Cool! I've never encountered the impact theory. I bought a large number of them ($5, at a post-xmas sale.) The vanes were glass cover-slips, one side soot, the other side white powder (perhaps smoke from burned Mg ribbon?) They had a variety of pressures (and glow differently for the same Tesla coil.) They also have a variety of frictions (spin them mechanically, let them coast down to 60RPM, then count the revolutions until they halt.) With twenty radiometers you can sort them, then keep the best ones (and give the others as xmas gifts.) During evening twilight, with only sunset-glow as illumination, the ones which had the brightest Tesla-coil glow and lowest friction are the ones which spin fastest. Some wouldn't spin at all under those conditions. I've seen versions which appear to have peeled-mica vanes and only one side blackened. Yours in the video looks like soot-and-mica to me. Not metal. (Does the silvery side reflect a red laser as much as aluminum foil does?) I've also seen some apparently using round glass cover-slips. Round microscope-slips are less common, but do exist. All of these are thermal insulators, and since the radiation-cooling is approximately the same for all coated surfaces, no major differences exist between black and white (and so the black side would be greatly heated via visible light, if both surfaces have near-identical cooling.) A metal-vane radiometer with thermally-conductive vanes would be very interesting! If you think you have one with metal vanes, break it open and check by bending the vanes (vent the vacuum, then scratch with a glass-cutter, to avoid imploding the bulb and destroying the vanes.) Or, if to the eye it looks exactly like rough mica, probably that's what it is. These are cheap, $12 each from sciplus if you buy two at a time (glass vanes, diamond shape. Some were mistreated in shipping, and occasionally a corner of a vane is cracked off and missing. Metal foil doesn't shatter like that.) Just now I FINALLY tried laser pointer. Green few-mW laser barely turns the more sensitive ones. But my violet laser 395nM gets it slowly spinning. TEST: aim the violet laser-spot at the center of the vane, observe the motion. Then move the laser-dot closer to the edge, near the upper tip of the diamond-shape (NOT to the outer tip, so no change in leverage.) Yes, it then spins slightly faster, barely noticeable, maybe under 10% faster? Aim at the center again, it visibly slows down (but hard to notice unless watching carefully.) I'll have to set up a laser-holder, rather than doing it by hand. Also, count the revolutions with a stopwatch, to get actual numbers for RPM changes, rather than trying to eyeball it. Put radiometer inside a freezer, and it does go backwards, only to stop after a minute ...then when removed, it slowly spins forward for half a minute. Also, freeze-spray the glass bulb. Or, big slab of dry ice held close. PS, if your radiometer has poor friction, you can bias it with diffuse window-sunlight, just to the threshold where it starts turning extremely slowly. That gets it past the static friction, so a small extra light-source may produce much more obvious results. (Heh, or buy all of them from a museum store, sort for sensitivity, then return the rest back to the store? Or just take a UV flashlight to the store, to open boxes and test RPMs.) ALSO, the glass bearing may have "good spots," and by bouncing the radiometer to reposition the needle, sometimes the sensitivity to light sources will change dramatically.

the electrostatic test is the most important test that says it all

Started off my morning right.

It was years since I had asked. Thank you!

I've been fascinated by those for decades. I almost bought one on a high school trip to the Chicago museum of Science and Industry in the 70's. I've never bought into their theories as to what made them spin and still don't. I've always wondered if they would spin in space without a glass dome and I suspect that's been tested. I also wonder if it's just a smaller version of our solar system since our earth spins in a semi vacuum too for no apparent reason that makes sense to me. A couple of observations; your light demonstration always directed the energy towards the black surface rather than testing both surfaces in a similar fashion. Your electrostatic demonstration focused on a flow of electrons from the generator poles which were both located on the white vane sides. That experiment in total darkness might prove interesting to see if it causes rotation without any light differences between colors. Your laser is a smaller version of our sun and it gives off heat from it's light which you concentrated on only the black surface. Would a more powerful laser with a color selected that reacts with white more than black and directed only on the white surfaces cause counter rotation? Would a DC high voltage source such as a flyback coil cause the same rotation when presented to one side or the other? Since radio waves can be focused; I wonder if there's been any attempt to build one of these devices with a pair of antennas inside so they could transmit radio waves from one side of the device to the other? They could then switch antennas and transmit in the opposite direction to see if there's any difference. I suspect this has all been tried by first year university students in colleges all around the world decade after decade but I never had the luxury of being one of these students. The curious mind is an animal that's always hungry.

Thermal transpiration is the most accurate theory by far, it follows the physics properly.

Indeed, very interesting. Wonder if the movement due to light has anything to do with Nitrogen gas spectral absorption lines?

No, a Crooke's radiometer works with hot water poured over it or cold ice pack draped over it if you wish so it is not dependent on light.