Sirius is setting earlier and earlier so I’d better ask this now.
I see stars twinkle and I see stars with color, but to me, Sirius is the only one I see that twinkles with flashes of color, like a diamond with good dispersion. Why is it the only one? I was thinking that, being the brightest star in the heavens* it was the only one bright enough to tickle the cones in my eye. If that were the case, stars that show color, like Betelgeuse, being dimmer shouldn’t show color, just look grey or something.
So, why is Sirius the only one that truly looks like a diamond in the sky? Or am I just seeing things?
Canopus does the same. It’s a southern star, but you should be able to see it from Mesa in the winter, low in the sky to the south. It is bright enough to look like an aircraft approaching to land, and it flashes color too.
That is not it. You are not seeing stars with rod vision, you are seeing them with cones anyway. It is just that, in most cases, they are activating all three cone classes about equally, so they look white. If you were seeing them with the rods, you would not be able to see them at all if you looked right at them, and they would be impossible to localize accurately.
Betelgeuse looks red (to the extent that it does to the naked eye) because it is very red, but Sirius is pretty much white anyway. As others have already surmised, if you are seeing flashes of color it is almost certainly caused by refraction through inhomogeneities in the atmosphere, which are themselves caused by air currents of different temperatures. This is what causes star twinkling in general. I must admit that I have never noticed flashes of color from Sirius or any other star or planet myself, but I suppose it wold be possible in principle for the atmospheric refraction to cause color effects too, by splitting the spectrum. I expect you only see it for Sirius because it is so much brighter than most stars.
Possibly it might also have to do with cone distribution in your eyes. The red and the green sensitive cones are very unevenly distributed, in random clumps of each type, and the actual distribution is different in different individuals. (Blue sensitive cones are distributed more regularly, consistently, and evenly.) Possibly your red and green cones are even more clumped than most other people’s are. It would not affect your color vision noticeably in most conditions, but it seems conceivable that it might produce the sort of effect you describe under the rather special conditions of looking at a bright, twinkling star, if the point of starlight were falling first on one cone clump and then on another (moving either because of the atmospheric refraction or because of the involuntary small movements of your eyes).
I haven’t looked for Canopus. It is pretty low in the south. I’m not sure it would show above the murk and sky glow rampant here. Good point on the dimmer stars disappearing when looked at directly if only rod vision was involved. They definitely do not.
Phil Plait, The Bad Astronomer, has an explanation here (toward the bottom of the post) and he posted a neat time-lapse photo here that was taken by another astronomer.
Coolness. At least the photo proves it’s not my imagination. He doesn’t really answer the question of why the phenomena is more easily seen with brighter stars though, just mentions that it is.