves open: how all colors are produced from the responses of the three cone types, told through the 19th century's color-matching experiments and the Young–Helmholtz theory, and why screens use only three colors. BI-11 then moves to the second stage of the comparison: opponent process theory, with its red–green, blue–yellow, and light–dark channels, and why we can't imagine a color like "reddish green."