Why purple (violet), well ...
I've been doing more thinking ... it can be a dangerous thing, but bear with me. The most perplexing question about the whole thing is why is the color purple (or something that sort of looks like purple)? I think that the answer might have something to do with what the index of refraction curve looks like for optical glass. Crown glass and flint glass are two types of glass commonly used in optics. Higher quality optics use both mated together in order to take advantage of their complementary refractive characteristics in order to minimize chromatic aberrations. Cheap optics usually just use crown glass or maybe only both types for the objective element only.
I discovered that the index of refraction curve for crown glass is not the nice gentle curve that I expected, but instead changes slope very abruptly at the short wavelengths (blue end of the spectrum). See the graph below.
please log in to view hosted photos in full size.
The index of refraction to the right of the green line (approximately 400 nanometers which is blue-violet) is approximately a linear function of wavelength. At the infrared end of the graph (approximately 1600 nanometers), the index of refraction is approximately 1.50. The longest wavelength visible light is approximately 700 nanometers and the index of refraction is still approximately 1.50. At 400 nanometers, the index of refraction is approximately 1.53.
Between the green line and the red line (approximately 250 nanometers, which is slightly past the range of of visible light and into ultraviolet) the index of refraction changes much more rapidly with wavelength. The index of refraction at the red line is approximately 1.57.
The left end of the x=axis represents a wavelength of approximately 200 nanometers and the index of refraction is somewhat greater than 1.6. Why mention light outside the range of human vision? Well, film and camera sensors are not human and they can see things that we can't. However the filters will take care of most of this ultraviolet light.
The most important thing from the graph is that the amount of dispersion at the blue-violet end of the visible spectrum is significant compared to the rest of the spectrum, so while corrective optics could address chromatic aberration at longer wavelengths, there would still be residual fringing at the violet end of the spectrum. The effect is weak so it would be masked in a multicolored image, but would be much more visible as a fringe around the perimeter of an area of white light.
Thanks for your indulgence and now I am ready for the funny farm. Geez, this is worse than pixel peeping, sensor dust, and crop factor all rolled into one.
Purple fringing can be also seen when looking through a refractor astronomical telescope even without bad CA and has been well collimated. Since reflectors don't have CA, I suppose that they also do not have purple fringing.
