I wrote this article the mid-1990s before many of the wonderful online sources of information that are now available. For more accurate and in-depth information you may wish to refer to:
A purple Vinca
a “purple vinca” flower

What makes one color different than another? Sure, we can all easily distinguish red from yellow, and we may hold an affinity for some colors more than others. But when we think analytically about the nature of colors, aren't they all essentially the same? Is chartreuse any less of a color than blue just because we may be less familiar with it?

As it turns out, almost all colors are in fact just as real as any other color--with the exception of purple. Purple is mostly an artifact of the way our eyes and brains perceive colors and try to make sense out of a world which doesn't always make sense. The explanation of what makes purple radically different than say orange is rather easy. As we begin to explore the colors (or more precisely the hues), lets start with something that most of us are familiar with, the standard color wheel.

The standard
color wheel

You can notice several things from looking at the color wheel. The first is how smoothly one color blends into the next. But more subtly is how different regions appear different in more ways than just their color. For instance, the redish region appears much larger in size than the tiny yellow one. Also the green and yellow regions appear much brighter than the blue region. However, the colors are perfectly geometrically aligned and are at the same intensities or illuminosity levels. Those differences we see are in fact mostly artifacts of how our eyes and brains work and are not inherent to the colors themselves. Now lets look at pure colors from a more scientific model, the color frequency spectrum of visible light (which is what we see in a rainbow).

The visible light spectrum

At first glance you may think we just took the color wheel, snipped it open and stretched it out straight. But look again, do you notice that something seems to be missing? All of the purple region which was in the color wheel is nowhere to be found in the rainbow spectrum. You may be wondering about the so called ultraviolet light. Actually the name ultraviolet is a misnomer; it is not visible to the human eye, and so does not appear violet. That term probably arose from the observation that when ultraviolet light hits some materials it causes them to glow with an apparent purple color. Thus is it not the ultraviolet light itself which you see, but rather a different form of light which is emitted by those excited objects.

We now know that purple can not be a single pure frequency of light as it does not appear in the visible spectrum. But we also know that we can see something which we call purple, and it is definitely not blue or green or even orange. Purple must be something, but what is it exactly? It turns out that what we observe as purple is always a mixture of other colors, usually some form of red and some form of blue. So purple is not pure in that it is not a single color. We can see pure orange light, but there is no such thing as pure puple light. Of course this should not shock us too much as most people know that the color "white" has the same property; it is not a single pure color either but is a blended mixture of many other colors.

Let us now consider colors from a different perspective. There is a whole field of study in color sciences, and yes, you can actually get an advanced degree in the subject of color. The color wheel is an easy way to demonstrate the basic relationships between colors, but is not very precise. A more frequently used aid in colorimetry is the CIE Chromaticity Diagram, a standardized color model formulated by the Commission Internationale de l'Éclairage, after a large and carefully measured study of human sight. Actually there are two standards, a so called 2° tristimuli model and a 10° model. The former is most useful as it models color vision in normal daylight brightness, whereas the later is beter suited for modeling human vision at very low light levels, such as at night. Now for the diagram:

The CIE
color plane
The 2° CIE Chromaticity Diagram

It should be said before going much further that the above figure is not the whole CIE diagram. The real diagram is actually a complex 3-dimensional object which is somewhat conical in shape. What you see in the above figure is a cross sectional slice near the base or fat part of the cone. The cross section above contains all the hues but not all of the shades and tints, which includes variations in brightness or darkness. The whole CIE diagram is supposed to represent all colors which are visible to the normal human eye, and the borders of the shape represent the boundary at which human vision can no longer observe color. Upon close inspection this diagram reveals quite a lot about the nature of human vision. An equivalent diagram for insects would look quite a bit different.

Other than the asymmetrical shape, one of the first things you notice about the CIE diagram is how large the green area is, while the blue area is rather tiny. Why is that? As taught in high school biology, the human retinaEye retinal response curves is packed with two kinds of light sensors, cones and rods. The cones primarily serve to detect color or hue, while the rods are mainly used for brightness levels only. Well, among the cones there are three subtypes, which are generally called red, green, and blue cones. Mainly due to differing organic pigments, those cones are highly sensitive to redish, greenish, or bluish light, but otherwise insensitive to other colors. They are most sensitive to the color frequencies of 430, 530, and 560 nm respectively, but at differing sensitivities (refer to the diagram at the right). The three colors of cones do not come in equal numbers. For every blue cone there may be upward of 40 or more cones of the other two colors. This fact along with other subtle properties means that the human eye is many times more sensitive to green light than it is to blue light, with red being just somewhat less sensitive than green. This is why the blue region is so much smaller in the CIE diagram; the human eye is just not very good at seeing blue.

Before we get back to the discussion on purple it is worth looking at one other observation about the CIE diagram. Notice the brighter triangular region in the middle of the diagram. That triangle is not really part of the true diagram, but rather is a highlight to indicates the set of colors which a typical computer monitor or color television can accurately reproduce. The colors outside of this triangle are observable by the human eye, but are physically impossible for a monitor to display (since you are viewing this on a computer screen what you are seeing is already being distorted by the monitor). So even though an expensive computer graphics card may boast of being able to display several million colors, it can never come close to being able to display all the colors which the human eye can distinguish. The three corners of the triangle represent the precise colors produced by each of the red, green, and blue phosphor dots which make up the screen. In general, the set of colors which a device can reproduce is called its gamut. Other devices such as color inkjet printers, scanners, or even digital cameras have their own gamuts; and although they mostly overlap, they are almost never the same. This is one reason why it can be so hard to print out some pictures which look great on your monitor but turn out with distorted or muddy colors once put on paper (the gamut of most inkjet printers is much smaller than even that of the RGB computer monitor).