Actually, I have run across a couple sources, to include the last one that I posted above, that show that the red cone does have an island of blue wavelengths to which it is sensitive.
Sunspace: thanks for the good links.
MEBuckner: no problem, here comes more.
Punoqllads: Apparently many people agree with you on the cone names. A lot of those studying the mechanism of color perception refer to the as S, M and L cones (short, medium and long wavelength cones).
Here are some more links. (Never let it be said that I did not cite.) The first one looks at the two leading theories in color perception: the Trichromatic theory which is our additive and subtractive color mixing theory and the Opponent-Process theory which leaps farther away from our wavelength-grounded notions to opposing colors. The second link also refers to these two theories and a whole lot more. From these links you will see that both theories are old and have good points and bad points.
(On the first link: click on the 3 different slides.)
(On the second link follow the Parts at the bottom the page to the slides. I think there are over 40 slides. All of the slides have interesting information on color perception, however, after Part 3, the info is much less pertinent to this discussion. In Parts 1 and 3, there is a lot of tangential information, too. Use the big red arrows at the bottom of the slides for easy navigation.)
For a different twist, read Oliver Sacks’ An Antropologist on Mars. The book tells of (I believe) 7 people with unusual neurological conditions. One is an artist, who experiences a mild stroke, and loses his conception of color. No injury to his eyes. In fact, he is extremely attuned to the variations of grey. He also presents a fascinating patient because, as an artist, he is very educated as to the characteristics of various colors in terms of wavelength and pigment. But he loses his very memory of color. Describing a banana as “yellow” means nothing to him. Darned intersting reading.
Sacks wrote another book, The Island of the Colorblind concerning a couple of communities in the S Pacific where a significnat percentage of the population shares a genetic trait of complete colorblindness.
If you have not read his books, I heartily recommend them.
No, the web page you note is showing the CIE color-matching funcitons x-bar[sub]lambda[/sub], y-bar[sub]lambda[/sub], and z-bar[sub]lambda[/sub], for the 1931 CIE X, Y, and Z primaries. You can tell this isn’t the spectral-response curves of the cones by seeing how high the blue-violet hump is, compared to the other humps. The S cone (thanks for finding that term, I much prefer it) is significantly weaker in its response curve than the M and L cones.
The cone responses are gaussian humps, centered at about 440 nm, 545 nm, and 580 nm for the S, M, and L cones, respectively. However, the S cone’s response at its center is only about 10% as strong as the M and L cones’ response at their centers.
My reference for the above information is not online, alas, but you should be able to find it in any bookstore that has a decent computer graphics section. It is Computer Graphics, Principles and Practice, 2nd edition, by Foley, van Dam, et.al. Chapter 13, section 2 (pp. 574-84) is on chromatic color, and figures 13.18 and 13.22 show the spectral-response of the cones, and the CIE color-matching functions, respectively.
The eye has three different types of color receptors. So you can think of the full set of visible colors as a three-sided pyramid with one edge representing red, one edge representing green, and one edge representing blue. The position of any particular color within this “color solid” is determined by how much it stimulates each of the red, green and blue receptors. So, for example, various intensities of pure red are located on the red axis, and various intensities of white are located along a line that runs though the center of the solid.
Now, imagine that we highlight all the colors in the color solid that represent monochromatic light – light that consists of a single wavelength. The color of monochromatic light is a function of two variables: wavelength and intensity. This means that the set of all pure colors will be a 2-D dimensional sheet embedded in the 3-D dimensional color solid.
As it happens this sheet is U-shaped. The U starts in the red part of the color solid, loops down through green and doubles back to blue. The colors lying in the open part of the U are the colors you’re talking about: the pinks and purples.
But your eye doesn’t know anything about pure wavelengths. All it knows is that it’s getting three signals of different intensities from the three different types of receptor. If the red and the blue receptors both fire equally the eye sees purple. If the red and the green receptors both fire equally the eye sees yellow. But the brain can’t tell whether the red and green stimulus is caused by a single yellow light or a mixture of red and green lights – both will cause the same response in the eys. For purple there’s such no monochromatic light – the only way to get it is to mix red and blue.
So the answer to your question is: We see color as a wheel because we don’t see the true spectrum – only varying values of three sample signals. We see red-purple-blue next to each other because the true spectrum cuts a U through our 3-receptor color space. Red is at the top of one arm of the U, yellow is halfway down, green is at the base, cyan is halfway up the other arm, and blue is at the top of the other arm. And purple lies in the gap.